Add files using upload-large-folder tool

.venv/lib/python3.11/site-packages/numpy/f2py/src/fortranobject.c

@@ -0,0 +1,1423 @@
+#define FORTRANOBJECT_C
+#include "fortranobject.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stdarg.h>
+#include <stdlib.h>
+#include <string.h>
+
+/*
+  This file implements: FortranObject, array_from_pyobj, copy_ND_array
+
+  Author: Pearu Peterson <[email protected]>
+  $Revision: 1.52 $
+  $Date: 2005/07/11 07:44:20 $
+*/
+
+int
+F2PyDict_SetItemString(PyObject *dict, char *name, PyObject *obj)
+{
+    if (obj == NULL) {
+        fprintf(stderr, "Error loading %s\n", name);
+        if (PyErr_Occurred()) {
+            PyErr_Print();
+            PyErr_Clear();
+        }
+        return -1;
+    }
+    return PyDict_SetItemString(dict, name, obj);
+}
+
+/*
+ * Python-only fallback for thread-local callback pointers
+ */
+void *
+F2PySwapThreadLocalCallbackPtr(char *key, void *ptr)
+{
+    PyObject *local_dict, *value;
+    void *prev;
+
+    local_dict = PyThreadState_GetDict();
+    if (local_dict == NULL) {
+        Py_FatalError(
+                "F2PySwapThreadLocalCallbackPtr: PyThreadState_GetDict "
+                "failed");
+    }
+
+    value = PyDict_GetItemString(local_dict, key);
+    if (value != NULL) {
+        prev = PyLong_AsVoidPtr(value);
+        if (PyErr_Occurred()) {
+            Py_FatalError(
+                    "F2PySwapThreadLocalCallbackPtr: PyLong_AsVoidPtr failed");
+        }
+    }
+    else {
+        prev = NULL;
+    }
+
+    value = PyLong_FromVoidPtr((void *)ptr);
+    if (value == NULL) {
+        Py_FatalError(
+                "F2PySwapThreadLocalCallbackPtr: PyLong_FromVoidPtr failed");
+    }
+
+    if (PyDict_SetItemString(local_dict, key, value) != 0) {
+        Py_FatalError(
+                "F2PySwapThreadLocalCallbackPtr: PyDict_SetItemString failed");
+    }
+
+    Py_DECREF(value);
+
+    return prev;
+}
+
+void *
+F2PyGetThreadLocalCallbackPtr(char *key)
+{
+    PyObject *local_dict, *value;
+    void *prev;
+
+    local_dict = PyThreadState_GetDict();
+    if (local_dict == NULL) {
+        Py_FatalError(
+                "F2PyGetThreadLocalCallbackPtr: PyThreadState_GetDict failed");
+    }
+
+    value = PyDict_GetItemString(local_dict, key);
+    if (value != NULL) {
+        prev = PyLong_AsVoidPtr(value);
+        if (PyErr_Occurred()) {
+            Py_FatalError(
+                    "F2PyGetThreadLocalCallbackPtr: PyLong_AsVoidPtr failed");
+        }
+    }
+    else {
+        prev = NULL;
+    }
+
+    return prev;
+}
+
+static PyArray_Descr *
+get_descr_from_type_and_elsize(const int type_num, const int elsize)  {
+  PyArray_Descr * descr = PyArray_DescrFromType(type_num);
+  if (type_num == NPY_STRING) {
+    // PyArray_DescrFromType returns descr with elsize = 0.
+    PyArray_DESCR_REPLACE(descr);
+    if (descr == NULL) {
+      return NULL;
+    }
+    descr->elsize = elsize;
+  }
+  return descr;
+}
+
+/************************* FortranObject *******************************/
+
+typedef PyObject *(*fortranfunc)(PyObject *, PyObject *, PyObject *, void *);
+
+PyObject *
+PyFortranObject_New(FortranDataDef *defs, f2py_void_func init)
+{
+    int i;
+    PyFortranObject *fp = NULL;
+    PyObject *v = NULL;
+    if (init != NULL) { /* Initialize F90 module objects */
+        (*(init))();
+    }
+    fp = PyObject_New(PyFortranObject, &PyFortran_Type);
+    if (fp == NULL) {
+        return NULL;
+    }
+    if ((fp->dict = PyDict_New()) == NULL) {
+        Py_DECREF(fp);
+        return NULL;
+    }
+    fp->len = 0;
+    while (defs[fp->len].name != NULL) {
+        fp->len++;
+    }
+    if (fp->len == 0) {
+        goto fail;
+    }
+    fp->defs = defs;
+    for (i = 0; i < fp->len; i++) {
+        if (fp->defs[i].rank == -1) { /* Is Fortran routine */
+            v = PyFortranObject_NewAsAttr(&(fp->defs[i]));
+            if (v == NULL) {
+                goto fail;
+            }
+            PyDict_SetItemString(fp->dict, fp->defs[i].name, v);
+            Py_XDECREF(v);
+        }
+        else if ((fp->defs[i].data) !=
+                 NULL) { /* Is Fortran variable or array (not allocatable) */
+            PyArray_Descr *
+            descr = get_descr_from_type_and_elsize(fp->defs[i].type,
+                                                   fp->defs[i].elsize);
+            if (descr == NULL) {
+                goto fail;
+            }
+            v = PyArray_NewFromDescr(&PyArray_Type, descr, fp->defs[i].rank,
+                                     fp->defs[i].dims.d, NULL, fp->defs[i].data,
+                                     NPY_ARRAY_FARRAY, NULL);
+            if (v == NULL) {
+                Py_DECREF(descr);
+                goto fail;
+            }
+            PyDict_SetItemString(fp->dict, fp->defs[i].name, v);
+            Py_XDECREF(v);
+        }
+    }
+    return (PyObject *)fp;
+fail:
+    Py_XDECREF(fp);
+    return NULL;
+}
+
+PyObject *
+PyFortranObject_NewAsAttr(FortranDataDef *defs)
+{ /* used for calling F90 module routines */
+    PyFortranObject *fp = NULL;
+    fp = PyObject_New(PyFortranObject, &PyFortran_Type);
+    if (fp == NULL)
+        return NULL;
+    if ((fp->dict = PyDict_New()) == NULL) {
+        PyObject_Del(fp);
+        return NULL;
+    }
+    fp->len = 1;
+    fp->defs = defs;
+    if (defs->rank == -1) {
+      PyDict_SetItemString(fp->dict, "__name__", PyUnicode_FromFormat("function %s", defs->name));
+    } else if (defs->rank == 0) {
+      PyDict_SetItemString(fp->dict, "__name__", PyUnicode_FromFormat("scalar %s", defs->name));
+    } else {
+      PyDict_SetItemString(fp->dict, "__name__", PyUnicode_FromFormat("array %s", defs->name));
+    }
+    return (PyObject *)fp;
+}
+
+/* Fortran methods */
+
+static void
+fortran_dealloc(PyFortranObject *fp)
+{
+    Py_XDECREF(fp->dict);
+    PyObject_Del(fp);
+}
+
+/* Returns number of bytes consumed from buf, or -1 on error. */
+static Py_ssize_t
+format_def(char *buf, Py_ssize_t size, FortranDataDef def)
+{
+    char *p = buf;
+    int i;
+    npy_intp n;
+
+    n = PyOS_snprintf(p, size, "array(%" NPY_INTP_FMT, def.dims.d[0]);
+    if (n < 0 || n >= size) {
+        return -1;
+    }
+    p += n;
+    size -= n;
+
+    for (i = 1; i < def.rank; i++) {
+        n = PyOS_snprintf(p, size, ",%" NPY_INTP_FMT, def.dims.d[i]);
+        if (n < 0 || n >= size) {
+            return -1;
+        }
+        p += n;
+        size -= n;
+    }
+
+    if (size <= 0) {
+        return -1;
+    }
+
+    *p++ = ')';
+    size--;
+
+    if (def.data == NULL) {
+        static const char notalloc[] = ", not allocated";
+        if ((size_t)size < sizeof(notalloc)) {
+            return -1;
+        }
+        memcpy(p, notalloc, sizeof(notalloc));
+        p += sizeof(notalloc);
+        size -= sizeof(notalloc);
+    }
+
+    return p - buf;
+}
+
+static PyObject *
+fortran_doc(FortranDataDef def)
+{
+    char *buf, *p;
+    PyObject *s = NULL;
+    Py_ssize_t n, origsize, size = 100;
+
+    if (def.doc != NULL) {
+        size += strlen(def.doc);
+    }
+    origsize = size;
+    buf = p = (char *)PyMem_Malloc(size);
+    if (buf == NULL) {
+        return PyErr_NoMemory();
+    }
+
+    if (def.rank == -1) {
+        if (def.doc) {
+            n = strlen(def.doc);
+            if (n > size) {
+                goto fail;
+            }
+            memcpy(p, def.doc, n);
+            p += n;
+            size -= n;
+        }
+        else {
+            n = PyOS_snprintf(p, size, "%s - no docs available", def.name);
+            if (n < 0 || n >= size) {
+                goto fail;
+            }
+            p += n;
+            size -= n;
+        }
+    }
+    else {
+        PyArray_Descr *d = PyArray_DescrFromType(def.type);
+        n = PyOS_snprintf(p, size, "%s : '%c'-", def.name, d->type);
+        Py_DECREF(d);
+        if (n < 0 || n >= size) {
+            goto fail;
+        }
+        p += n;
+        size -= n;
+
+        if (def.data == NULL) {
+            n = format_def(p, size, def);
+            if (n < 0) {
+                goto fail;
+            }
+            p += n;
+            size -= n;
+        }
+        else if (def.rank > 0) {
+            n = format_def(p, size, def);
+            if (n < 0) {
+                goto fail;
+            }
+            p += n;
+            size -= n;
+        }
+        else {
+            n = strlen("scalar");
+            if (size < n) {
+                goto fail;
+            }
+            memcpy(p, "scalar", n);
+            p += n;
+            size -= n;
+        }
+    }
+    if (size <= 1) {
+        goto fail;
+    }
+    *p++ = '\n';
+    size--;
+
+    /* p now points one beyond the last character of the string in buf */
+    s = PyUnicode_FromStringAndSize(buf, p - buf);
+
+    PyMem_Free(buf);
+    return s;
+
+fail:
+    fprintf(stderr,
+            "fortranobject.c: fortran_doc: len(p)=%zd>%zd=size:"
+            " too long docstring required, increase size\n",
+            p - buf, origsize);
+    PyMem_Free(buf);
+    return NULL;
+}
+
+static FortranDataDef *save_def; /* save pointer of an allocatable array */
+static void
+set_data(char *d, npy_intp *f)
+{           /* callback from Fortran */
+    if (*f) /* In fortran f=allocated(d) */
+        save_def->data = d;
+    else
+        save_def->data = NULL;
+    /* printf("set_data: d=%p,f=%d\n",d,*f); */
+}
+
+static PyObject *
+fortran_getattr(PyFortranObject *fp, char *name)
+{
+    int i, j, k, flag;
+    if (fp->dict != NULL) {
+        PyObject *v = _PyDict_GetItemStringWithError(fp->dict, name);
+        if (v == NULL && PyErr_Occurred()) {
+            return NULL;
+        }
+        else if (v != NULL) {
+            Py_INCREF(v);
+            return v;
+        }
+    }
+    for (i = 0, j = 1; i < fp->len && (j = strcmp(name, fp->defs[i].name));
+         i++)
+        ;
+    if (j == 0)
+        if (fp->defs[i].rank != -1) { /* F90 allocatable array */
+            if (fp->defs[i].func == NULL)
+                return NULL;
+            for (k = 0; k < fp->defs[i].rank; ++k) fp->defs[i].dims.d[k] = -1;
+            save_def = &fp->defs[i];
+            (*(fp->defs[i].func))(&fp->defs[i].rank, fp->defs[i].dims.d,
+                                  set_data, &flag);
+            if (flag == 2)
+                k = fp->defs[i].rank + 1;
+            else
+                k = fp->defs[i].rank;
+            if (fp->defs[i].data != NULL) { /* array is allocated */
+                PyObject *v = PyArray_New(
+                        &PyArray_Type, k, fp->defs[i].dims.d, fp->defs[i].type,
+                        NULL, fp->defs[i].data, 0, NPY_ARRAY_FARRAY, NULL);
+                if (v == NULL)
+                    return NULL;
+                /* Py_INCREF(v); */
+                return v;
+            }
+            else { /* array is not allocated */
+                Py_RETURN_NONE;
+            }
+        }
+    if (strcmp(name, "__dict__") == 0) {
+        Py_INCREF(fp->dict);
+        return fp->dict;
+    }
+    if (strcmp(name, "__doc__") == 0) {
+        PyObject *s = PyUnicode_FromString(""), *s2, *s3;
+        for (i = 0; i < fp->len; i++) {
+            s2 = fortran_doc(fp->defs[i]);
+            s3 = PyUnicode_Concat(s, s2);
+            Py_DECREF(s2);
+            Py_DECREF(s);
+            s = s3;
+        }
+        if (PyDict_SetItemString(fp->dict, name, s))
+            return NULL;
+        return s;
+    }
+    if ((strcmp(name, "_cpointer") == 0) && (fp->len == 1)) {
+        PyObject *cobj =
+                F2PyCapsule_FromVoidPtr((void *)(fp->defs[0].data), NULL);
+        if (PyDict_SetItemString(fp->dict, name, cobj))
+            return NULL;
+        return cobj;
+    }
+    PyObject *str, *ret;
+    str = PyUnicode_FromString(name);
+    ret = PyObject_GenericGetAttr((PyObject *)fp, str);
+    Py_DECREF(str);
+    return ret;
+}
+
+static int
+fortran_setattr(PyFortranObject *fp, char *name, PyObject *v)
+{
+    int i, j, flag;
+    PyArrayObject *arr = NULL;
+    for (i = 0, j = 1; i < fp->len && (j = strcmp(name, fp->defs[i].name));
+         i++)
+        ;
+    if (j == 0) {
+        if (fp->defs[i].rank == -1) {
+            PyErr_SetString(PyExc_AttributeError,
+                            "over-writing fortran routine");
+            return -1;
+        }
+        if (fp->defs[i].func != NULL) { /* is allocatable array */
+            npy_intp dims[F2PY_MAX_DIMS];
+            int k;
+            save_def = &fp->defs[i];
+            if (v != Py_None) { /* set new value (reallocate if needed --
+                                   see f2py generated code for more
+                                   details ) */
+                for (k = 0; k < fp->defs[i].rank; k++) dims[k] = -1;
+                if ((arr = array_from_pyobj(fp->defs[i].type, dims,
+                                            fp->defs[i].rank, F2PY_INTENT_IN,
+                                            v)) == NULL)
+                    return -1;
+                (*(fp->defs[i].func))(&fp->defs[i].rank, PyArray_DIMS(arr),
+                                      set_data, &flag);
+            }
+            else { /* deallocate */
+                for (k = 0; k < fp->defs[i].rank; k++) dims[k] = 0;
+                (*(fp->defs[i].func))(&fp->defs[i].rank, dims, set_data,
+                                      &flag);
+                for (k = 0; k < fp->defs[i].rank; k++) dims[k] = -1;
+            }
+            memcpy(fp->defs[i].dims.d, dims,
+                   fp->defs[i].rank * sizeof(npy_intp));
+        }
+        else { /* not allocatable array */
+            if ((arr = array_from_pyobj(fp->defs[i].type, fp->defs[i].dims.d,
+                                        fp->defs[i].rank, F2PY_INTENT_IN,
+                                        v)) == NULL)
+                return -1;
+        }
+        if (fp->defs[i].data !=
+            NULL) { /* copy Python object to Fortran array */
+            npy_intp s = PyArray_MultiplyList(fp->defs[i].dims.d,
+                                              PyArray_NDIM(arr));
+            if (s == -1)
+                s = PyArray_MultiplyList(PyArray_DIMS(arr), PyArray_NDIM(arr));
+            if (s < 0 || (memcpy(fp->defs[i].data, PyArray_DATA(arr),
+                                 s * PyArray_ITEMSIZE(arr))) == NULL) {
+                if ((PyObject *)arr != v) {
+                    Py_DECREF(arr);
+                }
+                return -1;
+            }
+            if ((PyObject *)arr != v) {
+                Py_DECREF(arr);
+            }
+        }
+        else
+            return (fp->defs[i].func == NULL ? -1 : 0);
+        return 0; /* successful */
+    }
+    if (fp->dict == NULL) {
+        fp->dict = PyDict_New();
+        if (fp->dict == NULL)
+            return -1;
+    }
+    if (v == NULL) {
+        int rv = PyDict_DelItemString(fp->dict, name);
+        if (rv < 0)
+            PyErr_SetString(PyExc_AttributeError,
+                            "delete non-existing fortran attribute");
+        return rv;
+    }
+    else
+        return PyDict_SetItemString(fp->dict, name, v);
+}
+
+static PyObject *
+fortran_call(PyFortranObject *fp, PyObject *arg, PyObject *kw)
+{
+    int i = 0;
+    /*  printf("fortran call
+        name=%s,func=%p,data=%p,%p\n",fp->defs[i].name,
+        fp->defs[i].func,fp->defs[i].data,&fp->defs[i].data); */
+    if (fp->defs[i].rank == -1) { /* is Fortran routine */
+        if (fp->defs[i].func == NULL) {
+            PyErr_Format(PyExc_RuntimeError, "no function to call");
+            return NULL;
+        }
+        else if (fp->defs[i].data == NULL)
+            /* dummy routine */
+            return (*((fortranfunc)(fp->defs[i].func)))((PyObject *)fp, arg,
+                                                        kw, NULL);
+        else
+            return (*((fortranfunc)(fp->defs[i].func)))(
+                    (PyObject *)fp, arg, kw, (void *)fp->defs[i].data);
+    }
+    PyErr_Format(PyExc_TypeError, "this fortran object is not callable");
+    return NULL;
+}
+
+static PyObject *
+fortran_repr(PyFortranObject *fp)
+{
+    PyObject *name = NULL, *repr = NULL;
+    name = PyObject_GetAttrString((PyObject *)fp, "__name__");
+    PyErr_Clear();
+    if (name != NULL && PyUnicode_Check(name)) {
+        repr = PyUnicode_FromFormat("<fortran %U>", name);
+    }
+    else {
+        repr = PyUnicode_FromString("<fortran object>");
+    }
+    Py_XDECREF(name);
+    return repr;
+}
+
+PyTypeObject PyFortran_Type = {
+        PyVarObject_HEAD_INIT(NULL, 0).tp_name = "fortran",
+        .tp_basicsize = sizeof(PyFortranObject),
+        .tp_dealloc = (destructor)fortran_dealloc,
+        .tp_getattr = (getattrfunc)fortran_getattr,
+        .tp_setattr = (setattrfunc)fortran_setattr,
+        .tp_repr = (reprfunc)fortran_repr,
+        .tp_call = (ternaryfunc)fortran_call,
+};
+
+/************************* f2py_report_atexit *******************************/
+
+#ifdef F2PY_REPORT_ATEXIT
+static int passed_time = 0;
+static int passed_counter = 0;
+static int passed_call_time = 0;
+static struct timeb start_time;
+static struct timeb stop_time;
+static struct timeb start_call_time;
+static struct timeb stop_call_time;
+static int cb_passed_time = 0;
+static int cb_passed_counter = 0;
+static int cb_passed_call_time = 0;
+static struct timeb cb_start_time;
+static struct timeb cb_stop_time;
+static struct timeb cb_start_call_time;
+static struct timeb cb_stop_call_time;
+
+extern void
+f2py_start_clock(void)
+{
+    ftime(&start_time);
+}
+extern void
+f2py_start_call_clock(void)
+{
+    f2py_stop_clock();
+    ftime(&start_call_time);
+}
+extern void
+f2py_stop_clock(void)
+{
+    ftime(&stop_time);
+    passed_time += 1000 * (stop_time.time - start_time.time);
+    passed_time += stop_time.millitm - start_time.millitm;
+}
+extern void
+f2py_stop_call_clock(void)
+{
+    ftime(&stop_call_time);
+    passed_call_time += 1000 * (stop_call_time.time - start_call_time.time);
+    passed_call_time += stop_call_time.millitm - start_call_time.millitm;
+    passed_counter += 1;
+    f2py_start_clock();
+}
+
+extern void
+f2py_cb_start_clock(void)
+{
+    ftime(&cb_start_time);
+}
+extern void
+f2py_cb_start_call_clock(void)
+{
+    f2py_cb_stop_clock();
+    ftime(&cb_start_call_time);
+}
+extern void
+f2py_cb_stop_clock(void)
+{
+    ftime(&cb_stop_time);
+    cb_passed_time += 1000 * (cb_stop_time.time - cb_start_time.time);
+    cb_passed_time += cb_stop_time.millitm - cb_start_time.millitm;
+}
+extern void
+f2py_cb_stop_call_clock(void)
+{
+    ftime(&cb_stop_call_time);
+    cb_passed_call_time +=
+            1000 * (cb_stop_call_time.time - cb_start_call_time.time);
+    cb_passed_call_time +=
+            cb_stop_call_time.millitm - cb_start_call_time.millitm;
+    cb_passed_counter += 1;
+    f2py_cb_start_clock();
+}
+
+static int f2py_report_on_exit_been_here = 0;
+extern void
+f2py_report_on_exit(int exit_flag, void *name)
+{
+    if (f2py_report_on_exit_been_here) {
+        fprintf(stderr, "             %s\n", (char *)name);
+        return;
+    }
+    f2py_report_on_exit_been_here = 1;
+    fprintf(stderr, "                      /-----------------------\\\n");
+    fprintf(stderr, "                     < F2PY performance report >\n");
+    fprintf(stderr, "                      \\-----------------------/\n");
+    fprintf(stderr, "Overall time spent in ...\n");
+    fprintf(stderr, "(a) wrapped (Fortran/C) functions           : %8d msec\n",
+            passed_call_time);
+    fprintf(stderr, "(b) f2py interface,           %6d calls  : %8d msec\n",
+            passed_counter, passed_time);
+    fprintf(stderr, "(c) call-back (Python) functions            : %8d msec\n",
+            cb_passed_call_time);
+    fprintf(stderr, "(d) f2py call-back interface, %6d calls  : %8d msec\n",
+            cb_passed_counter, cb_passed_time);
+
+    fprintf(stderr,
+            "(e) wrapped (Fortran/C) functions (actual) : %8d msec\n\n",
+            passed_call_time - cb_passed_call_time - cb_passed_time);
+    fprintf(stderr,
+            "Use -DF2PY_REPORT_ATEXIT_DISABLE to disable this message.\n");
+    fprintf(stderr, "Exit status: %d\n", exit_flag);
+    fprintf(stderr, "Modules    : %s\n", (char *)name);
+}
+#endif
+
+/********************** report on array copy ****************************/
+
+#ifdef F2PY_REPORT_ON_ARRAY_COPY
+static void
+f2py_report_on_array_copy(PyArrayObject *arr)
+{
+    const npy_intp arr_size = PyArray_Size((PyObject *)arr);
+    if (arr_size > F2PY_REPORT_ON_ARRAY_COPY) {
+        fprintf(stderr,
+                "copied an array: size=%ld, elsize=%" NPY_INTP_FMT "\n",
+                arr_size, (npy_intp)PyArray_ITEMSIZE(arr));
+    }
+}
+static void
+f2py_report_on_array_copy_fromany(void)
+{
+    fprintf(stderr, "created an array from object\n");
+}
+
+#define F2PY_REPORT_ON_ARRAY_COPY_FROMARR \
+    f2py_report_on_array_copy((PyArrayObject *)arr)
+#define F2PY_REPORT_ON_ARRAY_COPY_FROMANY f2py_report_on_array_copy_fromany()
+#else
+#define F2PY_REPORT_ON_ARRAY_COPY_FROMARR
+#define F2PY_REPORT_ON_ARRAY_COPY_FROMANY
+#endif
+
+/************************* array_from_obj *******************************/
+
+/*
+ * File: array_from_pyobj.c
+ *
+ * Description:
+ * ------------
+ * Provides array_from_pyobj function that returns a contiguous array
+ * object with the given dimensions and required storage order, either
+ * in row-major (C) or column-major (Fortran) order. The function
+ * array_from_pyobj is very flexible about its Python object argument
+ * that can be any number, list, tuple, or array.
+ *
+ * array_from_pyobj is used in f2py generated Python extension
+ * modules.
+ *
+ * Author: Pearu Peterson <[email protected]>
+ * Created: 13-16 January 2002
+ * $Id: fortranobject.c,v 1.52 2005/07/11 07:44:20 pearu Exp $
+ */
+
+static int check_and_fix_dimensions(const PyArrayObject* arr,
+                                    const int rank,
+                                    npy_intp *dims,
+                                    const char *errmess);
+
+static int
+find_first_negative_dimension(const int rank, const npy_intp *dims)
+{
+    int i;
+    for (i = 0; i < rank; ++i) {
+        if (dims[i] < 0) {
+            return i;
+        }
+    }
+    return -1;
+}
+
+#ifdef DEBUG_COPY_ND_ARRAY
+void
+dump_dims(int rank, npy_intp const *dims)
+{
+    int i;
+    printf("[");
+    for (i = 0; i < rank; ++i) {
+        printf("%3" NPY_INTP_FMT, dims[i]);
+    }
+    printf("]\n");
+}
+void
+dump_attrs(const PyArrayObject *obj)
+{
+    const PyArrayObject_fields *arr = (const PyArrayObject_fields *)obj;
+    int rank = PyArray_NDIM(arr);
+    npy_intp size = PyArray_Size((PyObject *)arr);
+    printf("\trank = %d, flags = %d, size = %" NPY_INTP_FMT "\n", rank,
+           arr->flags, size);
+    printf("\tstrides = ");
+    dump_dims(rank, arr->strides);
+    printf("\tdimensions = ");
+    dump_dims(rank, arr->dimensions);
+}
+#endif
+
+#define SWAPTYPE(a, b, t) \
+    {                     \
+        t c;              \
+        c = (a);          \
+        (a) = (b);        \
+        (b) = c;          \
+    }
+
+static int
+swap_arrays(PyArrayObject *obj1, PyArrayObject *obj2)
+{
+    PyArrayObject_fields *arr1 = (PyArrayObject_fields *)obj1,
+                         *arr2 = (PyArrayObject_fields *)obj2;
+    SWAPTYPE(arr1->data, arr2->data, char *);
+    SWAPTYPE(arr1->nd, arr2->nd, int);
+    SWAPTYPE(arr1->dimensions, arr2->dimensions, npy_intp *);
+    SWAPTYPE(arr1->strides, arr2->strides, npy_intp *);
+    SWAPTYPE(arr1->base, arr2->base, PyObject *);
+    SWAPTYPE(arr1->descr, arr2->descr, PyArray_Descr *);
+    SWAPTYPE(arr1->flags, arr2->flags, int);
+    /* SWAPTYPE(arr1->weakreflist,arr2->weakreflist,PyObject*); */
+    return 0;
+}
+
+#define ARRAY_ISCOMPATIBLE(arr,type_num)                                \
+    ((PyArray_ISINTEGER(arr) && PyTypeNum_ISINTEGER(type_num)) ||     \
+     (PyArray_ISFLOAT(arr) && PyTypeNum_ISFLOAT(type_num)) ||         \
+     (PyArray_ISCOMPLEX(arr) && PyTypeNum_ISCOMPLEX(type_num)) ||     \
+     (PyArray_ISBOOL(arr) && PyTypeNum_ISBOOL(type_num)) ||           \
+     (PyArray_ISSTRING(arr) && PyTypeNum_ISSTRING(type_num)))
+
+static int
+get_elsize(PyObject *obj) {
+  /*
+    get_elsize determines array itemsize from a Python object.  Returns
+    elsize if successful, -1 otherwise.
+
+    Supported types of the input are: numpy.ndarray, bytes, str, tuple,
+    list.
+  */
+
+  if (PyArray_Check(obj)) {
+    return PyArray_DESCR((PyArrayObject *)obj)->elsize;
+  } else if (PyBytes_Check(obj)) {
+    return PyBytes_GET_SIZE(obj);
+  } else if (PyUnicode_Check(obj)) {
+    return PyUnicode_GET_LENGTH(obj);
+  } else if (PySequence_Check(obj)) {
+    PyObject* fast = PySequence_Fast(obj, "f2py:fortranobject.c:get_elsize");
+    if (fast != NULL) {
+      Py_ssize_t i, n = PySequence_Fast_GET_SIZE(fast);
+      int sz, elsize = 0;
+      for (i=0; i<n; i++) {
+        sz = get_elsize(PySequence_Fast_GET_ITEM(fast, i) /* borrowed */);
+        if (sz > elsize) {
+          elsize = sz;
+        }
+      }
+      Py_DECREF(fast);
+      return elsize;
+    }
+  }
+  return -1;
+}
+
+extern PyArrayObject *
+ndarray_from_pyobj(const int type_num,
+                   const int elsize_,
+                   npy_intp *dims,
+                   const int rank,
+                   const int intent,
+                   PyObject *obj,
+                   const char *errmess) {
+    /*
+     * Return an array with given element type and shape from a Python
+     * object while taking into account the usage intent of the array.
+     *
+     * - element type is defined by type_num and elsize
+     * - shape is defined by dims and rank
+     *
+     * ndarray_from_pyobj is used to convert Python object arguments
+     * to numpy ndarrays with given type and shape that data is passed
+     * to interfaced Fortran or C functions.
+     *
+     * errmess (if not NULL), contains a prefix of an error message
+     * for an exception to be triggered within this function.
+     *
+     * Negative elsize value means that elsize is to be determined
+     * from the Python object in runtime.
+     *
+     * Note on strings
+     * ---------------
+     *
+     * String type (type_num == NPY_STRING) does not have fixed
+     * element size and, by default, the type object sets it to
+     * 0. Therefore, for string types, one has to use elsize
+     * argument. For other types, elsize value is ignored.
+     *
+     * NumPy defines the type of a fixed-width string as
+     * dtype('S<width>'). In addition, there is also dtype('c'), that
+     * appears as dtype('S1') (these have the same type_num value),
+     * but is actually different (.char attribute is either 'S' or
+     * 'c', respecitely).
+     *
+     * In Fortran, character arrays and strings are different
+     * concepts.  The relation between Fortran types, NumPy dtypes,
+     * and type_num-elsize pairs, is defined as follows:
+     *
+     * character*5 foo     | dtype('S5')  | elsize=5, shape=()
+     * character(5) foo    | dtype('S1')  | elsize=1, shape=(5)
+     * character*5 foo(n)  | dtype('S5')  | elsize=5, shape=(n,)
+     * character(5) foo(n) | dtype('S1')  | elsize=1, shape=(5, n)
+     * character*(*) foo   | dtype('S')   | elsize=-1, shape=()
+     *
+     * Note about reference counting
+     * -----------------------------
+     *
+     * If the caller returns the array to Python, it must be done with
+     * Py_BuildValue("N",arr).  Otherwise, if obj!=arr then the caller
+     * must call Py_DECREF(arr).
+     *
+     * Note on intent(cache,out,..)
+     * ----------------------------
+     * Don't expect correct data when returning intent(cache) array.
+     *
+     */
+    char mess[F2PY_MESSAGE_BUFFER_SIZE];
+    PyArrayObject *arr = NULL;
+    int elsize = (elsize_ < 0 ? get_elsize(obj) : elsize_);
+    if (elsize < 0) {
+      if (errmess != NULL) {
+        strcpy(mess, errmess);
+      }
+      sprintf(mess + strlen(mess),
+              " -- failed to determine element size from %s",
+              Py_TYPE(obj)->tp_name);
+      PyErr_SetString(PyExc_SystemError, mess);
+      return NULL;
+    }
+    PyArray_Descr * descr = get_descr_from_type_and_elsize(type_num, elsize);  // new reference
+    if (descr == NULL) {
+      return NULL;
+    }
+    elsize = descr->elsize;
+    if ((intent & F2PY_INTENT_HIDE)
+        || ((intent & F2PY_INTENT_CACHE) && (obj == Py_None))
+        || ((intent & F2PY_OPTIONAL) && (obj == Py_None))
+        ) {
+        /* intent(cache), optional, intent(hide) */
+        int ineg = find_first_negative_dimension(rank, dims);
+        if (ineg >= 0) {
+            int i;
+            strcpy(mess, "failed to create intent(cache|hide)|optional array"
+                   "-- must have defined dimensions but got (");
+            for(i = 0; i < rank; ++i)
+                sprintf(mess + strlen(mess), "%" NPY_INTP_FMT ",", dims[i]);
+            strcat(mess, ")");
+            PyErr_SetString(PyExc_ValueError, mess);
+            Py_DECREF(descr);
+            return NULL;
+        }
+        arr = (PyArrayObject *)                                      \
+          PyArray_NewFromDescr(&PyArray_Type, descr, rank, dims,
+                               NULL, NULL, !(intent & F2PY_INTENT_C), NULL);
+        if (arr == NULL) {
+          Py_DECREF(descr);
+          return NULL;
+        }
+        if (PyArray_ITEMSIZE(arr) != elsize) {
+          strcpy(mess, "failed to create intent(cache|hide)|optional array");
+          sprintf(mess+strlen(mess)," -- expected elsize=%d got %" NPY_INTP_FMT, elsize, (npy_intp)PyArray_ITEMSIZE(arr));
+          PyErr_SetString(PyExc_ValueError,mess);
+          Py_DECREF(arr);
+          return NULL;
+        }
+        if (!(intent & F2PY_INTENT_CACHE)) {
+          PyArray_FILLWBYTE(arr, 0);
+        }
+        return arr;
+    }
+
+    if (PyArray_Check(obj)) {
+        arr = (PyArrayObject *)obj;
+        if (intent & F2PY_INTENT_CACHE) {
+            /* intent(cache) */
+            if (PyArray_ISONESEGMENT(arr)
+                && PyArray_ITEMSIZE(arr) >= elsize) {
+                if (check_and_fix_dimensions(arr, rank, dims, errmess)) {
+                  Py_DECREF(descr);
+                  return NULL;
+                }
+                if (intent & F2PY_INTENT_OUT)
+                  Py_INCREF(arr);
+                Py_DECREF(descr);
+                return arr;
+            }
+            strcpy(mess, "failed to initialize intent(cache) array");
+            if (!PyArray_ISONESEGMENT(arr))
+                strcat(mess, " -- input must be in one segment");
+            if (PyArray_ITEMSIZE(arr) < elsize)
+                sprintf(mess + strlen(mess),
+                        " -- expected at least elsize=%d but got "
+                        "%" NPY_INTP_FMT,
+                        elsize, (npy_intp)PyArray_ITEMSIZE(arr));
+            PyErr_SetString(PyExc_ValueError, mess);
+            Py_DECREF(descr);
+            return NULL;
+        }
+
+        /* here we have always intent(in) or intent(inout) or intent(inplace)
+         */
+
+        if (check_and_fix_dimensions(arr, rank, dims, errmess)) {
+          Py_DECREF(descr);
+          return NULL;
+        }
+        /*
+        printf("intent alignment=%d\n", F2PY_GET_ALIGNMENT(intent));
+        printf("alignment check=%d\n", F2PY_CHECK_ALIGNMENT(arr, intent));
+        int i;
+        for (i=1;i<=16;i++)
+          printf("i=%d isaligned=%d\n", i, ARRAY_ISALIGNED(arr, i));
+        */
+        if ((! (intent & F2PY_INTENT_COPY)) &&
+            PyArray_ITEMSIZE(arr) == elsize &&
+            ARRAY_ISCOMPATIBLE(arr,type_num) &&
+            F2PY_CHECK_ALIGNMENT(arr, intent)) {
+            if ((intent & F2PY_INTENT_INOUT || intent & F2PY_INTENT_INPLACE)
+              ? ((intent & F2PY_INTENT_C) ? PyArray_ISCARRAY(arr) : PyArray_ISFARRAY(arr))
+              : ((intent & F2PY_INTENT_C) ? PyArray_ISCARRAY_RO(arr) : PyArray_ISFARRAY_RO(arr))) {
+                if ((intent & F2PY_INTENT_OUT)) {
+                    Py_INCREF(arr);
+                }
+                /* Returning input array */
+                Py_DECREF(descr);
+                return arr;
+            }
+        }
+        if (intent & F2PY_INTENT_INOUT) {
+            strcpy(mess, "failed to initialize intent(inout) array");
+            /* Must use PyArray_IS*ARRAY because intent(inout) requires
+             * writable input */
+            if ((intent & F2PY_INTENT_C) && !PyArray_ISCARRAY(arr))
+                strcat(mess, " -- input not contiguous");
+            if (!(intent & F2PY_INTENT_C) && !PyArray_ISFARRAY(arr))
+                strcat(mess, " -- input not fortran contiguous");
+            if (PyArray_ITEMSIZE(arr) != elsize)
+                sprintf(mess + strlen(mess),
+                        " -- expected elsize=%d but got %" NPY_INTP_FMT,
+                        elsize,
+                        (npy_intp)PyArray_ITEMSIZE(arr)
+                        );
+            if (!(ARRAY_ISCOMPATIBLE(arr, type_num))) {
+                sprintf(mess + strlen(mess),
+                        " -- input '%c' not compatible to '%c'",
+                        PyArray_DESCR(arr)->type, descr->type);
+            }
+            if (!(F2PY_CHECK_ALIGNMENT(arr, intent)))
+                sprintf(mess + strlen(mess), " -- input not %d-aligned",
+                        F2PY_GET_ALIGNMENT(intent));
+            PyErr_SetString(PyExc_ValueError, mess);
+            Py_DECREF(descr);
+            return NULL;
+        }
+
+        /* here we have always intent(in) or intent(inplace) */
+
+        {
+          PyArrayObject * retarr = (PyArrayObject *)                    \
+            PyArray_NewFromDescr(&PyArray_Type, descr, PyArray_NDIM(arr), PyArray_DIMS(arr),
+                                 NULL, NULL, !(intent & F2PY_INTENT_C), NULL);
+          if (retarr==NULL) {
+            Py_DECREF(descr);
+            return NULL;
+          }
+          F2PY_REPORT_ON_ARRAY_COPY_FROMARR;
+          if (PyArray_CopyInto(retarr, arr)) {
+            Py_DECREF(retarr);
+            return NULL;
+          }
+          if (intent & F2PY_INTENT_INPLACE) {
+            if (swap_arrays(arr,retarr)) {
+              Py_DECREF(retarr);
+              return NULL; /* XXX: set exception */
+            }
+            Py_XDECREF(retarr);
+            if (intent & F2PY_INTENT_OUT)
+              Py_INCREF(arr);
+          } else {
+            arr = retarr;
+          }
+        }
+        return arr;
+    }
+
+    if ((intent & F2PY_INTENT_INOUT) || (intent & F2PY_INTENT_INPLACE) ||
+        (intent & F2PY_INTENT_CACHE)) {
+        PyErr_Format(PyExc_TypeError,
+                     "failed to initialize intent(inout|inplace|cache) "
+                     "array, input '%s' object is not an array",
+                     Py_TYPE(obj)->tp_name);
+        Py_DECREF(descr);
+        return NULL;
+    }
+
+    {
+        F2PY_REPORT_ON_ARRAY_COPY_FROMANY;
+        arr = (PyArrayObject *)PyArray_FromAny(
+                obj, descr, 0, 0,
+                ((intent & F2PY_INTENT_C) ? NPY_ARRAY_CARRAY
+                                          : NPY_ARRAY_FARRAY) |
+                        NPY_ARRAY_FORCECAST,
+                NULL);
+        // Warning: in the case of NPY_STRING, PyArray_FromAny may
+        // reset descr->elsize, e.g. dtype('S0') becomes dtype('S1').
+        if (arr == NULL) {
+          Py_DECREF(descr);
+          return NULL;
+        }
+        if (type_num != NPY_STRING && PyArray_ITEMSIZE(arr) != elsize) {
+          // This is internal sanity tests: elsize has been set to
+          // descr->elsize in the beginning of this function.
+          strcpy(mess, "failed to initialize intent(in) array");
+          sprintf(mess + strlen(mess),
+                  " -- expected elsize=%d got %" NPY_INTP_FMT, elsize,
+                  (npy_intp)PyArray_ITEMSIZE(arr));
+          PyErr_SetString(PyExc_ValueError, mess);
+          Py_DECREF(arr);
+          return NULL;
+        }
+        if (check_and_fix_dimensions(arr, rank, dims, errmess)) {
+          Py_DECREF(arr);
+          return NULL;
+        }
+        return arr;
+    }
+}
+
+extern PyArrayObject *
+array_from_pyobj(const int type_num,
+                                npy_intp *dims,
+                                const int rank,
+                                const int intent,
+                                PyObject *obj) {
+  /*
+    Same as ndarray_from_pyobj but with elsize determined from type,
+    if possible. Provided for backward compatibility.
+   */
+  PyArray_Descr* descr = PyArray_DescrFromType(type_num);
+  int elsize = descr->elsize;
+  Py_DECREF(descr);
+  return ndarray_from_pyobj(type_num, elsize, dims, rank, intent, obj, NULL);
+}
+
+/*****************************************/
+/* Helper functions for array_from_pyobj */
+/*****************************************/
+
+static int
+check_and_fix_dimensions(const PyArrayObject* arr, const int rank,
+                         npy_intp *dims, const char *errmess)
+{
+    /*
+     * This function fills in blanks (that are -1's) in dims list using
+     * the dimensions from arr. It also checks that non-blank dims will
+     * match with the corresponding values in arr dimensions.
+     *
+     * Returns 0 if the function is successful.
+     *
+     * If an error condition is detected, an exception is set and 1 is
+     * returned.
+     */
+    char mess[F2PY_MESSAGE_BUFFER_SIZE];
+    const npy_intp arr_size =
+            (PyArray_NDIM(arr)) ? PyArray_Size((PyObject *)arr) : 1;
+#ifdef DEBUG_COPY_ND_ARRAY
+    dump_attrs(arr);
+    printf("check_and_fix_dimensions:init: dims=");
+    dump_dims(rank, dims);
+#endif
+    if (rank > PyArray_NDIM(arr)) { /* [1,2] -> [[1],[2]]; 1 -> [[1]]  */
+        npy_intp new_size = 1;
+        int free_axe = -1;
+        int i;
+        npy_intp d;
+        /* Fill dims where -1 or 0; check dimensions; calc new_size; */
+        for (i = 0; i < PyArray_NDIM(arr); ++i) {
+            d = PyArray_DIM(arr, i);
+            if (dims[i] >= 0) {
+                if (d > 1 && dims[i] != d) {
+                    PyErr_Format(
+                            PyExc_ValueError,
+                            "%d-th dimension must be fixed to %" NPY_INTP_FMT
+                            " but got %" NPY_INTP_FMT "\n",
+                            i, dims[i], d);
+                    return 1;
+                }
+                if (!dims[i])
+                    dims[i] = 1;
+            }
+            else {
+                dims[i] = d ? d : 1;
+            }
+            new_size *= dims[i];
+        }
+        for (i = PyArray_NDIM(arr); i < rank; ++i)
+            if (dims[i] > 1) {
+                PyErr_Format(PyExc_ValueError,
+                             "%d-th dimension must be %" NPY_INTP_FMT
+                             " but got 0 (not defined).\n",
+                             i, dims[i]);
+                return 1;
+            }
+            else if (free_axe < 0)
+                free_axe = i;
+            else
+                dims[i] = 1;
+        if (free_axe >= 0) {
+            dims[free_axe] = arr_size / new_size;
+            new_size *= dims[free_axe];
+        }
+        if (new_size != arr_size) {
+            PyErr_Format(PyExc_ValueError,
+                         "unexpected array size: new_size=%" NPY_INTP_FMT
+                         ", got array with arr_size=%" NPY_INTP_FMT
+                         " (maybe too many free indices)\n",
+                         new_size, arr_size);
+            return 1;
+        }
+    }
+    else if (rank == PyArray_NDIM(arr)) {
+        npy_intp new_size = 1;
+        int i;
+        npy_intp d;
+        for (i = 0; i < rank; ++i) {
+            d = PyArray_DIM(arr, i);
+            if (dims[i] >= 0) {
+                if (d > 1 && d != dims[i]) {
+                    if (errmess != NULL) {
+                        strcpy(mess, errmess);
+                    }
+                    sprintf(mess + strlen(mess),
+                            " -- %d-th dimension must be fixed to %"
+                            NPY_INTP_FMT " but got %" NPY_INTP_FMT,
+                            i, dims[i], d);
+                    PyErr_SetString(PyExc_ValueError, mess);
+                    return 1;
+                }
+                if (!dims[i])
+                    dims[i] = 1;
+            }
+            else
+                dims[i] = d;
+            new_size *= dims[i];
+        }
+        if (new_size != arr_size) {
+            PyErr_Format(PyExc_ValueError,
+                         "unexpected array size: new_size=%" NPY_INTP_FMT
+                         ", got array with arr_size=%" NPY_INTP_FMT "\n",
+                         new_size, arr_size);
+            return 1;
+        }
+    }
+    else { /* [[1,2]] -> [[1],[2]] */
+        int i, j;
+        npy_intp d;
+        int effrank;
+        npy_intp size;
+        for (i = 0, effrank = 0; i < PyArray_NDIM(arr); ++i)
+            if (PyArray_DIM(arr, i) > 1)
+                ++effrank;
+        if (dims[rank - 1] >= 0)
+            if (effrank > rank) {
+                PyErr_Format(PyExc_ValueError,
+                             "too many axes: %d (effrank=%d), "
+                             "expected rank=%d\n",
+                             PyArray_NDIM(arr), effrank, rank);
+                return 1;
+            }
+
+        for (i = 0, j = 0; i < rank; ++i) {
+            while (j < PyArray_NDIM(arr) && PyArray_DIM(arr, j) < 2) ++j;
+            if (j >= PyArray_NDIM(arr))
+                d = 1;
+            else
+                d = PyArray_DIM(arr, j++);
+            if (dims[i] >= 0) {
+                if (d > 1 && d != dims[i]) {
+                    if (errmess != NULL) {
+                        strcpy(mess, errmess);
+                    }
+                    sprintf(mess + strlen(mess),
+                            " -- %d-th dimension must be fixed to %"
+                            NPY_INTP_FMT " but got %" NPY_INTP_FMT
+                            " (real index=%d)\n",
+                            i, dims[i], d, j-1);
+                    PyErr_SetString(PyExc_ValueError, mess);
+                    return 1;
+                }
+                if (!dims[i])
+                    dims[i] = 1;
+            }
+            else
+                dims[i] = d;
+        }
+
+        for (i = rank; i < PyArray_NDIM(arr);
+             ++i) { /* [[1,2],[3,4]] -> [1,2,3,4] */
+            while (j < PyArray_NDIM(arr) && PyArray_DIM(arr, j) < 2) ++j;
+            if (j >= PyArray_NDIM(arr))
+                d = 1;
+            else
+                d = PyArray_DIM(arr, j++);
+            dims[rank - 1] *= d;
+        }
+        for (i = 0, size = 1; i < rank; ++i) size *= dims[i];
+        if (size != arr_size) {
+            char msg[200];
+            int len;
+            snprintf(msg, sizeof(msg),
+                     "unexpected array size: size=%" NPY_INTP_FMT
+                     ", arr_size=%" NPY_INTP_FMT
+                     ", rank=%d, effrank=%d, arr.nd=%d, dims=[",
+                     size, arr_size, rank, effrank, PyArray_NDIM(arr));
+            for (i = 0; i < rank; ++i) {
+                len = strlen(msg);
+                snprintf(msg + len, sizeof(msg) - len, " %" NPY_INTP_FMT,
+                         dims[i]);
+            }
+            len = strlen(msg);
+            snprintf(msg + len, sizeof(msg) - len, " ], arr.dims=[");
+            for (i = 0; i < PyArray_NDIM(arr); ++i) {
+                len = strlen(msg);
+                snprintf(msg + len, sizeof(msg) - len, " %" NPY_INTP_FMT,
+                         PyArray_DIM(arr, i));
+            }
+            len = strlen(msg);
+            snprintf(msg + len, sizeof(msg) - len, " ]\n");
+            PyErr_SetString(PyExc_ValueError, msg);
+            return 1;
+        }
+    }
+#ifdef DEBUG_COPY_ND_ARRAY
+    printf("check_and_fix_dimensions:end: dims=");
+    dump_dims(rank, dims);
+#endif
+    return 0;
+}
+
+/* End of file: array_from_pyobj.c */
+
+/************************* copy_ND_array *******************************/
+
+extern int
+copy_ND_array(const PyArrayObject *arr, PyArrayObject *out)
+{
+    F2PY_REPORT_ON_ARRAY_COPY_FROMARR;
+    return PyArray_CopyInto(out, (PyArrayObject *)arr);
+}
+
+/********************* Various utility functions ***********************/
+
+extern int
+f2py_describe(PyObject *obj, char *buf) {
+  /*
+    Write the description of a Python object to buf. The caller must
+    provide buffer with size sufficient to write the description.
+
+    Return 1 on success.
+  */
+  char localbuf[F2PY_MESSAGE_BUFFER_SIZE];
+  if (PyBytes_Check(obj)) {
+    sprintf(localbuf, "%d-%s", (npy_int)PyBytes_GET_SIZE(obj), Py_TYPE(obj)->tp_name);
+  } else if (PyUnicode_Check(obj)) {
+    sprintf(localbuf, "%d-%s", (npy_int)PyUnicode_GET_LENGTH(obj), Py_TYPE(obj)->tp_name);
+  } else if (PyArray_CheckScalar(obj)) {
+    PyArrayObject* arr = (PyArrayObject*)obj;
+    sprintf(localbuf, "%c%" NPY_INTP_FMT "-%s-scalar", PyArray_DESCR(arr)->kind, PyArray_ITEMSIZE(arr), Py_TYPE(obj)->tp_name);
+  } else if (PyArray_Check(obj)) {
+    int i;
+    PyArrayObject* arr = (PyArrayObject*)obj;
+    strcpy(localbuf, "(");
+    for (i=0; i<PyArray_NDIM(arr); i++) {
+      if (i) {
+        strcat(localbuf, " ");
+      }
+      sprintf(localbuf + strlen(localbuf), "%" NPY_INTP_FMT ",", PyArray_DIM(arr, i));
+    }
+    sprintf(localbuf + strlen(localbuf), ")-%c%" NPY_INTP_FMT "-%s", PyArray_DESCR(arr)->kind, PyArray_ITEMSIZE(arr), Py_TYPE(obj)->tp_name);
+  } else if (PySequence_Check(obj)) {
+    sprintf(localbuf, "%d-%s", (npy_int)PySequence_Length(obj), Py_TYPE(obj)->tp_name);
+  } else {
+    sprintf(localbuf, "%s instance", Py_TYPE(obj)->tp_name);
+  }
+  // TODO: detect the size of buf and make sure that size(buf) >= size(localbuf).
+  strcpy(buf, localbuf);
+  return 1;
+}
+
+extern npy_intp
+f2py_size_impl(PyArrayObject* var, ...)
+{
+  npy_intp sz = 0;
+  npy_intp dim;
+  npy_intp rank;
+  va_list argp;
+  va_start(argp, var);
+  dim = va_arg(argp, npy_int);
+  if (dim==-1)
+    {
+      sz = PyArray_SIZE(var);
+    }
+  else
+    {
+      rank = PyArray_NDIM(var);
+      if (dim>=1 && dim<=rank)
+        sz = PyArray_DIM(var, dim-1);
+      else
+        fprintf(stderr, "f2py_size: 2nd argument value=%" NPY_INTP_FMT
+                " fails to satisfy 1<=value<=%" NPY_INTP_FMT
+                ". Result will be 0.\n", dim, rank);
+    }
+  va_end(argp);
+  return sz;
+}
+
+/*********************************************/
+/* Compatibility functions for Python >= 3.0 */
+/*********************************************/
+
+PyObject *
+F2PyCapsule_FromVoidPtr(void *ptr, void (*dtor)(PyObject *))
+{
+    PyObject *ret = PyCapsule_New(ptr, NULL, dtor);
+    if (ret == NULL) {
+        PyErr_Clear();
+    }
+    return ret;
+}
+
+void *
+F2PyCapsule_AsVoidPtr(PyObject *obj)
+{
+    void *ret = PyCapsule_GetPointer(obj, NULL);
+    if (ret == NULL) {
+        PyErr_Clear();
+    }
+    return ret;
+}
+
+int
+F2PyCapsule_Check(PyObject *ptr)
+{
+    return PyCapsule_CheckExact(ptr);
+}
+
+#ifdef __cplusplus
+}
+#endif
+/************************* EOF fortranobject.c *******************************/

.venv/lib/python3.11/site-packages/numpy/f2py/src/fortranobject.h

@@ -0,0 +1,173 @@
+#ifndef Py_FORTRANOBJECT_H
+#define Py_FORTRANOBJECT_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <Python.h>
+
+#ifndef NPY_NO_DEPRECATED_API
+#define NPY_NO_DEPRECATED_API NPY_API_VERSION
+#endif
+#ifdef FORTRANOBJECT_C
+#define NO_IMPORT_ARRAY
+#endif
+#define PY_ARRAY_UNIQUE_SYMBOL _npy_f2py_ARRAY_API
+#include "numpy/arrayobject.h"
+#include "numpy/npy_3kcompat.h"
+
+#ifdef F2PY_REPORT_ATEXIT
+#include <sys/timeb.h>
+// clang-format off
+extern void f2py_start_clock(void);
+extern void f2py_stop_clock(void);
+extern void f2py_start_call_clock(void);
+extern void f2py_stop_call_clock(void);
+extern void f2py_cb_start_clock(void);
+extern void f2py_cb_stop_clock(void);
+extern void f2py_cb_start_call_clock(void);
+extern void f2py_cb_stop_call_clock(void);
+extern void f2py_report_on_exit(int, void *);
+// clang-format on
+#endif
+
+#ifdef DMALLOC
+#include "dmalloc.h"
+#endif
+
+/* Fortran object interface */
+
+/*
+123456789-123456789-123456789-123456789-123456789-123456789-123456789-12
+
+PyFortranObject represents various Fortran objects:
+Fortran (module) routines, COMMON blocks, module data.
+
+Author: Pearu Peterson <[email protected]>
+*/
+
+#define F2PY_MAX_DIMS 40
+#define F2PY_MESSAGE_BUFFER_SIZE 300  // Increase on "stack smashing detected"
+
+typedef void (*f2py_set_data_func)(char *, npy_intp *);
+typedef void (*f2py_void_func)(void);
+typedef void (*f2py_init_func)(int *, npy_intp *, f2py_set_data_func, int *);
+
+/*typedef void* (*f2py_c_func)(void*,...);*/
+
+typedef void *(*f2pycfunc)(void);
+
+typedef struct {
+    char *name; /* attribute (array||routine) name */
+    int rank;   /* array rank, 0 for scalar, max is F2PY_MAX_DIMS,
+                   || rank=-1 for Fortran routine */
+    struct {
+        npy_intp d[F2PY_MAX_DIMS];
+    } dims;              /* dimensions of the array, || not used */
+    int type;            /* PyArray_<type> || not used */
+    int elsize;                /* Element size || not used */
+    char *data;          /* pointer to array || Fortran routine */
+    f2py_init_func func; /* initialization function for
+                            allocatable arrays:
+                            func(&rank,dims,set_ptr_func,name,len(name))
+                            || C/API wrapper for Fortran routine */
+    char *doc;           /* documentation string; only recommended
+                            for routines. */
+} FortranDataDef;
+
+typedef struct {
+    PyObject_HEAD
+    int len;              /* Number of attributes */
+    FortranDataDef *defs; /* An array of FortranDataDef's */
+    PyObject *dict;       /* Fortran object attribute dictionary */
+} PyFortranObject;
+
+#define PyFortran_Check(op) (Py_TYPE(op) == &PyFortran_Type)
+#define PyFortran_Check1(op) (0 == strcmp(Py_TYPE(op)->tp_name, "fortran"))
+
+extern PyTypeObject PyFortran_Type;
+extern int
+F2PyDict_SetItemString(PyObject *dict, char *name, PyObject *obj);
+extern PyObject *
+PyFortranObject_New(FortranDataDef *defs, f2py_void_func init);
+extern PyObject *
+PyFortranObject_NewAsAttr(FortranDataDef *defs);
+
+PyObject *
+F2PyCapsule_FromVoidPtr(void *ptr, void (*dtor)(PyObject *));
+void *
+F2PyCapsule_AsVoidPtr(PyObject *obj);
+int
+F2PyCapsule_Check(PyObject *ptr);
+
+extern void *
+F2PySwapThreadLocalCallbackPtr(char *key, void *ptr);
+extern void *
+F2PyGetThreadLocalCallbackPtr(char *key);
+
+#define ISCONTIGUOUS(m) (PyArray_FLAGS(m) & NPY_ARRAY_C_CONTIGUOUS)
+#define F2PY_INTENT_IN 1
+#define F2PY_INTENT_INOUT 2
+#define F2PY_INTENT_OUT 4
+#define F2PY_INTENT_HIDE 8
+#define F2PY_INTENT_CACHE 16
+#define F2PY_INTENT_COPY 32
+#define F2PY_INTENT_C 64
+#define F2PY_OPTIONAL 128
+#define F2PY_INTENT_INPLACE 256
+#define F2PY_INTENT_ALIGNED4 512
+#define F2PY_INTENT_ALIGNED8 1024
+#define F2PY_INTENT_ALIGNED16 2048
+
+#define ARRAY_ISALIGNED(ARR, SIZE) ((size_t)(PyArray_DATA(ARR)) % (SIZE) == 0)
+#define F2PY_ALIGN4(intent) (intent & F2PY_INTENT_ALIGNED4)
+#define F2PY_ALIGN8(intent) (intent & F2PY_INTENT_ALIGNED8)
+#define F2PY_ALIGN16(intent) (intent & F2PY_INTENT_ALIGNED16)
+
+#define F2PY_GET_ALIGNMENT(intent) \
+    (F2PY_ALIGN4(intent)           \
+             ? 4                   \
+             : (F2PY_ALIGN8(intent) ? 8 : (F2PY_ALIGN16(intent) ? 16 : 1)))
+#define F2PY_CHECK_ALIGNMENT(arr, intent) \
+    ARRAY_ISALIGNED(arr, F2PY_GET_ALIGNMENT(intent))
+#define F2PY_ARRAY_IS_CHARACTER_COMPATIBLE(arr) ((PyArray_DESCR(arr)->type_num == NPY_STRING && PyArray_DESCR(arr)->elsize >= 1) \
+                                                 || PyArray_DESCR(arr)->type_num == NPY_UINT8)
+#define F2PY_IS_UNICODE_ARRAY(arr) (PyArray_DESCR(arr)->type_num == NPY_UNICODE)
+
+extern PyArrayObject *
+ndarray_from_pyobj(const int type_num, const int elsize_, npy_intp *dims,
+                   const int rank, const int intent, PyObject *obj,
+                   const char *errmess);
+
+extern PyArrayObject *
+array_from_pyobj(const int type_num, npy_intp *dims, const int rank,
+                 const int intent, PyObject *obj);
+extern int
+copy_ND_array(const PyArrayObject *in, PyArrayObject *out);
+
+#ifdef DEBUG_COPY_ND_ARRAY
+extern void
+dump_attrs(const PyArrayObject *arr);
+#endif
+
+  extern int f2py_describe(PyObject *obj, char *buf);
+
+  /* Utility CPP macros and functions that can be used in signature file
+     expressions. See signature-file.rst for documentation.
+  */
+
+#define f2py_itemsize(var) (PyArray_DESCR((capi_ ## var ## _as_array))->elsize)
+#define f2py_size(var, ...) f2py_size_impl((PyArrayObject *)(capi_ ## var ## _as_array), ## __VA_ARGS__, -1)
+#define f2py_rank(var) var ## _Rank
+#define f2py_shape(var,dim) var ## _Dims[dim]
+#define f2py_len(var) f2py_shape(var,0)
+#define f2py_fshape(var,dim) f2py_shape(var,rank(var)-dim-1)
+#define f2py_flen(var) f2py_fshape(var,0)
+#define f2py_slen(var) capi_ ## var ## _len
+
+  extern npy_intp f2py_size_impl(PyArrayObject* var, ...);
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_FORTRANOBJECT_H */

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/abstract_interface/foo.f90

@@ -0,0 +1,34 @@
+module ops_module
+
+  abstract interface
+    subroutine op(x, y, z)
+      integer, intent(in) :: x, y
+      integer, intent(out) :: z
+    end subroutine
+  end interface
+
+contains
+
+  subroutine foo(x, y, r1, r2)
+    integer, intent(in) :: x, y
+    integer, intent(out) :: r1, r2
+    procedure (op) add1, add2
+    procedure (op), pointer::p
+    p=>add1
+    call p(x, y, r1)
+    p=>add2
+    call p(x, y, r2)
+  end subroutine
+end module
+
+subroutine add1(x, y, z)
+  integer, intent(in) :: x, y
+  integer, intent(out) :: z
+  z = x + y
+end subroutine
+
+subroutine add2(x, y, z)
+  integer, intent(in) :: x, y
+  integer, intent(out) :: z
+  z = x + 2 * y
+end subroutine

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/abstract_interface/gh18403_mod.f90

@@ -0,0 +1,6 @@
+module test
+  abstract interface
+    subroutine foo()
+    end subroutine
+  end interface
+end module test

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/block_docstring/foo.f

@@ -0,0 +1,6 @@
+      SUBROUTINE FOO()
+      INTEGER BAR(2, 3)
+
+      COMMON  /BLOCK/ BAR
+      RETURN
+      END

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/common/block.f

@@ -0,0 +1,11 @@
+      SUBROUTINE INITCB
+      DOUBLE PRECISION LONG
+      CHARACTER        STRING
+      INTEGER          OK
+    
+      COMMON  /BLOCK/ LONG, STRING, OK
+      LONG = 1.0
+      STRING = '2'
+      OK = 3
+      RETURN
+      END

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/common/gh19161.f90

@@ -0,0 +1,10 @@
+module typedefmod
+  use iso_fortran_env, only: real32
+end module typedefmod
+
+module data
+  use typedefmod, only: real32
+  implicit none
+  real(kind=real32) :: x
+  common/test/x
+end module data

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/isocintrin/isoCtests.f90

@@ -0,0 +1,34 @@
+  module coddity
+    use iso_c_binding, only: c_double, c_int, c_int64_t
+    implicit none
+    contains
+      subroutine c_add(a, b, c) bind(c, name="c_add")
+        real(c_double), intent(in) :: a, b
+        real(c_double), intent(out) :: c
+        c = a + b
+      end subroutine c_add
+      ! gh-9693
+      function wat(x, y) result(z) bind(c)
+          integer(c_int), intent(in) :: x, y
+          integer(c_int) :: z
+
+          z = x + 7
+      end function wat
+      ! gh-25207
+      subroutine c_add_int64(a, b, c) bind(c)
+        integer(c_int64_t), intent(in) :: a, b
+        integer(c_int64_t), intent(out) :: c
+        c = a + b
+      end subroutine c_add_int64
+      ! gh-25207
+      subroutine add_arr(A, B, C)
+         integer(c_int64_t), intent(in) :: A(3)
+         integer(c_int64_t), intent(in) :: B(3)
+         integer(c_int64_t), intent(out) :: C(3)
+         integer :: j
+
+         do j = 1, 3
+            C(j) = A(j)+B(j)
+         end do
+      end subroutine
+  end module coddity

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/mixed/foo.f

@@ -0,0 +1,5 @@
+      subroutine bar11(a)
+cf2py intent(out) a
+      integer a
+      a = 11
+      end

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/mixed/foo_fixed.f90

@@ -0,0 +1,8 @@
+      module foo_fixed
+      contains
+        subroutine bar12(a)
+!f2py intent(out) a
+          integer a
+          a = 12
+        end subroutine bar12
+      end module foo_fixed

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/mixed/foo_free.f90

@@ -0,0 +1,8 @@
+module foo_free
+contains
+  subroutine bar13(a)
+    !f2py intent(out) a
+    integer a
+    a = 13
+  end subroutine bar13
+end module foo_free

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/module_data/module_data_docstring.f90

@@ -0,0 +1,12 @@
+module mod
+  integer :: i
+  integer :: x(4)
+  real, dimension(2,3) :: a
+  real, allocatable, dimension(:,:) :: b
+contains
+  subroutine foo
+    integer :: k
+    k = 1
+    a(1,2) = a(1,2)+3
+  end subroutine foo
+end module mod

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/regression/gh25337/data.f90

@@ -0,0 +1,8 @@
+module data
+   real(8) :: shift
+contains
+   subroutine set_shift(in_shift)
+      real(8), intent(in) :: in_shift
+      shift = in_shift
+   end subroutine set_shift
+end module data

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/regression/gh25337/use_data.f90

@@ -0,0 +1,6 @@
+subroutine shift_a(dim_a, a)
+    use data, only: shift
+    integer, intent(in) :: dim_a
+    real(8), intent(inout), dimension(dim_a) :: a
+    a = a + shift
+end subroutine shift_a

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/regression/inout.f90

@@ -0,0 +1,9 @@
+! Check that intent(in out) translates as intent(inout).
+! The separation seems to be a common usage.
+      subroutine foo(x)
+          implicit none
+          real(4), intent(in out) :: x
+          dimension x(3)
+          x(1) = x(1) + x(2) + x(3)
+          return
+      end

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_complex/foo77.f

@@ -0,0 +1,45 @@
+       function t0(value)
+         complex value
+         complex t0
+         t0 = value
+       end
+       function t8(value)
+         complex*8 value
+         complex*8 t8
+         t8 = value
+       end
+       function t16(value)
+         complex*16 value
+         complex*16 t16
+         t16 = value
+       end
+       function td(value)
+         double complex value
+         double complex td
+         td = value
+       end
+
+       subroutine s0(t0,value)
+         complex value
+         complex t0
+cf2py    intent(out) t0
+         t0 = value
+       end
+       subroutine s8(t8,value)
+         complex*8 value
+         complex*8 t8
+cf2py    intent(out) t8
+         t8 = value
+       end
+       subroutine s16(t16,value)
+         complex*16 value
+         complex*16 t16
+cf2py    intent(out) t16
+         t16 = value
+       end
+       subroutine sd(td,value)
+         double complex value
+         double complex td
+cf2py    intent(out) td
+         td = value
+       end

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_complex/foo90.f90

@@ -0,0 +1,48 @@
+module f90_return_complex
+  contains
+       function t0(value)
+         complex :: value
+         complex :: t0
+         t0 = value
+       end function t0
+       function t8(value)
+         complex(kind=4) :: value
+         complex(kind=4) :: t8
+         t8 = value
+       end function t8
+       function t16(value)
+         complex(kind=8) :: value
+         complex(kind=8) :: t16
+         t16 = value
+       end function t16
+       function td(value)
+         double complex :: value
+         double complex :: td
+         td = value
+       end function td
+
+       subroutine s0(t0,value)
+         complex :: value
+         complex :: t0
+!f2py    intent(out) t0
+         t0 = value
+       end subroutine s0
+       subroutine s8(t8,value)
+         complex(kind=4) :: value
+         complex(kind=4) :: t8
+!f2py    intent(out) t8
+         t8 = value
+       end subroutine s8
+       subroutine s16(t16,value)
+         complex(kind=8) :: value
+         complex(kind=8) :: t16
+!f2py    intent(out) t16
+         t16 = value
+       end subroutine s16
+       subroutine sd(td,value)
+         double complex :: value
+         double complex :: td
+!f2py    intent(out) td
+         td = value
+       end subroutine sd
+end module f90_return_complex

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_logical/foo77.f

@@ -0,0 +1,56 @@
+       function t0(value)
+         logical value
+         logical t0
+         t0 = value
+       end
+       function t1(value)
+         logical*1 value
+         logical*1 t1
+         t1 = value
+       end
+       function t2(value)
+         logical*2 value
+         logical*2 t2
+         t2 = value
+       end
+       function t4(value)
+         logical*4 value
+         logical*4 t4
+         t4 = value
+       end
+c       function t8(value)
+c         logical*8 value
+c         logical*8 t8
+c         t8 = value
+c       end
+
+       subroutine s0(t0,value)
+         logical value
+         logical t0
+cf2py    intent(out) t0
+         t0 = value
+       end
+       subroutine s1(t1,value)
+         logical*1 value
+         logical*1 t1
+cf2py    intent(out) t1
+         t1 = value
+       end
+       subroutine s2(t2,value)
+         logical*2 value
+         logical*2 t2
+cf2py    intent(out) t2
+         t2 = value
+       end
+       subroutine s4(t4,value)
+         logical*4 value
+         logical*4 t4
+cf2py    intent(out) t4
+         t4 = value
+       end
+c       subroutine s8(t8,value)
+c         logical*8 value
+c         logical*8 t8
+cf2py    intent(out) t8
+c         t8 = value
+c       end

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_logical/foo90.f90

@@ -0,0 +1,59 @@
+module f90_return_logical
+  contains
+       function t0(value)
+         logical :: value
+         logical :: t0
+         t0 = value
+       end function t0
+       function t1(value)
+         logical(kind=1) :: value
+         logical(kind=1) :: t1
+         t1 = value
+       end function t1
+       function t2(value)
+         logical(kind=2) :: value
+         logical(kind=2) :: t2
+         t2 = value
+       end function t2
+       function t4(value)
+         logical(kind=4) :: value
+         logical(kind=4) :: t4
+         t4 = value
+       end function t4
+       function t8(value)
+         logical(kind=8) :: value
+         logical(kind=8) :: t8
+         t8 = value
+       end function t8
+
+       subroutine s0(t0,value)
+         logical :: value
+         logical :: t0
+!f2py    intent(out) t0
+         t0 = value
+       end subroutine s0
+       subroutine s1(t1,value)
+         logical(kind=1) :: value
+         logical(kind=1) :: t1
+!f2py    intent(out) t1
+         t1 = value
+       end subroutine s1
+       subroutine s2(t2,value)
+         logical(kind=2) :: value
+         logical(kind=2) :: t2
+!f2py    intent(out) t2
+         t2 = value
+       end subroutine s2
+       subroutine s4(t4,value)
+         logical(kind=4) :: value
+         logical(kind=4) :: t4
+!f2py    intent(out) t4
+         t4 = value
+       end subroutine s4
+       subroutine s8(t8,value)
+         logical(kind=8) :: value
+         logical(kind=8) :: t8
+!f2py    intent(out) t8
+         t8 = value
+       end subroutine s8
+end module f90_return_logical

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_real/foo77.f

@@ -0,0 +1,45 @@
+       function t0(value)
+         real value
+         real t0
+         t0 = value
+       end
+       function t4(value)
+         real*4 value
+         real*4 t4
+         t4 = value
+       end
+       function t8(value)
+         real*8 value
+         real*8 t8
+         t8 = value
+       end
+       function td(value)
+         double precision value
+         double precision td
+         td = value
+       end
+
+       subroutine s0(t0,value)
+         real value
+         real t0
+cf2py    intent(out) t0
+         t0 = value
+       end
+       subroutine s4(t4,value)
+         real*4 value
+         real*4 t4
+cf2py    intent(out) t4
+         t4 = value
+       end
+       subroutine s8(t8,value)
+         real*8 value
+         real*8 t8
+cf2py    intent(out) t8
+         t8 = value
+       end
+       subroutine sd(td,value)
+         double precision value
+         double precision td
+cf2py    intent(out) td
+         td = value
+       end

.venv/lib/python3.11/site-packages/numpy/f2py/tests/src/return_real/foo90.f90

@@ -0,0 +1,48 @@
+module f90_return_real
+  contains
+       function t0(value)
+         real :: value
+         real :: t0
+         t0 = value
+       end function t0
+       function t4(value)
+         real(kind=4) :: value
+         real(kind=4) :: t4
+         t4 = value
+       end function t4
+       function t8(value)
+         real(kind=8) :: value
+         real(kind=8) :: t8
+         t8 = value
+       end function t8
+       function td(value)
+         double precision :: value
+         double precision :: td
+         td = value
+       end function td
+
+       subroutine s0(t0,value)
+         real :: value
+         real :: t0
+!f2py    intent(out) t0
+         t0 = value
+       end subroutine s0
+       subroutine s4(t4,value)
+         real(kind=4) :: value
+         real(kind=4) :: t4
+!f2py    intent(out) t4
+         t4 = value
+       end subroutine s4
+       subroutine s8(t8,value)
+         real(kind=8) :: value
+         real(kind=8) :: t8
+!f2py    intent(out) t8
+         t8 = value
+       end subroutine s8
+       subroutine sd(td,value)
+         double precision :: value
+         double precision :: td
+!f2py    intent(out) td
+         td = value
+       end subroutine sd
+end module f90_return_real

.venv/lib/python3.11/site-packages/numpy/matrixlib/__init__.py

@@ -0,0 +1,11 @@
+"""Sub-package containing the matrix class and related functions.
+
+"""
+from . import defmatrix
+from .defmatrix import *
+
+__all__ = defmatrix.__all__
+
+from numpy._pytesttester import PytestTester
+test = PytestTester(__name__)
+del PytestTester

.venv/lib/python3.11/site-packages/numpy/matrixlib/__init__.pyi

@@ -0,0 +1,15 @@
+from numpy._pytesttester import PytestTester
+
+from numpy import (
+    matrix as matrix,
+)
+
+from numpy.matrixlib.defmatrix import (
+    bmat as bmat,
+    mat as mat,
+    asmatrix as asmatrix,
+)
+
+__all__: list[str]
+__path__: list[str]
+test: PytestTester

.venv/lib/python3.11/site-packages/numpy/matrixlib/defmatrix.py

@@ -0,0 +1,1114 @@
+__all__ = ['matrix', 'bmat', 'mat', 'asmatrix']
+
+import sys
+import warnings
+import ast
+
+from .._utils import set_module
+import numpy.core.numeric as N
+from numpy.core.numeric import concatenate, isscalar
+# While not in __all__, matrix_power used to be defined here, so we import
+# it for backward compatibility.
+from numpy.linalg import matrix_power
+
+
+def _convert_from_string(data):
+    for char in '[]':
+        data = data.replace(char, '')
+
+    rows = data.split(';')
+    newdata = []
+    count = 0
+    for row in rows:
+        trow = row.split(',')
+        newrow = []
+        for col in trow:
+            temp = col.split()
+            newrow.extend(map(ast.literal_eval, temp))
+        if count == 0:
+            Ncols = len(newrow)
+        elif len(newrow) != Ncols:
+            raise ValueError("Rows not the same size.")
+        count += 1
+        newdata.append(newrow)
+    return newdata
+
+
+@set_module('numpy')
+def asmatrix(data, dtype=None):
+    """
+    Interpret the input as a matrix.
+
+    Unlike `matrix`, `asmatrix` does not make a copy if the input is already
+    a matrix or an ndarray.  Equivalent to ``matrix(data, copy=False)``.
+
+    Parameters
+    ----------
+    data : array_like
+        Input data.
+    dtype : data-type
+       Data-type of the output matrix.
+
+    Returns
+    -------
+    mat : matrix
+        `data` interpreted as a matrix.
+
+    Examples
+    --------
+    >>> x = np.array([[1, 2], [3, 4]])
+
+    >>> m = np.asmatrix(x)
+
+    >>> x[0,0] = 5
+
+    >>> m
+    matrix([[5, 2],
+            [3, 4]])
+
+    """
+    return matrix(data, dtype=dtype, copy=False)
+
+
+@set_module('numpy')
+class matrix(N.ndarray):
+    """
+    matrix(data, dtype=None, copy=True)
+
+    .. note:: It is no longer recommended to use this class, even for linear
+              algebra. Instead use regular arrays. The class may be removed
+              in the future.
+
+    Returns a matrix from an array-like object, or from a string of data.
+    A matrix is a specialized 2-D array that retains its 2-D nature
+    through operations.  It has certain special operators, such as ``*``
+    (matrix multiplication) and ``**`` (matrix power).
+
+    Parameters
+    ----------
+    data : array_like or string
+       If `data` is a string, it is interpreted as a matrix with commas
+       or spaces separating columns, and semicolons separating rows.
+    dtype : data-type
+       Data-type of the output matrix.
+    copy : bool
+       If `data` is already an `ndarray`, then this flag determines
+       whether the data is copied (the default), or whether a view is
+       constructed.
+
+    See Also
+    --------
+    array
+
+    Examples
+    --------
+    >>> a = np.matrix('1 2; 3 4')
+    >>> a
+    matrix([[1, 2],
+            [3, 4]])
+
+    >>> np.matrix([[1, 2], [3, 4]])
+    matrix([[1, 2],
+            [3, 4]])
+
+    """
+    __array_priority__ = 10.0
+    def __new__(subtype, data, dtype=None, copy=True):
+        warnings.warn('the matrix subclass is not the recommended way to '
+                      'represent matrices or deal with linear algebra (see '
+                      'https://docs.scipy.org/doc/numpy/user/'
+                      'numpy-for-matlab-users.html). '
+                      'Please adjust your code to use regular ndarray.',
+                      PendingDeprecationWarning, stacklevel=2)
+        if isinstance(data, matrix):
+            dtype2 = data.dtype
+            if (dtype is None):
+                dtype = dtype2
+            if (dtype2 == dtype) and (not copy):
+                return data
+            return data.astype(dtype)
+
+        if isinstance(data, N.ndarray):
+            if dtype is None:
+                intype = data.dtype
+            else:
+                intype = N.dtype(dtype)
+            new = data.view(subtype)
+            if intype != data.dtype:
+                return new.astype(intype)
+            if copy: return new.copy()
+            else: return new
+
+        if isinstance(data, str):
+            data = _convert_from_string(data)
+
+        # now convert data to an array
+        arr = N.array(data, dtype=dtype, copy=copy)
+        ndim = arr.ndim
+        shape = arr.shape
+        if (ndim > 2):
+            raise ValueError("matrix must be 2-dimensional")
+        elif ndim == 0:
+            shape = (1, 1)
+        elif ndim == 1:
+            shape = (1, shape[0])
+
+        order = 'C'
+        if (ndim == 2) and arr.flags.fortran:
+            order = 'F'
+
+        if not (order or arr.flags.contiguous):
+            arr = arr.copy()
+
+        ret = N.ndarray.__new__(subtype, shape, arr.dtype,
+                                buffer=arr,
+                                order=order)
+        return ret
+
+    def __array_finalize__(self, obj):
+        self._getitem = False
+        if (isinstance(obj, matrix) and obj._getitem): return
+        ndim = self.ndim
+        if (ndim == 2):
+            return
+        if (ndim > 2):
+            newshape = tuple([x for x in self.shape if x > 1])
+            ndim = len(newshape)
+            if ndim == 2:
+                self.shape = newshape
+                return
+            elif (ndim > 2):
+                raise ValueError("shape too large to be a matrix.")
+        else:
+            newshape = self.shape
+        if ndim == 0:
+            self.shape = (1, 1)
+        elif ndim == 1:
+            self.shape = (1, newshape[0])
+        return
+
+    def __getitem__(self, index):
+        self._getitem = True
+
+        try:
+            out = N.ndarray.__getitem__(self, index)
+        finally:
+            self._getitem = False
+
+        if not isinstance(out, N.ndarray):
+            return out
+
+        if out.ndim == 0:
+            return out[()]
+        if out.ndim == 1:
+            sh = out.shape[0]
+            # Determine when we should have a column array
+            try:
+                n = len(index)
+            except Exception:
+                n = 0
+            if n > 1 and isscalar(index[1]):
+                out.shape = (sh, 1)
+            else:
+                out.shape = (1, sh)
+        return out
+
+    def __mul__(self, other):
+        if isinstance(other, (N.ndarray, list, tuple)) :
+            # This promotes 1-D vectors to row vectors
+            return N.dot(self, asmatrix(other))
+        if isscalar(other) or not hasattr(other, '__rmul__') :
+            return N.dot(self, other)
+        return NotImplemented
+
+    def __rmul__(self, other):
+        return N.dot(other, self)
+
+    def __imul__(self, other):
+        self[:] = self * other
+        return self
+
+    def __pow__(self, other):
+        return matrix_power(self, other)
+
+    def __ipow__(self, other):
+        self[:] = self ** other
+        return self
+
+    def __rpow__(self, other):
+        return NotImplemented
+
+    def _align(self, axis):
+        """A convenience function for operations that need to preserve axis
+        orientation.
+        """
+        if axis is None:
+            return self[0, 0]
+        elif axis==0:
+            return self
+        elif axis==1:
+            return self.transpose()
+        else:
+            raise ValueError("unsupported axis")
+
+    def _collapse(self, axis):
+        """A convenience function for operations that want to collapse
+        to a scalar like _align, but are using keepdims=True
+        """
+        if axis is None:
+            return self[0, 0]
+        else:
+            return self
+
+    # Necessary because base-class tolist expects dimension
+    #  reduction by x[0]
+    def tolist(self):
+        """
+        Return the matrix as a (possibly nested) list.
+
+        See `ndarray.tolist` for full documentation.
+
+        See Also
+        --------
+        ndarray.tolist
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.tolist()
+        [[0, 1, 2, 3], [4, 5, 6, 7], [8, 9, 10, 11]]
+
+        """
+        return self.__array__().tolist()
+
+    # To preserve orientation of result...
+    def sum(self, axis=None, dtype=None, out=None):
+        """
+        Returns the sum of the matrix elements, along the given axis.
+
+        Refer to `numpy.sum` for full documentation.
+
+        See Also
+        --------
+        numpy.sum
+
+        Notes
+        -----
+        This is the same as `ndarray.sum`, except that where an `ndarray` would
+        be returned, a `matrix` object is returned instead.
+
+        Examples
+        --------
+        >>> x = np.matrix([[1, 2], [4, 3]])
+        >>> x.sum()
+        10
+        >>> x.sum(axis=1)
+        matrix([[3],
+                [7]])
+        >>> x.sum(axis=1, dtype='float')
+        matrix([[3.],
+                [7.]])
+        >>> out = np.zeros((2, 1), dtype='float')
+        >>> x.sum(axis=1, dtype='float', out=np.asmatrix(out))
+        matrix([[3.],
+                [7.]])
+
+        """
+        return N.ndarray.sum(self, axis, dtype, out, keepdims=True)._collapse(axis)
+
+
+    # To update docstring from array to matrix...
+    def squeeze(self, axis=None):
+        """
+        Return a possibly reshaped matrix.
+
+        Refer to `numpy.squeeze` for more documentation.
+
+        Parameters
+        ----------
+        axis : None or int or tuple of ints, optional
+            Selects a subset of the axes of length one in the shape.
+            If an axis is selected with shape entry greater than one,
+            an error is raised.
+
+        Returns
+        -------
+        squeezed : matrix
+            The matrix, but as a (1, N) matrix if it had shape (N, 1).
+
+        See Also
+        --------
+        numpy.squeeze : related function
+
+        Notes
+        -----
+        If `m` has a single column then that column is returned
+        as the single row of a matrix.  Otherwise `m` is returned.
+        The returned matrix is always either `m` itself or a view into `m`.
+        Supplying an axis keyword argument will not affect the returned matrix
+        but it may cause an error to be raised.
+
+        Examples
+        --------
+        >>> c = np.matrix([[1], [2]])
+        >>> c
+        matrix([[1],
+                [2]])
+        >>> c.squeeze()
+        matrix([[1, 2]])
+        >>> r = c.T
+        >>> r
+        matrix([[1, 2]])
+        >>> r.squeeze()
+        matrix([[1, 2]])
+        >>> m = np.matrix([[1, 2], [3, 4]])
+        >>> m.squeeze()
+        matrix([[1, 2],
+                [3, 4]])
+
+        """
+        return N.ndarray.squeeze(self, axis=axis)
+
+
+    # To update docstring from array to matrix...
+    def flatten(self, order='C'):
+        """
+        Return a flattened copy of the matrix.
+
+        All `N` elements of the matrix are placed into a single row.
+
+        Parameters
+        ----------
+        order : {'C', 'F', 'A', 'K'}, optional
+            'C' means to flatten in row-major (C-style) order. 'F' means to
+            flatten in column-major (Fortran-style) order. 'A' means to
+            flatten in column-major order if `m` is Fortran *contiguous* in
+            memory, row-major order otherwise. 'K' means to flatten `m` in
+            the order the elements occur in memory. The default is 'C'.
+
+        Returns
+        -------
+        y : matrix
+            A copy of the matrix, flattened to a `(1, N)` matrix where `N`
+            is the number of elements in the original matrix.
+
+        See Also
+        --------
+        ravel : Return a flattened array.
+        flat : A 1-D flat iterator over the matrix.
+
+        Examples
+        --------
+        >>> m = np.matrix([[1,2], [3,4]])
+        >>> m.flatten()
+        matrix([[1, 2, 3, 4]])
+        >>> m.flatten('F')
+        matrix([[1, 3, 2, 4]])
+
+        """
+        return N.ndarray.flatten(self, order=order)
+
+    def mean(self, axis=None, dtype=None, out=None):
+        """
+        Returns the average of the matrix elements along the given axis.
+
+        Refer to `numpy.mean` for full documentation.
+
+        See Also
+        --------
+        numpy.mean
+
+        Notes
+        -----
+        Same as `ndarray.mean` except that, where that returns an `ndarray`,
+        this returns a `matrix` object.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3, 4)))
+        >>> x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.mean()
+        5.5
+        >>> x.mean(0)
+        matrix([[4., 5., 6., 7.]])
+        >>> x.mean(1)
+        matrix([[ 1.5],
+                [ 5.5],
+                [ 9.5]])
+
+        """
+        return N.ndarray.mean(self, axis, dtype, out, keepdims=True)._collapse(axis)
+
+    def std(self, axis=None, dtype=None, out=None, ddof=0):
+        """
+        Return the standard deviation of the array elements along the given axis.
+
+        Refer to `numpy.std` for full documentation.
+
+        See Also
+        --------
+        numpy.std
+
+        Notes
+        -----
+        This is the same as `ndarray.std`, except that where an `ndarray` would
+        be returned, a `matrix` object is returned instead.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3, 4)))
+        >>> x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.std()
+        3.4520525295346629 # may vary
+        >>> x.std(0)
+        matrix([[ 3.26598632,  3.26598632,  3.26598632,  3.26598632]]) # may vary
+        >>> x.std(1)
+        matrix([[ 1.11803399],
+                [ 1.11803399],
+                [ 1.11803399]])
+
+        """
+        return N.ndarray.std(self, axis, dtype, out, ddof, keepdims=True)._collapse(axis)
+
+    def var(self, axis=None, dtype=None, out=None, ddof=0):
+        """
+        Returns the variance of the matrix elements, along the given axis.
+
+        Refer to `numpy.var` for full documentation.
+
+        See Also
+        --------
+        numpy.var
+
+        Notes
+        -----
+        This is the same as `ndarray.var`, except that where an `ndarray` would
+        be returned, a `matrix` object is returned instead.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3, 4)))
+        >>> x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.var()
+        11.916666666666666
+        >>> x.var(0)
+        matrix([[ 10.66666667,  10.66666667,  10.66666667,  10.66666667]]) # may vary
+        >>> x.var(1)
+        matrix([[1.25],
+                [1.25],
+                [1.25]])
+
+        """
+        return N.ndarray.var(self, axis, dtype, out, ddof, keepdims=True)._collapse(axis)
+
+    def prod(self, axis=None, dtype=None, out=None):
+        """
+        Return the product of the array elements over the given axis.
+
+        Refer to `prod` for full documentation.
+
+        See Also
+        --------
+        prod, ndarray.prod
+
+        Notes
+        -----
+        Same as `ndarray.prod`, except, where that returns an `ndarray`, this
+        returns a `matrix` object instead.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.prod()
+        0
+        >>> x.prod(0)
+        matrix([[  0,  45, 120, 231]])
+        >>> x.prod(1)
+        matrix([[   0],
+                [ 840],
+                [7920]])
+
+        """
+        return N.ndarray.prod(self, axis, dtype, out, keepdims=True)._collapse(axis)
+
+    def any(self, axis=None, out=None):
+        """
+        Test whether any array element along a given axis evaluates to True.
+
+        Refer to `numpy.any` for full documentation.
+
+        Parameters
+        ----------
+        axis : int, optional
+            Axis along which logical OR is performed
+        out : ndarray, optional
+            Output to existing array instead of creating new one, must have
+            same shape as expected output
+
+        Returns
+        -------
+            any : bool, ndarray
+                Returns a single bool if `axis` is ``None``; otherwise,
+                returns `ndarray`
+
+        """
+        return N.ndarray.any(self, axis, out, keepdims=True)._collapse(axis)
+
+    def all(self, axis=None, out=None):
+        """
+        Test whether all matrix elements along a given axis evaluate to True.
+
+        Parameters
+        ----------
+        See `numpy.all` for complete descriptions
+
+        See Also
+        --------
+        numpy.all
+
+        Notes
+        -----
+        This is the same as `ndarray.all`, but it returns a `matrix` object.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> y = x[0]; y
+        matrix([[0, 1, 2, 3]])
+        >>> (x == y)
+        matrix([[ True,  True,  True,  True],
+                [False, False, False, False],
+                [False, False, False, False]])
+        >>> (x == y).all()
+        False
+        >>> (x == y).all(0)
+        matrix([[False, False, False, False]])
+        >>> (x == y).all(1)
+        matrix([[ True],
+                [False],
+                [False]])
+
+        """
+        return N.ndarray.all(self, axis, out, keepdims=True)._collapse(axis)
+
+    def max(self, axis=None, out=None):
+        """
+        Return the maximum value along an axis.
+
+        Parameters
+        ----------
+        See `amax` for complete descriptions
+
+        See Also
+        --------
+        amax, ndarray.max
+
+        Notes
+        -----
+        This is the same as `ndarray.max`, but returns a `matrix` object
+        where `ndarray.max` would return an ndarray.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.max()
+        11
+        >>> x.max(0)
+        matrix([[ 8,  9, 10, 11]])
+        >>> x.max(1)
+        matrix([[ 3],
+                [ 7],
+                [11]])
+
+        """
+        return N.ndarray.max(self, axis, out, keepdims=True)._collapse(axis)
+
+    def argmax(self, axis=None, out=None):
+        """
+        Indexes of the maximum values along an axis.
+
+        Return the indexes of the first occurrences of the maximum values
+        along the specified axis.  If axis is None, the index is for the
+        flattened matrix.
+
+        Parameters
+        ----------
+        See `numpy.argmax` for complete descriptions
+
+        See Also
+        --------
+        numpy.argmax
+
+        Notes
+        -----
+        This is the same as `ndarray.argmax`, but returns a `matrix` object
+        where `ndarray.argmax` would return an `ndarray`.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.argmax()
+        11
+        >>> x.argmax(0)
+        matrix([[2, 2, 2, 2]])
+        >>> x.argmax(1)
+        matrix([[3],
+                [3],
+                [3]])
+
+        """
+        return N.ndarray.argmax(self, axis, out)._align(axis)
+
+    def min(self, axis=None, out=None):
+        """
+        Return the minimum value along an axis.
+
+        Parameters
+        ----------
+        See `amin` for complete descriptions.
+
+        See Also
+        --------
+        amin, ndarray.min
+
+        Notes
+        -----
+        This is the same as `ndarray.min`, but returns a `matrix` object
+        where `ndarray.min` would return an ndarray.
+
+        Examples
+        --------
+        >>> x = -np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[  0,  -1,  -2,  -3],
+                [ -4,  -5,  -6,  -7],
+                [ -8,  -9, -10, -11]])
+        >>> x.min()
+        -11
+        >>> x.min(0)
+        matrix([[ -8,  -9, -10, -11]])
+        >>> x.min(1)
+        matrix([[ -3],
+                [ -7],
+                [-11]])
+
+        """
+        return N.ndarray.min(self, axis, out, keepdims=True)._collapse(axis)
+
+    def argmin(self, axis=None, out=None):
+        """
+        Indexes of the minimum values along an axis.
+
+        Return the indexes of the first occurrences of the minimum values
+        along the specified axis.  If axis is None, the index is for the
+        flattened matrix.
+
+        Parameters
+        ----------
+        See `numpy.argmin` for complete descriptions.
+
+        See Also
+        --------
+        numpy.argmin
+
+        Notes
+        -----
+        This is the same as `ndarray.argmin`, but returns a `matrix` object
+        where `ndarray.argmin` would return an `ndarray`.
+
+        Examples
+        --------
+        >>> x = -np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[  0,  -1,  -2,  -3],
+                [ -4,  -5,  -6,  -7],
+                [ -8,  -9, -10, -11]])
+        >>> x.argmin()
+        11
+        >>> x.argmin(0)
+        matrix([[2, 2, 2, 2]])
+        >>> x.argmin(1)
+        matrix([[3],
+                [3],
+                [3]])
+
+        """
+        return N.ndarray.argmin(self, axis, out)._align(axis)
+
+    def ptp(self, axis=None, out=None):
+        """
+        Peak-to-peak (maximum - minimum) value along the given axis.
+
+        Refer to `numpy.ptp` for full documentation.
+
+        See Also
+        --------
+        numpy.ptp
+
+        Notes
+        -----
+        Same as `ndarray.ptp`, except, where that would return an `ndarray` object,
+        this returns a `matrix` object.
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.ptp()
+        11
+        >>> x.ptp(0)
+        matrix([[8, 8, 8, 8]])
+        >>> x.ptp(1)
+        matrix([[3],
+                [3],
+                [3]])
+
+        """
+        return N.ndarray.ptp(self, axis, out)._align(axis)
+
+    @property
+    def I(self):
+        """
+        Returns the (multiplicative) inverse of invertible `self`.
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        ret : matrix object
+            If `self` is non-singular, `ret` is such that ``ret * self`` ==
+            ``self * ret`` == ``np.matrix(np.eye(self[0,:].size))`` all return
+            ``True``.
+
+        Raises
+        ------
+        numpy.linalg.LinAlgError: Singular matrix
+            If `self` is singular.
+
+        See Also
+        --------
+        linalg.inv
+
+        Examples
+        --------
+        >>> m = np.matrix('[1, 2; 3, 4]'); m
+        matrix([[1, 2],
+                [3, 4]])
+        >>> m.getI()
+        matrix([[-2. ,  1. ],
+                [ 1.5, -0.5]])
+        >>> m.getI() * m
+        matrix([[ 1.,  0.], # may vary
+                [ 0.,  1.]])
+
+        """
+        M, N = self.shape
+        if M == N:
+            from numpy.linalg import inv as func
+        else:
+            from numpy.linalg import pinv as func
+        return asmatrix(func(self))
+
+    @property
+    def A(self):
+        """
+        Return `self` as an `ndarray` object.
+
+        Equivalent to ``np.asarray(self)``.
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        ret : ndarray
+            `self` as an `ndarray`
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.getA()
+        array([[ 0,  1,  2,  3],
+               [ 4,  5,  6,  7],
+               [ 8,  9, 10, 11]])
+
+        """
+        return self.__array__()
+
+    @property
+    def A1(self):
+        """
+        Return `self` as a flattened `ndarray`.
+
+        Equivalent to ``np.asarray(x).ravel()``
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        ret : ndarray
+            `self`, 1-D, as an `ndarray`
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4))); x
+        matrix([[ 0,  1,  2,  3],
+                [ 4,  5,  6,  7],
+                [ 8,  9, 10, 11]])
+        >>> x.getA1()
+        array([ 0,  1,  2, ...,  9, 10, 11])
+
+
+        """
+        return self.__array__().ravel()
+
+
+    def ravel(self, order='C'):
+        """
+        Return a flattened matrix.
+
+        Refer to `numpy.ravel` for more documentation.
+
+        Parameters
+        ----------
+        order : {'C', 'F', 'A', 'K'}, optional
+            The elements of `m` are read using this index order. 'C' means to
+            index the elements in C-like order, with the last axis index
+            changing fastest, back to the first axis index changing slowest.
+            'F' means to index the elements in Fortran-like index order, with
+            the first index changing fastest, and the last index changing
+            slowest. Note that the 'C' and 'F' options take no account of the
+            memory layout of the underlying array, and only refer to the order
+            of axis indexing.  'A' means to read the elements in Fortran-like
+            index order if `m` is Fortran *contiguous* in memory, C-like order
+            otherwise.  'K' means to read the elements in the order they occur
+            in memory, except for reversing the data when strides are negative.
+            By default, 'C' index order is used.
+
+        Returns
+        -------
+        ret : matrix
+            Return the matrix flattened to shape `(1, N)` where `N`
+            is the number of elements in the original matrix.
+            A copy is made only if necessary.
+
+        See Also
+        --------
+        matrix.flatten : returns a similar output matrix but always a copy
+        matrix.flat : a flat iterator on the array.
+        numpy.ravel : related function which returns an ndarray
+
+        """
+        return N.ndarray.ravel(self, order=order)
+
+    @property
+    def T(self):
+        """
+        Returns the transpose of the matrix.
+
+        Does *not* conjugate!  For the complex conjugate transpose, use ``.H``.
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        ret : matrix object
+            The (non-conjugated) transpose of the matrix.
+
+        See Also
+        --------
+        transpose, getH
+
+        Examples
+        --------
+        >>> m = np.matrix('[1, 2; 3, 4]')
+        >>> m
+        matrix([[1, 2],
+                [3, 4]])
+        >>> m.getT()
+        matrix([[1, 3],
+                [2, 4]])
+
+        """
+        return self.transpose()
+
+    @property
+    def H(self):
+        """
+        Returns the (complex) conjugate transpose of `self`.
+
+        Equivalent to ``np.transpose(self)`` if `self` is real-valued.
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        ret : matrix object
+            complex conjugate transpose of `self`
+
+        Examples
+        --------
+        >>> x = np.matrix(np.arange(12).reshape((3,4)))
+        >>> z = x - 1j*x; z
+        matrix([[  0. +0.j,   1. -1.j,   2. -2.j,   3. -3.j],
+                [  4. -4.j,   5. -5.j,   6. -6.j,   7. -7.j],
+                [  8. -8.j,   9. -9.j,  10.-10.j,  11.-11.j]])
+        >>> z.getH()
+        matrix([[ 0. -0.j,  4. +4.j,  8. +8.j],
+                [ 1. +1.j,  5. +5.j,  9. +9.j],
+                [ 2. +2.j,  6. +6.j, 10.+10.j],
+                [ 3. +3.j,  7. +7.j, 11.+11.j]])
+
+        """
+        if issubclass(self.dtype.type, N.complexfloating):
+            return self.transpose().conjugate()
+        else:
+            return self.transpose()
+
+    # kept for compatibility
+    getT = T.fget
+    getA = A.fget
+    getA1 = A1.fget
+    getH = H.fget
+    getI = I.fget
+
+def _from_string(str, gdict, ldict):
+    rows = str.split(';')
+    rowtup = []
+    for row in rows:
+        trow = row.split(',')
+        newrow = []
+        for x in trow:
+            newrow.extend(x.split())
+        trow = newrow
+        coltup = []
+        for col in trow:
+            col = col.strip()
+            try:
+                thismat = ldict[col]
+            except KeyError:
+                try:
+                    thismat = gdict[col]
+                except KeyError as e:
+                    raise NameError(f"name {col!r} is not defined") from None
+
+            coltup.append(thismat)
+        rowtup.append(concatenate(coltup, axis=-1))
+    return concatenate(rowtup, axis=0)
+
+
+@set_module('numpy')
+def bmat(obj, ldict=None, gdict=None):
+    """
+    Build a matrix object from a string, nested sequence, or array.
+
+    Parameters
+    ----------
+    obj : str or array_like
+        Input data. If a string, variables in the current scope may be
+        referenced by name.
+    ldict : dict, optional
+        A dictionary that replaces local operands in current frame.
+        Ignored if `obj` is not a string or `gdict` is None.
+    gdict : dict, optional
+        A dictionary that replaces global operands in current frame.
+        Ignored if `obj` is not a string.
+
+    Returns
+    -------
+    out : matrix
+        Returns a matrix object, which is a specialized 2-D array.
+
+    See Also
+    --------
+    block :
+        A generalization of this function for N-d arrays, that returns normal
+        ndarrays.
+
+    Examples
+    --------
+    >>> A = np.mat('1 1; 1 1')
+    >>> B = np.mat('2 2; 2 2')
+    >>> C = np.mat('3 4; 5 6')
+    >>> D = np.mat('7 8; 9 0')
+
+    All the following expressions construct the same block matrix:
+
+    >>> np.bmat([[A, B], [C, D]])
+    matrix([[1, 1, 2, 2],
+            [1, 1, 2, 2],
+            [3, 4, 7, 8],
+            [5, 6, 9, 0]])
+    >>> np.bmat(np.r_[np.c_[A, B], np.c_[C, D]])
+    matrix([[1, 1, 2, 2],
+            [1, 1, 2, 2],
+            [3, 4, 7, 8],
+            [5, 6, 9, 0]])
+    >>> np.bmat('A,B; C,D')
+    matrix([[1, 1, 2, 2],
+            [1, 1, 2, 2],
+            [3, 4, 7, 8],
+            [5, 6, 9, 0]])
+
+    """
+    if isinstance(obj, str):
+        if gdict is None:
+            # get previous frame
+            frame = sys._getframe().f_back
+            glob_dict = frame.f_globals
+            loc_dict = frame.f_locals
+        else:
+            glob_dict = gdict
+            loc_dict = ldict
+
+        return matrix(_from_string(obj, glob_dict, loc_dict))
+
+    if isinstance(obj, (tuple, list)):
+        # [[A,B],[C,D]]
+        arr_rows = []
+        for row in obj:
+            if isinstance(row, N.ndarray):  # not 2-d
+                return matrix(concatenate(obj, axis=-1))
+            else:
+                arr_rows.append(concatenate(row, axis=-1))
+        return matrix(concatenate(arr_rows, axis=0))
+    if isinstance(obj, N.ndarray):
+        return matrix(obj)
+
+mat = asmatrix

.venv/lib/python3.11/site-packages/numpy/matrixlib/defmatrix.pyi

@@ -0,0 +1,16 @@
+from collections.abc import Sequence, Mapping
+from typing import Any
+from numpy import matrix as matrix
+from numpy._typing import ArrayLike, DTypeLike, NDArray
+
+__all__: list[str]
+
+def bmat(
+    obj: str | Sequence[ArrayLike] | NDArray[Any],
+    ldict: None | Mapping[str, Any] = ...,
+    gdict: None | Mapping[str, Any] = ...,
+) -> matrix[Any, Any]: ...
+
+def asmatrix(data: ArrayLike, dtype: DTypeLike = ...) -> matrix[Any, Any]: ...
+
+mat = asmatrix

.venv/lib/python3.11/site-packages/numpy/matrixlib/setup.py

@@ -0,0 +1,12 @@
+#!/usr/bin/env python3
+def configuration(parent_package='', top_path=None):
+    from numpy.distutils.misc_util import Configuration
+    config = Configuration('matrixlib', parent_package, top_path)
+    config.add_subpackage('tests')
+    config.add_data_files('*.pyi')
+    return config
+
+if __name__ == "__main__":
+    from numpy.distutils.core import setup
+    config = configuration(top_path='').todict()
+    setup(**config)

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_defmatrix.py

@@ -0,0 +1,453 @@
+import collections.abc
+
+import numpy as np
+from numpy import matrix, asmatrix, bmat
+from numpy.testing import (
+    assert_, assert_equal, assert_almost_equal, assert_array_equal,
+    assert_array_almost_equal, assert_raises
+    )
+from numpy.linalg import matrix_power
+from numpy.matrixlib import mat
+
+class TestCtor:
+    def test_basic(self):
+        A = np.array([[1, 2], [3, 4]])
+        mA = matrix(A)
+        assert_(np.all(mA.A == A))
+
+        B = bmat("A,A;A,A")
+        C = bmat([[A, A], [A, A]])
+        D = np.array([[1, 2, 1, 2],
+                      [3, 4, 3, 4],
+                      [1, 2, 1, 2],
+                      [3, 4, 3, 4]])
+        assert_(np.all(B.A == D))
+        assert_(np.all(C.A == D))
+
+        E = np.array([[5, 6], [7, 8]])
+        AEresult = matrix([[1, 2, 5, 6], [3, 4, 7, 8]])
+        assert_(np.all(bmat([A, E]) == AEresult))
+
+        vec = np.arange(5)
+        mvec = matrix(vec)
+        assert_(mvec.shape == (1, 5))
+
+    def test_exceptions(self):
+        # Check for ValueError when called with invalid string data.
+        assert_raises(ValueError, matrix, "invalid")
+
+    def test_bmat_nondefault_str(self):
+        A = np.array([[1, 2], [3, 4]])
+        B = np.array([[5, 6], [7, 8]])
+        Aresult = np.array([[1, 2, 1, 2],
+                            [3, 4, 3, 4],
+                            [1, 2, 1, 2],
+                            [3, 4, 3, 4]])
+        mixresult = np.array([[1, 2, 5, 6],
+                              [3, 4, 7, 8],
+                              [5, 6, 1, 2],
+                              [7, 8, 3, 4]])
+        assert_(np.all(bmat("A,A;A,A") == Aresult))
+        assert_(np.all(bmat("A,A;A,A", ldict={'A':B}) == Aresult))
+        assert_raises(TypeError, bmat, "A,A;A,A", gdict={'A':B})
+        assert_(
+            np.all(bmat("A,A;A,A", ldict={'A':A}, gdict={'A':B}) == Aresult))
+        b2 = bmat("A,B;C,D", ldict={'A':A,'B':B}, gdict={'C':B,'D':A})
+        assert_(np.all(b2 == mixresult))
+
+
+class TestProperties:
+    def test_sum(self):
+        """Test whether matrix.sum(axis=1) preserves orientation.
+        Fails in NumPy <= 0.9.6.2127.
+        """
+        M = matrix([[1, 2, 0, 0],
+                   [3, 4, 0, 0],
+                   [1, 2, 1, 2],
+                   [3, 4, 3, 4]])
+        sum0 = matrix([8, 12, 4, 6])
+        sum1 = matrix([3, 7, 6, 14]).T
+        sumall = 30
+        assert_array_equal(sum0, M.sum(axis=0))
+        assert_array_equal(sum1, M.sum(axis=1))
+        assert_equal(sumall, M.sum())
+
+        assert_array_equal(sum0, np.sum(M, axis=0))
+        assert_array_equal(sum1, np.sum(M, axis=1))
+        assert_equal(sumall, np.sum(M))
+
+    def test_prod(self):
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(x.prod(), 720)
+        assert_equal(x.prod(0), matrix([[4, 10, 18]]))
+        assert_equal(x.prod(1), matrix([[6], [120]]))
+
+        assert_equal(np.prod(x), 720)
+        assert_equal(np.prod(x, axis=0), matrix([[4, 10, 18]]))
+        assert_equal(np.prod(x, axis=1), matrix([[6], [120]]))
+
+        y = matrix([0, 1, 3])
+        assert_(y.prod() == 0)
+
+    def test_max(self):
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(x.max(), 6)
+        assert_equal(x.max(0), matrix([[4, 5, 6]]))
+        assert_equal(x.max(1), matrix([[3], [6]]))
+
+        assert_equal(np.max(x), 6)
+        assert_equal(np.max(x, axis=0), matrix([[4, 5, 6]]))
+        assert_equal(np.max(x, axis=1), matrix([[3], [6]]))
+
+    def test_min(self):
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(x.min(), 1)
+        assert_equal(x.min(0), matrix([[1, 2, 3]]))
+        assert_equal(x.min(1), matrix([[1], [4]]))
+
+        assert_equal(np.min(x), 1)
+        assert_equal(np.min(x, axis=0), matrix([[1, 2, 3]]))
+        assert_equal(np.min(x, axis=1), matrix([[1], [4]]))
+
+    def test_ptp(self):
+        x = np.arange(4).reshape((2, 2))
+        assert_(x.ptp() == 3)
+        assert_(np.all(x.ptp(0) == np.array([2, 2])))
+        assert_(np.all(x.ptp(1) == np.array([1, 1])))
+
+    def test_var(self):
+        x = np.arange(9).reshape((3, 3))
+        mx = x.view(np.matrix)
+        assert_equal(x.var(ddof=0), mx.var(ddof=0))
+        assert_equal(x.var(ddof=1), mx.var(ddof=1))
+
+    def test_basic(self):
+        import numpy.linalg as linalg
+
+        A = np.array([[1., 2.],
+                      [3., 4.]])
+        mA = matrix(A)
+        assert_(np.allclose(linalg.inv(A), mA.I))
+        assert_(np.all(np.array(np.transpose(A) == mA.T)))
+        assert_(np.all(np.array(np.transpose(A) == mA.H)))
+        assert_(np.all(A == mA.A))
+
+        B = A + 2j*A
+        mB = matrix(B)
+        assert_(np.allclose(linalg.inv(B), mB.I))
+        assert_(np.all(np.array(np.transpose(B) == mB.T)))
+        assert_(np.all(np.array(np.transpose(B).conj() == mB.H)))
+
+    def test_pinv(self):
+        x = matrix(np.arange(6).reshape(2, 3))
+        xpinv = matrix([[-0.77777778,  0.27777778],
+                        [-0.11111111,  0.11111111],
+                        [ 0.55555556, -0.05555556]])
+        assert_almost_equal(x.I, xpinv)
+
+    def test_comparisons(self):
+        A = np.arange(100).reshape(10, 10)
+        mA = matrix(A)
+        mB = matrix(A) + 0.1
+        assert_(np.all(mB == A+0.1))
+        assert_(np.all(mB == matrix(A+0.1)))
+        assert_(not np.any(mB == matrix(A-0.1)))
+        assert_(np.all(mA < mB))
+        assert_(np.all(mA <= mB))
+        assert_(np.all(mA <= mA))
+        assert_(not np.any(mA < mA))
+
+        assert_(not np.any(mB < mA))
+        assert_(np.all(mB >= mA))
+        assert_(np.all(mB >= mB))
+        assert_(not np.any(mB > mB))
+
+        assert_(np.all(mA == mA))
+        assert_(not np.any(mA == mB))
+        assert_(np.all(mB != mA))
+
+        assert_(not np.all(abs(mA) > 0))
+        assert_(np.all(abs(mB > 0)))
+
+    def test_asmatrix(self):
+        A = np.arange(100).reshape(10, 10)
+        mA = asmatrix(A)
+        A[0, 0] = -10
+        assert_(A[0, 0] == mA[0, 0])
+
+    def test_noaxis(self):
+        A = matrix([[1, 0], [0, 1]])
+        assert_(A.sum() == matrix(2))
+        assert_(A.mean() == matrix(0.5))
+
+    def test_repr(self):
+        A = matrix([[1, 0], [0, 1]])
+        assert_(repr(A) == "matrix([[1, 0],\n        [0, 1]])")
+
+    def test_make_bool_matrix_from_str(self):
+        A = matrix('True; True; False')
+        B = matrix([[True], [True], [False]])
+        assert_array_equal(A, B)
+
+class TestCasting:
+    def test_basic(self):
+        A = np.arange(100).reshape(10, 10)
+        mA = matrix(A)
+
+        mB = mA.copy()
+        O = np.ones((10, 10), np.float64) * 0.1
+        mB = mB + O
+        assert_(mB.dtype.type == np.float64)
+        assert_(np.all(mA != mB))
+        assert_(np.all(mB == mA+0.1))
+
+        mC = mA.copy()
+        O = np.ones((10, 10), np.complex128)
+        mC = mC * O
+        assert_(mC.dtype.type == np.complex128)
+        assert_(np.all(mA != mB))
+
+
+class TestAlgebra:
+    def test_basic(self):
+        import numpy.linalg as linalg
+
+        A = np.array([[1., 2.], [3., 4.]])
+        mA = matrix(A)
+
+        B = np.identity(2)
+        for i in range(6):
+            assert_(np.allclose((mA ** i).A, B))
+            B = np.dot(B, A)
+
+        Ainv = linalg.inv(A)
+        B = np.identity(2)
+        for i in range(6):
+            assert_(np.allclose((mA ** -i).A, B))
+            B = np.dot(B, Ainv)
+
+        assert_(np.allclose((mA * mA).A, np.dot(A, A)))
+        assert_(np.allclose((mA + mA).A, (A + A)))
+        assert_(np.allclose((3*mA).A, (3*A)))
+
+        mA2 = matrix(A)
+        mA2 *= 3
+        assert_(np.allclose(mA2.A, 3*A))
+
+    def test_pow(self):
+        """Test raising a matrix to an integer power works as expected."""
+        m = matrix("1. 2.; 3. 4.")
+        m2 = m.copy()
+        m2 **= 2
+        mi = m.copy()
+        mi **= -1
+        m4 = m2.copy()
+        m4 **= 2
+        assert_array_almost_equal(m2, m**2)
+        assert_array_almost_equal(m4, np.dot(m2, m2))
+        assert_array_almost_equal(np.dot(mi, m), np.eye(2))
+
+    def test_scalar_type_pow(self):
+        m = matrix([[1, 2], [3, 4]])
+        for scalar_t in [np.int8, np.uint8]:
+            two = scalar_t(2)
+            assert_array_almost_equal(m ** 2, m ** two)
+
+    def test_notimplemented(self):
+        '''Check that 'not implemented' operations produce a failure.'''
+        A = matrix([[1., 2.],
+                    [3., 4.]])
+
+        # __rpow__
+        with assert_raises(TypeError):
+            1.0**A
+
+        # __mul__ with something not a list, ndarray, tuple, or scalar
+        with assert_raises(TypeError):
+            A*object()
+
+
+class TestMatrixReturn:
+    def test_instance_methods(self):
+        a = matrix([1.0], dtype='f8')
+        methodargs = {
+            'astype': ('intc',),
+            'clip': (0.0, 1.0),
+            'compress': ([1],),
+            'repeat': (1,),
+            'reshape': (1,),
+            'swapaxes': (0, 0),
+            'dot': np.array([1.0]),
+            }
+        excluded_methods = [
+            'argmin', 'choose', 'dump', 'dumps', 'fill', 'getfield',
+            'getA', 'getA1', 'item', 'nonzero', 'put', 'putmask', 'resize',
+            'searchsorted', 'setflags', 'setfield', 'sort',
+            'partition', 'argpartition',
+            'take', 'tofile', 'tolist', 'tostring', 'tobytes', 'all', 'any',
+            'sum', 'argmax', 'argmin', 'min', 'max', 'mean', 'var', 'ptp',
+            'prod', 'std', 'ctypes', 'itemset',
+            ]
+        for attrib in dir(a):
+            if attrib.startswith('_') or attrib in excluded_methods:
+                continue
+            f = getattr(a, attrib)
+            if isinstance(f, collections.abc.Callable):
+                # reset contents of a
+                a.astype('f8')
+                a.fill(1.0)
+                if attrib in methodargs:
+                    args = methodargs[attrib]
+                else:
+                    args = ()
+                b = f(*args)
+                assert_(type(b) is matrix, "%s" % attrib)
+        assert_(type(a.real) is matrix)
+        assert_(type(a.imag) is matrix)
+        c, d = matrix([0.0]).nonzero()
+        assert_(type(c) is np.ndarray)
+        assert_(type(d) is np.ndarray)
+
+
+class TestIndexing:
+    def test_basic(self):
+        x = asmatrix(np.zeros((3, 2), float))
+        y = np.zeros((3, 1), float)
+        y[:, 0] = [0.8, 0.2, 0.3]
+        x[:, 1] = y > 0.5
+        assert_equal(x, [[0, 1], [0, 0], [0, 0]])
+
+
+class TestNewScalarIndexing:
+    a = matrix([[1, 2], [3, 4]])
+
+    def test_dimesions(self):
+        a = self.a
+        x = a[0]
+        assert_equal(x.ndim, 2)
+
+    def test_array_from_matrix_list(self):
+        a = self.a
+        x = np.array([a, a])
+        assert_equal(x.shape, [2, 2, 2])
+
+    def test_array_to_list(self):
+        a = self.a
+        assert_equal(a.tolist(), [[1, 2], [3, 4]])
+
+    def test_fancy_indexing(self):
+        a = self.a
+        x = a[1, [0, 1, 0]]
+        assert_(isinstance(x, matrix))
+        assert_equal(x, matrix([[3,  4,  3]]))
+        x = a[[1, 0]]
+        assert_(isinstance(x, matrix))
+        assert_equal(x, matrix([[3,  4], [1, 2]]))
+        x = a[[[1], [0]], [[1, 0], [0, 1]]]
+        assert_(isinstance(x, matrix))
+        assert_equal(x, matrix([[4,  3], [1,  2]]))
+
+    def test_matrix_element(self):
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(x[0][0], matrix([[1, 2, 3]]))
+        assert_equal(x[0][0].shape, (1, 3))
+        assert_equal(x[0].shape, (1, 3))
+        assert_equal(x[:, 0].shape, (2, 1))
+
+        x = matrix(0)
+        assert_equal(x[0, 0], 0)
+        assert_equal(x[0], 0)
+        assert_equal(x[:, 0].shape, x.shape)
+
+    def test_scalar_indexing(self):
+        x = asmatrix(np.zeros((3, 2), float))
+        assert_equal(x[0, 0], x[0][0])
+
+    def test_row_column_indexing(self):
+        x = asmatrix(np.eye(2))
+        assert_array_equal(x[0,:], [[1, 0]])
+        assert_array_equal(x[1,:], [[0, 1]])
+        assert_array_equal(x[:, 0], [[1], [0]])
+        assert_array_equal(x[:, 1], [[0], [1]])
+
+    def test_boolean_indexing(self):
+        A = np.arange(6)
+        A.shape = (3, 2)
+        x = asmatrix(A)
+        assert_array_equal(x[:, np.array([True, False])], x[:, 0])
+        assert_array_equal(x[np.array([True, False, False]),:], x[0,:])
+
+    def test_list_indexing(self):
+        A = np.arange(6)
+        A.shape = (3, 2)
+        x = asmatrix(A)
+        assert_array_equal(x[:, [1, 0]], x[:, ::-1])
+        assert_array_equal(x[[2, 1, 0],:], x[::-1,:])
+
+
+class TestPower:
+    def test_returntype(self):
+        a = np.array([[0, 1], [0, 0]])
+        assert_(type(matrix_power(a, 2)) is np.ndarray)
+        a = mat(a)
+        assert_(type(matrix_power(a, 2)) is matrix)
+
+    def test_list(self):
+        assert_array_equal(matrix_power([[0, 1], [0, 0]], 2), [[0, 0], [0, 0]])
+
+
+class TestShape:
+
+    a = np.array([[1], [2]])
+    m = matrix([[1], [2]])
+
+    def test_shape(self):
+        assert_equal(self.a.shape, (2, 1))
+        assert_equal(self.m.shape, (2, 1))
+
+    def test_numpy_ravel(self):
+        assert_equal(np.ravel(self.a).shape, (2,))
+        assert_equal(np.ravel(self.m).shape, (2,))
+
+    def test_member_ravel(self):
+        assert_equal(self.a.ravel().shape, (2,))
+        assert_equal(self.m.ravel().shape, (1, 2))
+
+    def test_member_flatten(self):
+        assert_equal(self.a.flatten().shape, (2,))
+        assert_equal(self.m.flatten().shape, (1, 2))
+
+    def test_numpy_ravel_order(self):
+        x = np.array([[1, 2, 3], [4, 5, 6]])
+        assert_equal(np.ravel(x), [1, 2, 3, 4, 5, 6])
+        assert_equal(np.ravel(x, order='F'), [1, 4, 2, 5, 3, 6])
+        assert_equal(np.ravel(x.T), [1, 4, 2, 5, 3, 6])
+        assert_equal(np.ravel(x.T, order='A'), [1, 2, 3, 4, 5, 6])
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(np.ravel(x), [1, 2, 3, 4, 5, 6])
+        assert_equal(np.ravel(x, order='F'), [1, 4, 2, 5, 3, 6])
+        assert_equal(np.ravel(x.T), [1, 4, 2, 5, 3, 6])
+        assert_equal(np.ravel(x.T, order='A'), [1, 2, 3, 4, 5, 6])
+
+    def test_matrix_ravel_order(self):
+        x = matrix([[1, 2, 3], [4, 5, 6]])
+        assert_equal(x.ravel(), [[1, 2, 3, 4, 5, 6]])
+        assert_equal(x.ravel(order='F'), [[1, 4, 2, 5, 3, 6]])
+        assert_equal(x.T.ravel(), [[1, 4, 2, 5, 3, 6]])
+        assert_equal(x.T.ravel(order='A'), [[1, 2, 3, 4, 5, 6]])
+
+    def test_array_memory_sharing(self):
+        assert_(np.may_share_memory(self.a, self.a.ravel()))
+        assert_(not np.may_share_memory(self.a, self.a.flatten()))
+
+    def test_matrix_memory_sharing(self):
+        assert_(np.may_share_memory(self.m, self.m.ravel()))
+        assert_(not np.may_share_memory(self.m, self.m.flatten()))
+
+    def test_expand_dims_matrix(self):
+        # matrices are always 2d - so expand_dims only makes sense when the
+        # type is changed away from matrix.
+        a = np.arange(10).reshape((2, 5)).view(np.matrix)
+        expanded = np.expand_dims(a, axis=1)
+        assert_equal(expanded.ndim, 3)
+        assert_(not isinstance(expanded, np.matrix))

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_interaction.py

@@ -0,0 +1,354 @@
+"""Tests of interaction of matrix with other parts of numpy.
+
+Note that tests with MaskedArray and linalg are done in separate files.
+"""
+import pytest
+
+import textwrap
+import warnings
+
+import numpy as np
+from numpy.testing import (assert_, assert_equal, assert_raises,
+                           assert_raises_regex, assert_array_equal,
+                           assert_almost_equal, assert_array_almost_equal)
+
+
+def test_fancy_indexing():
+    # The matrix class messes with the shape. While this is always
+    # weird (getitem is not used, it does not have setitem nor knows
+    # about fancy indexing), this tests gh-3110
+    # 2018-04-29: moved here from core.tests.test_index.
+    m = np.matrix([[1, 2], [3, 4]])
+
+    assert_(isinstance(m[[0, 1, 0], :], np.matrix))
+
+    # gh-3110. Note the transpose currently because matrices do *not*
+    # support dimension fixing for fancy indexing correctly.
+    x = np.asmatrix(np.arange(50).reshape(5, 10))
+    assert_equal(x[:2, np.array(-1)], x[:2, -1].T)
+
+
+def test_polynomial_mapdomain():
+    # test that polynomial preserved matrix subtype.
+    # 2018-04-29: moved here from polynomial.tests.polyutils.
+    dom1 = [0, 4]
+    dom2 = [1, 3]
+    x = np.matrix([dom1, dom1])
+    res = np.polynomial.polyutils.mapdomain(x, dom1, dom2)
+    assert_(isinstance(res, np.matrix))
+
+
+def test_sort_matrix_none():
+    # 2018-04-29: moved here from core.tests.test_multiarray
+    a = np.matrix([[2, 1, 0]])
+    actual = np.sort(a, axis=None)
+    expected = np.matrix([[0, 1, 2]])
+    assert_equal(actual, expected)
+    assert_(type(expected) is np.matrix)
+
+
+def test_partition_matrix_none():
+    # gh-4301
+    # 2018-04-29: moved here from core.tests.test_multiarray
+    a = np.matrix([[2, 1, 0]])
+    actual = np.partition(a, 1, axis=None)
+    expected = np.matrix([[0, 1, 2]])
+    assert_equal(actual, expected)
+    assert_(type(expected) is np.matrix)
+
+
+def test_dot_scalar_and_matrix_of_objects():
+    # Ticket #2469
+    # 2018-04-29: moved here from core.tests.test_multiarray
+    arr = np.matrix([1, 2], dtype=object)
+    desired = np.matrix([[3, 6]], dtype=object)
+    assert_equal(np.dot(arr, 3), desired)
+    assert_equal(np.dot(3, arr), desired)
+
+
+def test_inner_scalar_and_matrix():
+    # 2018-04-29: moved here from core.tests.test_multiarray
+    for dt in np.typecodes['AllInteger'] + np.typecodes['AllFloat'] + '?':
+        sca = np.array(3, dtype=dt)[()]
+        arr = np.matrix([[1, 2], [3, 4]], dtype=dt)
+        desired = np.matrix([[3, 6], [9, 12]], dtype=dt)
+        assert_equal(np.inner(arr, sca), desired)
+        assert_equal(np.inner(sca, arr), desired)
+
+
+def test_inner_scalar_and_matrix_of_objects():
+    # Ticket #4482
+    # 2018-04-29: moved here from core.tests.test_multiarray
+    arr = np.matrix([1, 2], dtype=object)
+    desired = np.matrix([[3, 6]], dtype=object)
+    assert_equal(np.inner(arr, 3), desired)
+    assert_equal(np.inner(3, arr), desired)
+
+
+def test_iter_allocate_output_subtype():
+    # Make sure that the subtype with priority wins
+    # 2018-04-29: moved here from core.tests.test_nditer, given the
+    # matrix specific shape test.
+
+    # matrix vs ndarray
+    a = np.matrix([[1, 2], [3, 4]])
+    b = np.arange(4).reshape(2, 2).T
+    i = np.nditer([a, b, None], [],
+                  [['readonly'], ['readonly'], ['writeonly', 'allocate']])
+    assert_(type(i.operands[2]) is np.matrix)
+    assert_(type(i.operands[2]) is not np.ndarray)
+    assert_equal(i.operands[2].shape, (2, 2))
+
+    # matrix always wants things to be 2D
+    b = np.arange(4).reshape(1, 2, 2)
+    assert_raises(RuntimeError, np.nditer, [a, b, None], [],
+                  [['readonly'], ['readonly'], ['writeonly', 'allocate']])
+    # but if subtypes are disabled, the result can still work
+    i = np.nditer([a, b, None], [],
+                  [['readonly'], ['readonly'],
+                   ['writeonly', 'allocate', 'no_subtype']])
+    assert_(type(i.operands[2]) is np.ndarray)
+    assert_(type(i.operands[2]) is not np.matrix)
+    assert_equal(i.operands[2].shape, (1, 2, 2))
+
+
+def like_function():
+    # 2018-04-29: moved here from core.tests.test_numeric
+    a = np.matrix([[1, 2], [3, 4]])
+    for like_function in np.zeros_like, np.ones_like, np.empty_like:
+        b = like_function(a)
+        assert_(type(b) is np.matrix)
+
+        c = like_function(a, subok=False)
+        assert_(type(c) is not np.matrix)
+
+
+def test_array_astype():
+    # 2018-04-29: copied here from core.tests.test_api
+    # subok=True passes through a matrix
+    a = np.matrix([[0, 1, 2], [3, 4, 5]], dtype='f4')
+    b = a.astype('f4', subok=True, copy=False)
+    assert_(a is b)
+
+    # subok=True is default, and creates a subtype on a cast
+    b = a.astype('i4', copy=False)
+    assert_equal(a, b)
+    assert_equal(type(b), np.matrix)
+
+    # subok=False never returns a matrix
+    b = a.astype('f4', subok=False, copy=False)
+    assert_equal(a, b)
+    assert_(not (a is b))
+    assert_(type(b) is not np.matrix)
+
+
+def test_stack():
+    # 2018-04-29: copied here from core.tests.test_shape_base
+    # check np.matrix cannot be stacked
+    m = np.matrix([[1, 2], [3, 4]])
+    assert_raises_regex(ValueError, 'shape too large to be a matrix',
+                        np.stack, [m, m])
+
+
+def test_object_scalar_multiply():
+    # Tickets #2469 and #4482
+    # 2018-04-29: moved here from core.tests.test_ufunc
+    arr = np.matrix([1, 2], dtype=object)
+    desired = np.matrix([[3, 6]], dtype=object)
+    assert_equal(np.multiply(arr, 3), desired)
+    assert_equal(np.multiply(3, arr), desired)
+
+
+def test_nanfunctions_matrices():
+    # Check that it works and that type and
+    # shape are preserved
+    # 2018-04-29: moved here from core.tests.test_nanfunctions
+    mat = np.matrix(np.eye(3))
+    for f in [np.nanmin, np.nanmax]:
+        res = f(mat, axis=0)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (1, 3))
+        res = f(mat, axis=1)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (3, 1))
+        res = f(mat)
+        assert_(np.isscalar(res))
+    # check that rows of nan are dealt with for subclasses (#4628)
+    mat[1] = np.nan
+    for f in [np.nanmin, np.nanmax]:
+        with warnings.catch_warnings(record=True) as w:
+            warnings.simplefilter('always')
+            res = f(mat, axis=0)
+            assert_(isinstance(res, np.matrix))
+            assert_(not np.any(np.isnan(res)))
+            assert_(len(w) == 0)
+
+        with warnings.catch_warnings(record=True) as w:
+            warnings.simplefilter('always')
+            res = f(mat, axis=1)
+            assert_(isinstance(res, np.matrix))
+            assert_(np.isnan(res[1, 0]) and not np.isnan(res[0, 0])
+                    and not np.isnan(res[2, 0]))
+            assert_(len(w) == 1, 'no warning raised')
+            assert_(issubclass(w[0].category, RuntimeWarning))
+
+        with warnings.catch_warnings(record=True) as w:
+            warnings.simplefilter('always')
+            res = f(mat)
+            assert_(np.isscalar(res))
+            assert_(res != np.nan)
+            assert_(len(w) == 0)
+
+
+def test_nanfunctions_matrices_general():
+    # Check that it works and that type and
+    # shape are preserved
+    # 2018-04-29: moved here from core.tests.test_nanfunctions
+    mat = np.matrix(np.eye(3))
+    for f in (np.nanargmin, np.nanargmax, np.nansum, np.nanprod,
+              np.nanmean, np.nanvar, np.nanstd):
+        res = f(mat, axis=0)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (1, 3))
+        res = f(mat, axis=1)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (3, 1))
+        res = f(mat)
+        assert_(np.isscalar(res))
+
+    for f in np.nancumsum, np.nancumprod:
+        res = f(mat, axis=0)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (3, 3))
+        res = f(mat, axis=1)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (3, 3))
+        res = f(mat)
+        assert_(isinstance(res, np.matrix))
+        assert_(res.shape == (1, 3*3))
+
+
+def test_average_matrix():
+    # 2018-04-29: moved here from core.tests.test_function_base.
+    y = np.matrix(np.random.rand(5, 5))
+    assert_array_equal(y.mean(0), np.average(y, 0))
+
+    a = np.matrix([[1, 2], [3, 4]])
+    w = np.matrix([[1, 2], [3, 4]])
+
+    r = np.average(a, axis=0, weights=w)
+    assert_equal(type(r), np.matrix)
+    assert_equal(r, [[2.5, 10.0/3]])
+
+
+def test_trapz_matrix():
+    # Test to make sure matrices give the same answer as ndarrays
+    # 2018-04-29: moved here from core.tests.test_function_base.
+    x = np.linspace(0, 5)
+    y = x * x
+    r = np.trapz(y, x)
+    mx = np.matrix(x)
+    my = np.matrix(y)
+    mr = np.trapz(my, mx)
+    assert_almost_equal(mr, r)
+
+
+def test_ediff1d_matrix():
+    # 2018-04-29: moved here from core.tests.test_arraysetops.
+    assert(isinstance(np.ediff1d(np.matrix(1)), np.matrix))
+    assert(isinstance(np.ediff1d(np.matrix(1), to_begin=1), np.matrix))
+
+
+def test_apply_along_axis_matrix():
+    # this test is particularly malicious because matrix
+    # refuses to become 1d
+    # 2018-04-29: moved here from core.tests.test_shape_base.
+    def double(row):
+        return row * 2
+
+    m = np.matrix([[0, 1], [2, 3]])
+    expected = np.matrix([[0, 2], [4, 6]])
+
+    result = np.apply_along_axis(double, 0, m)
+    assert_(isinstance(result, np.matrix))
+    assert_array_equal(result, expected)
+
+    result = np.apply_along_axis(double, 1, m)
+    assert_(isinstance(result, np.matrix))
+    assert_array_equal(result, expected)
+
+
+def test_kron_matrix():
+    # 2018-04-29: moved here from core.tests.test_shape_base.
+    a = np.ones([2, 2])
+    m = np.asmatrix(a)
+    assert_equal(type(np.kron(a, a)), np.ndarray)
+    assert_equal(type(np.kron(m, m)), np.matrix)
+    assert_equal(type(np.kron(a, m)), np.matrix)
+    assert_equal(type(np.kron(m, a)), np.matrix)
+
+
+class TestConcatenatorMatrix:
+    # 2018-04-29: moved here from core.tests.test_index_tricks.
+    def test_matrix(self):
+        a = [1, 2]
+        b = [3, 4]
+
+        ab_r = np.r_['r', a, b]
+        ab_c = np.r_['c', a, b]
+
+        assert_equal(type(ab_r), np.matrix)
+        assert_equal(type(ab_c), np.matrix)
+
+        assert_equal(np.array(ab_r), [[1, 2, 3, 4]])
+        assert_equal(np.array(ab_c), [[1], [2], [3], [4]])
+
+        assert_raises(ValueError, lambda: np.r_['rc', a, b])
+
+    def test_matrix_scalar(self):
+        r = np.r_['r', [1, 2], 3]
+        assert_equal(type(r), np.matrix)
+        assert_equal(np.array(r), [[1, 2, 3]])
+
+    def test_matrix_builder(self):
+        a = np.array([1])
+        b = np.array([2])
+        c = np.array([3])
+        d = np.array([4])
+        actual = np.r_['a, b; c, d']
+        expected = np.bmat([[a, b], [c, d]])
+
+        assert_equal(actual, expected)
+        assert_equal(type(actual), type(expected))
+
+
+def test_array_equal_error_message_matrix():
+    # 2018-04-29: moved here from testing.tests.test_utils.
+    with pytest.raises(AssertionError) as exc_info:
+        assert_equal(np.array([1, 2]), np.matrix([1, 2]))
+    msg = str(exc_info.value)
+    msg_reference = textwrap.dedent("""\
+
+    Arrays are not equal
+
+    (shapes (2,), (1, 2) mismatch)
+     x: array([1, 2])
+     y: matrix([[1, 2]])""")
+    assert_equal(msg, msg_reference)
+
+
+def test_array_almost_equal_matrix():
+    # Matrix slicing keeps things 2-D, while array does not necessarily.
+    # See gh-8452.
+    # 2018-04-29: moved here from testing.tests.test_utils.
+    m1 = np.matrix([[1., 2.]])
+    m2 = np.matrix([[1., np.nan]])
+    m3 = np.matrix([[1., -np.inf]])
+    m4 = np.matrix([[np.nan, np.inf]])
+    m5 = np.matrix([[1., 2.], [np.nan, np.inf]])
+    for assert_func in assert_array_almost_equal, assert_almost_equal:
+        for m in m1, m2, m3, m4, m5:
+            assert_func(m, m)
+            a = np.array(m)
+            assert_func(a, m)
+            assert_func(m, a)

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_masked_matrix.py

@@ -0,0 +1,231 @@
+import numpy as np
+from numpy.testing import assert_warns
+from numpy.ma.testutils import (assert_, assert_equal, assert_raises,
+                                assert_array_equal)
+from numpy.ma.core import (masked_array, masked_values, masked, allequal,
+                           MaskType, getmask, MaskedArray, nomask,
+                           log, add, hypot, divide)
+from numpy.ma.extras import mr_
+from numpy.compat import pickle
+
+
+class MMatrix(MaskedArray, np.matrix,):
+
+    def __new__(cls, data, mask=nomask):
+        mat = np.matrix(data)
+        _data = MaskedArray.__new__(cls, data=mat, mask=mask)
+        return _data
+
+    def __array_finalize__(self, obj):
+        np.matrix.__array_finalize__(self, obj)
+        MaskedArray.__array_finalize__(self, obj)
+        return
+
+    @property
+    def _series(self):
+        _view = self.view(MaskedArray)
+        _view._sharedmask = False
+        return _view
+
+
+class TestMaskedMatrix:
+    def test_matrix_indexing(self):
+        # Tests conversions and indexing
+        x1 = np.matrix([[1, 2, 3], [4, 3, 2]])
+        x2 = masked_array(x1, mask=[[1, 0, 0], [0, 1, 0]])
+        x3 = masked_array(x1, mask=[[0, 1, 0], [1, 0, 0]])
+        x4 = masked_array(x1)
+        # test conversion to strings
+        str(x2)  # raises?
+        repr(x2)  # raises?
+        # tests of indexing
+        assert_(type(x2[1, 0]) is type(x1[1, 0]))
+        assert_(x1[1, 0] == x2[1, 0])
+        assert_(x2[1, 1] is masked)
+        assert_equal(x1[0, 2], x2[0, 2])
+        assert_equal(x1[0, 1:], x2[0, 1:])
+        assert_equal(x1[:, 2], x2[:, 2])
+        assert_equal(x1[:], x2[:])
+        assert_equal(x1[1:], x3[1:])
+        x1[0, 2] = 9
+        x2[0, 2] = 9
+        assert_equal(x1, x2)
+        x1[0, 1:] = 99
+        x2[0, 1:] = 99
+        assert_equal(x1, x2)
+        x2[0, 1] = masked
+        assert_equal(x1, x2)
+        x2[0, 1:] = masked
+        assert_equal(x1, x2)
+        x2[0, :] = x1[0, :]
+        x2[0, 1] = masked
+        assert_(allequal(getmask(x2), np.array([[0, 1, 0], [0, 1, 0]])))
+        x3[1, :] = masked_array([1, 2, 3], [1, 1, 0])
+        assert_(allequal(getmask(x3)[1], masked_array([1, 1, 0])))
+        assert_(allequal(getmask(x3[1]), masked_array([1, 1, 0])))
+        x4[1, :] = masked_array([1, 2, 3], [1, 1, 0])
+        assert_(allequal(getmask(x4[1]), masked_array([1, 1, 0])))
+        assert_(allequal(x4[1], masked_array([1, 2, 3])))
+        x1 = np.matrix(np.arange(5) * 1.0)
+        x2 = masked_values(x1, 3.0)
+        assert_equal(x1, x2)
+        assert_(allequal(masked_array([0, 0, 0, 1, 0], dtype=MaskType),
+                         x2.mask))
+        assert_equal(3.0, x2.fill_value)
+
+    def test_pickling_subbaseclass(self):
+        # Test pickling w/ a subclass of ndarray
+        a = masked_array(np.matrix(list(range(10))), mask=[1, 0, 1, 0, 0] * 2)
+        for proto in range(2, pickle.HIGHEST_PROTOCOL + 1):
+            a_pickled = pickle.loads(pickle.dumps(a, protocol=proto))
+            assert_equal(a_pickled._mask, a._mask)
+            assert_equal(a_pickled, a)
+            assert_(isinstance(a_pickled._data, np.matrix))
+
+    def test_count_mean_with_matrix(self):
+        m = masked_array(np.matrix([[1, 2], [3, 4]]), mask=np.zeros((2, 2)))
+
+        assert_equal(m.count(axis=0).shape, (1, 2))
+        assert_equal(m.count(axis=1).shape, (2, 1))
+
+        # Make sure broadcasting inside mean and var work
+        assert_equal(m.mean(axis=0), [[2., 3.]])
+        assert_equal(m.mean(axis=1), [[1.5], [3.5]])
+
+    def test_flat(self):
+        # Test that flat can return items even for matrices [#4585, #4615]
+        # test simple access
+        test = masked_array(np.matrix([[1, 2, 3]]), mask=[0, 0, 1])
+        assert_equal(test.flat[1], 2)
+        assert_equal(test.flat[2], masked)
+        assert_(np.all(test.flat[0:2] == test[0, 0:2]))
+        # Test flat on masked_matrices
+        test = masked_array(np.matrix([[1, 2, 3]]), mask=[0, 0, 1])
+        test.flat = masked_array([3, 2, 1], mask=[1, 0, 0])
+        control = masked_array(np.matrix([[3, 2, 1]]), mask=[1, 0, 0])
+        assert_equal(test, control)
+        # Test setting
+        test = masked_array(np.matrix([[1, 2, 3]]), mask=[0, 0, 1])
+        testflat = test.flat
+        testflat[:] = testflat[[2, 1, 0]]
+        assert_equal(test, control)
+        testflat[0] = 9
+        # test that matrices keep the correct shape (#4615)
+        a = masked_array(np.matrix(np.eye(2)), mask=0)
+        b = a.flat
+        b01 = b[:2]
+        assert_equal(b01.data, np.array([[1., 0.]]))
+        assert_equal(b01.mask, np.array([[False, False]]))
+
+    def test_allany_onmatrices(self):
+        x = np.array([[0.13, 0.26, 0.90],
+                      [0.28, 0.33, 0.63],
+                      [0.31, 0.87, 0.70]])
+        X = np.matrix(x)
+        m = np.array([[True, False, False],
+                      [False, False, False],
+                      [True, True, False]], dtype=np.bool_)
+        mX = masked_array(X, mask=m)
+        mXbig = (mX > 0.5)
+        mXsmall = (mX < 0.5)
+
+        assert_(not mXbig.all())
+        assert_(mXbig.any())
+        assert_equal(mXbig.all(0), np.matrix([False, False, True]))
+        assert_equal(mXbig.all(1), np.matrix([False, False, True]).T)
+        assert_equal(mXbig.any(0), np.matrix([False, False, True]))
+        assert_equal(mXbig.any(1), np.matrix([True, True, True]).T)
+
+        assert_(not mXsmall.all())
+        assert_(mXsmall.any())
+        assert_equal(mXsmall.all(0), np.matrix([True, True, False]))
+        assert_equal(mXsmall.all(1), np.matrix([False, False, False]).T)
+        assert_equal(mXsmall.any(0), np.matrix([True, True, False]))
+        assert_equal(mXsmall.any(1), np.matrix([True, True, False]).T)
+
+    def test_compressed(self):
+        a = masked_array(np.matrix([1, 2, 3, 4]), mask=[0, 0, 0, 0])
+        b = a.compressed()
+        assert_equal(b, a)
+        assert_(isinstance(b, np.matrix))
+        a[0, 0] = masked
+        b = a.compressed()
+        assert_equal(b, [[2, 3, 4]])
+
+    def test_ravel(self):
+        a = masked_array(np.matrix([1, 2, 3, 4, 5]), mask=[[0, 1, 0, 0, 0]])
+        aravel = a.ravel()
+        assert_equal(aravel.shape, (1, 5))
+        assert_equal(aravel._mask.shape, a.shape)
+
+    def test_view(self):
+        # Test view w/ flexible dtype
+        iterator = list(zip(np.arange(10), np.random.rand(10)))
+        data = np.array(iterator)
+        a = masked_array(iterator, dtype=[('a', float), ('b', float)])
+        a.mask[0] = (1, 0)
+        test = a.view((float, 2), np.matrix)
+        assert_equal(test, data)
+        assert_(isinstance(test, np.matrix))
+        assert_(not isinstance(test, MaskedArray))
+
+
+class TestSubclassing:
+    # Test suite for masked subclasses of ndarray.
+
+    def setup_method(self):
+        x = np.arange(5, dtype='float')
+        mx = MMatrix(x, mask=[0, 1, 0, 0, 0])
+        self.data = (x, mx)
+
+    def test_maskedarray_subclassing(self):
+        # Tests subclassing MaskedArray
+        (x, mx) = self.data
+        assert_(isinstance(mx._data, np.matrix))
+
+    def test_masked_unary_operations(self):
+        # Tests masked_unary_operation
+        (x, mx) = self.data
+        with np.errstate(divide='ignore'):
+            assert_(isinstance(log(mx), MMatrix))
+            assert_equal(log(x), np.log(x))
+
+    def test_masked_binary_operations(self):
+        # Tests masked_binary_operation
+        (x, mx) = self.data
+        # Result should be a MMatrix
+        assert_(isinstance(add(mx, mx), MMatrix))
+        assert_(isinstance(add(mx, x), MMatrix))
+        # Result should work
+        assert_equal(add(mx, x), mx+x)
+        assert_(isinstance(add(mx, mx)._data, np.matrix))
+        with assert_warns(DeprecationWarning):
+            assert_(isinstance(add.outer(mx, mx), MMatrix))
+        assert_(isinstance(hypot(mx, mx), MMatrix))
+        assert_(isinstance(hypot(mx, x), MMatrix))
+
+    def test_masked_binary_operations2(self):
+        # Tests domained_masked_binary_operation
+        (x, mx) = self.data
+        xmx = masked_array(mx.data.__array__(), mask=mx.mask)
+        assert_(isinstance(divide(mx, mx), MMatrix))
+        assert_(isinstance(divide(mx, x), MMatrix))
+        assert_equal(divide(mx, mx), divide(xmx, xmx))
+
+class TestConcatenator:
+    # Tests for mr_, the equivalent of r_ for masked arrays.
+
+    def test_matrix_builder(self):
+        assert_raises(np.ma.MAError, lambda: mr_['1, 2; 3, 4'])
+
+    def test_matrix(self):
+        # Test consistency with unmasked version.  If we ever deprecate
+        # matrix, this test should either still pass, or both actual and
+        # expected should fail to be build.
+        actual = mr_['r', 1, 2, 3]
+        expected = np.ma.array(np.r_['r', 1, 2, 3])
+        assert_array_equal(actual, expected)
+
+        # outer type is masked array, inner type is matrix
+        assert_equal(type(actual), type(expected))
+        assert_equal(type(actual.data), type(expected.data))

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_matrix_linalg.py

@@ -0,0 +1,93 @@
+""" Test functions for linalg module using the matrix class."""
+import numpy as np
+
+from numpy.linalg.tests.test_linalg import (
+    LinalgCase, apply_tag, TestQR as _TestQR, LinalgTestCase,
+    _TestNorm2D, _TestNormDoubleBase, _TestNormSingleBase, _TestNormInt64Base,
+    SolveCases, InvCases, EigvalsCases, EigCases, SVDCases, CondCases,
+    PinvCases, DetCases, LstsqCases)
+
+
+CASES = []
+
+# square test cases
+CASES += apply_tag('square', [
+    LinalgCase("0x0_matrix",
+               np.empty((0, 0), dtype=np.double).view(np.matrix),
+               np.empty((0, 1), dtype=np.double).view(np.matrix),
+               tags={'size-0'}),
+    LinalgCase("matrix_b_only",
+               np.array([[1., 2.], [3., 4.]]),
+               np.matrix([2., 1.]).T),
+    LinalgCase("matrix_a_and_b",
+               np.matrix([[1., 2.], [3., 4.]]),
+               np.matrix([2., 1.]).T),
+])
+
+# hermitian test-cases
+CASES += apply_tag('hermitian', [
+    LinalgCase("hmatrix_a_and_b",
+               np.matrix([[1., 2.], [2., 1.]]),
+               None),
+])
+# No need to make generalized or strided cases for matrices.
+
+
+class MatrixTestCase(LinalgTestCase):
+    TEST_CASES = CASES
+
+
+class TestSolveMatrix(SolveCases, MatrixTestCase):
+    pass
+
+
+class TestInvMatrix(InvCases, MatrixTestCase):
+    pass
+
+
+class TestEigvalsMatrix(EigvalsCases, MatrixTestCase):
+    pass
+
+
+class TestEigMatrix(EigCases, MatrixTestCase):
+    pass
+
+
+class TestSVDMatrix(SVDCases, MatrixTestCase):
+    pass
+
+
+class TestCondMatrix(CondCases, MatrixTestCase):
+    pass
+
+
+class TestPinvMatrix(PinvCases, MatrixTestCase):
+    pass
+
+
+class TestDetMatrix(DetCases, MatrixTestCase):
+    pass
+
+
+class TestLstsqMatrix(LstsqCases, MatrixTestCase):
+    pass
+
+
+class _TestNorm2DMatrix(_TestNorm2D):
+    array = np.matrix
+
+
+class TestNormDoubleMatrix(_TestNorm2DMatrix, _TestNormDoubleBase):
+    pass
+
+
+class TestNormSingleMatrix(_TestNorm2DMatrix, _TestNormSingleBase):
+    pass
+
+
+class TestNormInt64Matrix(_TestNorm2DMatrix, _TestNormInt64Base):
+    pass
+
+
+class TestQRMatrix(_TestQR):
+    array = np.matrix

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_multiarray.py

@@ -0,0 +1,16 @@
+import numpy as np
+from numpy.testing import assert_, assert_equal, assert_array_equal
+
+class TestView:
+    def test_type(self):
+        x = np.array([1, 2, 3])
+        assert_(isinstance(x.view(np.matrix), np.matrix))
+
+    def test_keywords(self):
+        x = np.array([(1, 2)], dtype=[('a', np.int8), ('b', np.int8)])
+        # We must be specific about the endianness here:
+        y = x.view(dtype='<i2', type=np.matrix)
+        assert_array_equal(y, [[513]])
+
+        assert_(isinstance(y, np.matrix))
+        assert_equal(y.dtype, np.dtype('<i2'))

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_numeric.py

@@ -0,0 +1,17 @@
+import numpy as np
+from numpy.testing import assert_equal
+
+class TestDot:
+    def test_matscalar(self):
+        b1 = np.matrix(np.ones((3, 3), dtype=complex))
+        assert_equal(b1*1.0, b1)
+
+
+def test_diagonal():
+    b1 = np.matrix([[1,2],[3,4]])
+    diag_b1 = np.matrix([[1, 4]])
+    array_b1 = np.array([1, 4])
+
+    assert_equal(b1.diagonal(), diag_b1)
+    assert_equal(np.diagonal(b1), array_b1)
+    assert_equal(np.diag(b1), array_b1)

.venv/lib/python3.11/site-packages/numpy/matrixlib/tests/test_regression.py

@@ -0,0 +1,31 @@
+import numpy as np
+from numpy.testing import assert_, assert_equal, assert_raises
+
+
+class TestRegression:
+    def test_kron_matrix(self):
+        # Ticket #71
+        x = np.matrix('[1 0; 1 0]')
+        assert_equal(type(np.kron(x, x)), type(x))
+
+    def test_matrix_properties(self):
+        # Ticket #125
+        a = np.matrix([1.0], dtype=float)
+        assert_(type(a.real) is np.matrix)
+        assert_(type(a.imag) is np.matrix)
+        c, d = np.matrix([0.0]).nonzero()
+        assert_(type(c) is np.ndarray)
+        assert_(type(d) is np.ndarray)
+
+    def test_matrix_multiply_by_1d_vector(self):
+        # Ticket #473
+        def mul():
+            np.mat(np.eye(2))*np.ones(2)
+
+        assert_raises(ValueError, mul)
+
+    def test_matrix_std_argmax(self):
+        # Ticket #83
+        x = np.asmatrix(np.random.uniform(0, 1, (3, 3)))
+        assert_equal(x.std().shape, ())
+        assert_equal(x.argmax().shape, ())

.venv/lib/python3.11/site-packages/numpy/polynomial/__init__.pyi

@@ -0,0 +1,22 @@
+from numpy._pytesttester import PytestTester
+
+from numpy.polynomial import (
+    chebyshev as chebyshev,
+    hermite as hermite,
+    hermite_e as hermite_e,
+    laguerre as laguerre,
+    legendre as legendre,
+    polynomial as polynomial,
+)
+from numpy.polynomial.chebyshev import Chebyshev as Chebyshev
+from numpy.polynomial.hermite import Hermite as Hermite
+from numpy.polynomial.hermite_e import HermiteE as HermiteE
+from numpy.polynomial.laguerre import Laguerre as Laguerre
+from numpy.polynomial.legendre import Legendre as Legendre
+from numpy.polynomial.polynomial import Polynomial as Polynomial
+
+__all__: list[str]
+__path__: list[str]
+test: PytestTester
+
+def set_default_printstyle(style): ...

.venv/lib/python3.11/site-packages/numpy/polynomial/chebyshev.pyi

@@ -0,0 +1,51 @@
+from typing import Any
+
+from numpy import ndarray, dtype, int_
+from numpy.polynomial._polybase import ABCPolyBase
+from numpy.polynomial.polyutils import trimcoef
+
+__all__: list[str]
+
+chebtrim = trimcoef
+
+def poly2cheb(pol): ...
+def cheb2poly(c): ...
+
+chebdomain: ndarray[Any, dtype[int_]]
+chebzero: ndarray[Any, dtype[int_]]
+chebone: ndarray[Any, dtype[int_]]
+chebx: ndarray[Any, dtype[int_]]
+
+def chebline(off, scl): ...
+def chebfromroots(roots): ...
+def chebadd(c1, c2): ...
+def chebsub(c1, c2): ...
+def chebmulx(c): ...
+def chebmul(c1, c2): ...
+def chebdiv(c1, c2): ...
+def chebpow(c, pow, maxpower=...): ...
+def chebder(c, m=..., scl=..., axis=...): ...
+def chebint(c, m=..., k = ..., lbnd=..., scl=..., axis=...): ...
+def chebval(x, c, tensor=...): ...
+def chebval2d(x, y, c): ...
+def chebgrid2d(x, y, c): ...
+def chebval3d(x, y, z, c): ...
+def chebgrid3d(x, y, z, c): ...
+def chebvander(x, deg): ...
+def chebvander2d(x, y, deg): ...
+def chebvander3d(x, y, z, deg): ...
+def chebfit(x, y, deg, rcond=..., full=..., w=...): ...
+def chebcompanion(c): ...
+def chebroots(c): ...
+def chebinterpolate(func, deg, args = ...): ...
+def chebgauss(deg): ...
+def chebweight(x): ...
+def chebpts1(npts): ...
+def chebpts2(npts): ...
+
+class Chebyshev(ABCPolyBase):
+    @classmethod
+    def interpolate(cls, func, deg, domain=..., args = ...): ...
+    domain: Any
+    window: Any
+    basis_name: Any

.venv/lib/python3.11/site-packages/numpy/polynomial/hermite.py

@@ -0,0 +1,1703 @@
+"""
+==============================================================
+Hermite Series, "Physicists" (:mod:`numpy.polynomial.hermite`)
+==============================================================
+
+This module provides a number of objects (mostly functions) useful for
+dealing with Hermite series, including a `Hermite` class that
+encapsulates the usual arithmetic operations.  (General information
+on how this module represents and works with such polynomials is in the
+docstring for its "parent" sub-package, `numpy.polynomial`).
+
+Classes
+-------
+.. autosummary::
+   :toctree: generated/
+
+   Hermite
+
+Constants
+---------
+.. autosummary::
+   :toctree: generated/
+
+   hermdomain
+   hermzero
+   hermone
+   hermx
+
+Arithmetic
+----------
+.. autosummary::
+   :toctree: generated/
+
+   hermadd
+   hermsub
+   hermmulx
+   hermmul
+   hermdiv
+   hermpow
+   hermval
+   hermval2d
+   hermval3d
+   hermgrid2d
+   hermgrid3d
+
+Calculus
+--------
+.. autosummary::
+   :toctree: generated/
+
+   hermder
+   hermint
+
+Misc Functions
+--------------
+.. autosummary::
+   :toctree: generated/
+
+   hermfromroots
+   hermroots
+   hermvander
+   hermvander2d
+   hermvander3d
+   hermgauss
+   hermweight
+   hermcompanion
+   hermfit
+   hermtrim
+   hermline
+   herm2poly
+   poly2herm
+
+See also
+--------
+`numpy.polynomial`
+
+"""
+import numpy as np
+import numpy.linalg as la
+from numpy.core.multiarray import normalize_axis_index
+
+from . import polyutils as pu
+from ._polybase import ABCPolyBase
+
+__all__ = [
+    'hermzero', 'hermone', 'hermx', 'hermdomain', 'hermline', 'hermadd',
+    'hermsub', 'hermmulx', 'hermmul', 'hermdiv', 'hermpow', 'hermval',
+    'hermder', 'hermint', 'herm2poly', 'poly2herm', 'hermfromroots',
+    'hermvander', 'hermfit', 'hermtrim', 'hermroots', 'Hermite',
+    'hermval2d', 'hermval3d', 'hermgrid2d', 'hermgrid3d', 'hermvander2d',
+    'hermvander3d', 'hermcompanion', 'hermgauss', 'hermweight']
+
+hermtrim = pu.trimcoef
+
+
+def poly2herm(pol):
+    """
+    poly2herm(pol)
+
+    Convert a polynomial to a Hermite series.
+
+    Convert an array representing the coefficients of a polynomial (relative
+    to the "standard" basis) ordered from lowest degree to highest, to an
+    array of the coefficients of the equivalent Hermite series, ordered
+    from lowest to highest degree.
+
+    Parameters
+    ----------
+    pol : array_like
+        1-D array containing the polynomial coefficients
+
+    Returns
+    -------
+    c : ndarray
+        1-D array containing the coefficients of the equivalent Hermite
+        series.
+
+    See Also
+    --------
+    herm2poly
+
+    Notes
+    -----
+    The easy way to do conversions between polynomial basis sets
+    is to use the convert method of a class instance.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import poly2herm
+    >>> poly2herm(np.arange(4))
+    array([1.   ,  2.75 ,  0.5  ,  0.375])
+
+    """
+    [pol] = pu.as_series([pol])
+    deg = len(pol) - 1
+    res = 0
+    for i in range(deg, -1, -1):
+        res = hermadd(hermmulx(res), pol[i])
+    return res
+
+
+def herm2poly(c):
+    """
+    Convert a Hermite series to a polynomial.
+
+    Convert an array representing the coefficients of a Hermite series,
+    ordered from lowest degree to highest, to an array of the coefficients
+    of the equivalent polynomial (relative to the "standard" basis) ordered
+    from lowest to highest degree.
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array containing the Hermite series coefficients, ordered
+        from lowest order term to highest.
+
+    Returns
+    -------
+    pol : ndarray
+        1-D array containing the coefficients of the equivalent polynomial
+        (relative to the "standard" basis) ordered from lowest order term
+        to highest.
+
+    See Also
+    --------
+    poly2herm
+
+    Notes
+    -----
+    The easy way to do conversions between polynomial basis sets
+    is to use the convert method of a class instance.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import herm2poly
+    >>> herm2poly([ 1.   ,  2.75 ,  0.5  ,  0.375])
+    array([0., 1., 2., 3.])
+
+    """
+    from .polynomial import polyadd, polysub, polymulx
+
+    [c] = pu.as_series([c])
+    n = len(c)
+    if n == 1:
+        return c
+    if n == 2:
+        c[1] *= 2
+        return c
+    else:
+        c0 = c[-2]
+        c1 = c[-1]
+        # i is the current degree of c1
+        for i in range(n - 1, 1, -1):
+            tmp = c0
+            c0 = polysub(c[i - 2], c1*(2*(i - 1)))
+            c1 = polyadd(tmp, polymulx(c1)*2)
+        return polyadd(c0, polymulx(c1)*2)
+
+#
+# These are constant arrays are of integer type so as to be compatible
+# with the widest range of other types, such as Decimal.
+#
+
+# Hermite
+hermdomain = np.array([-1, 1])
+
+# Hermite coefficients representing zero.
+hermzero = np.array([0])
+
+# Hermite coefficients representing one.
+hermone = np.array([1])
+
+# Hermite coefficients representing the identity x.
+hermx = np.array([0, 1/2])
+
+
+def hermline(off, scl):
+    """
+    Hermite series whose graph is a straight line.
+
+
+
+    Parameters
+    ----------
+    off, scl : scalars
+        The specified line is given by ``off + scl*x``.
+
+    Returns
+    -------
+    y : ndarray
+        This module's representation of the Hermite series for
+        ``off + scl*x``.
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyline
+    numpy.polynomial.chebyshev.chebline
+    numpy.polynomial.legendre.legline
+    numpy.polynomial.laguerre.lagline
+    numpy.polynomial.hermite_e.hermeline
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermline, hermval
+    >>> hermval(0,hermline(3, 2))
+    3.0
+    >>> hermval(1,hermline(3, 2))
+    5.0
+
+    """
+    if scl != 0:
+        return np.array([off, scl/2])
+    else:
+        return np.array([off])
+
+
+def hermfromroots(roots):
+    """
+    Generate a Hermite series with given roots.
+
+    The function returns the coefficients of the polynomial
+
+    .. math:: p(x) = (x - r_0) * (x - r_1) * ... * (x - r_n),
+
+    in Hermite form, where the `r_n` are the roots specified in `roots`.
+    If a zero has multiplicity n, then it must appear in `roots` n times.
+    For instance, if 2 is a root of multiplicity three and 3 is a root of
+    multiplicity 2, then `roots` looks something like [2, 2, 2, 3, 3]. The
+    roots can appear in any order.
+
+    If the returned coefficients are `c`, then
+
+    .. math:: p(x) = c_0 + c_1 * H_1(x) + ... +  c_n * H_n(x)
+
+    The coefficient of the last term is not generally 1 for monic
+    polynomials in Hermite form.
+
+    Parameters
+    ----------
+    roots : array_like
+        Sequence containing the roots.
+
+    Returns
+    -------
+    out : ndarray
+        1-D array of coefficients.  If all roots are real then `out` is a
+        real array, if some of the roots are complex, then `out` is complex
+        even if all the coefficients in the result are real (see Examples
+        below).
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyfromroots
+    numpy.polynomial.legendre.legfromroots
+    numpy.polynomial.laguerre.lagfromroots
+    numpy.polynomial.chebyshev.chebfromroots
+    numpy.polynomial.hermite_e.hermefromroots
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermfromroots, hermval
+    >>> coef = hermfromroots((-1, 0, 1))
+    >>> hermval((-1, 0, 1), coef)
+    array([0.,  0.,  0.])
+    >>> coef = hermfromroots((-1j, 1j))
+    >>> hermval((-1j, 1j), coef)
+    array([0.+0.j, 0.+0.j])
+
+    """
+    return pu._fromroots(hermline, hermmul, roots)
+
+
+def hermadd(c1, c2):
+    """
+    Add one Hermite series to another.
+
+    Returns the sum of two Hermite series `c1` + `c2`.  The arguments
+    are sequences of coefficients ordered from lowest order term to
+    highest, i.e., [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Array representing the Hermite series of their sum.
+
+    See Also
+    --------
+    hermsub, hermmulx, hermmul, hermdiv, hermpow
+
+    Notes
+    -----
+    Unlike multiplication, division, etc., the sum of two Hermite series
+    is a Hermite series (without having to "reproject" the result onto
+    the basis set) so addition, just like that of "standard" polynomials,
+    is simply "component-wise."
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermadd
+    >>> hermadd([1, 2, 3], [1, 2, 3, 4])
+    array([2., 4., 6., 4.])
+
+    """
+    return pu._add(c1, c2)
+
+
+def hermsub(c1, c2):
+    """
+    Subtract one Hermite series from another.
+
+    Returns the difference of two Hermite series `c1` - `c2`.  The
+    sequences of coefficients are from lowest order term to highest, i.e.,
+    [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Of Hermite series coefficients representing their difference.
+
+    See Also
+    --------
+    hermadd, hermmulx, hermmul, hermdiv, hermpow
+
+    Notes
+    -----
+    Unlike multiplication, division, etc., the difference of two Hermite
+    series is a Hermite series (without having to "reproject" the result
+    onto the basis set) so subtraction, just like that of "standard"
+    polynomials, is simply "component-wise."
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermsub
+    >>> hermsub([1, 2, 3, 4], [1, 2, 3])
+    array([0.,  0.,  0.,  4.])
+
+    """
+    return pu._sub(c1, c2)
+
+
+def hermmulx(c):
+    """Multiply a Hermite series by x.
+
+    Multiply the Hermite series `c` by x, where x is the independent
+    variable.
+
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Array representing the result of the multiplication.
+
+    See Also
+    --------
+    hermadd, hermsub, hermmul, hermdiv, hermpow
+
+    Notes
+    -----
+    The multiplication uses the recursion relationship for Hermite
+    polynomials in the form
+
+    .. math::
+
+        xP_i(x) = (P_{i + 1}(x)/2 + i*P_{i - 1}(x))
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermmulx
+    >>> hermmulx([1, 2, 3])
+    array([2. , 6.5, 1. , 1.5])
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    # The zero series needs special treatment
+    if len(c) == 1 and c[0] == 0:
+        return c
+
+    prd = np.empty(len(c) + 1, dtype=c.dtype)
+    prd[0] = c[0]*0
+    prd[1] = c[0]/2
+    for i in range(1, len(c)):
+        prd[i + 1] = c[i]/2
+        prd[i - 1] += c[i]*i
+    return prd
+
+
+def hermmul(c1, c2):
+    """
+    Multiply one Hermite series by another.
+
+    Returns the product of two Hermite series `c1` * `c2`.  The arguments
+    are sequences of coefficients, from lowest order "term" to highest,
+    e.g., [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Of Hermite series coefficients representing their product.
+
+    See Also
+    --------
+    hermadd, hermsub, hermmulx, hermdiv, hermpow
+
+    Notes
+    -----
+    In general, the (polynomial) product of two C-series results in terms
+    that are not in the Hermite polynomial basis set.  Thus, to express
+    the product as a Hermite series, it is necessary to "reproject" the
+    product onto said basis set, which may produce "unintuitive" (but
+    correct) results; see Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermmul
+    >>> hermmul([1, 2, 3], [0, 1, 2])
+    array([52.,  29.,  52.,   7.,   6.])
+
+    """
+    # s1, s2 are trimmed copies
+    [c1, c2] = pu.as_series([c1, c2])
+
+    if len(c1) > len(c2):
+        c = c2
+        xs = c1
+    else:
+        c = c1
+        xs = c2
+
+    if len(c) == 1:
+        c0 = c[0]*xs
+        c1 = 0
+    elif len(c) == 2:
+        c0 = c[0]*xs
+        c1 = c[1]*xs
+    else:
+        nd = len(c)
+        c0 = c[-2]*xs
+        c1 = c[-1]*xs
+        for i in range(3, len(c) + 1):
+            tmp = c0
+            nd = nd - 1
+            c0 = hermsub(c[-i]*xs, c1*(2*(nd - 1)))
+            c1 = hermadd(tmp, hermmulx(c1)*2)
+    return hermadd(c0, hermmulx(c1)*2)
+
+
+def hermdiv(c1, c2):
+    """
+    Divide one Hermite series by another.
+
+    Returns the quotient-with-remainder of two Hermite series
+    `c1` / `c2`.  The arguments are sequences of coefficients from lowest
+    order "term" to highest, e.g., [1,2,3] represents the series
+    ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    [quo, rem] : ndarrays
+        Of Hermite series coefficients representing the quotient and
+        remainder.
+
+    See Also
+    --------
+    hermadd, hermsub, hermmulx, hermmul, hermpow
+
+    Notes
+    -----
+    In general, the (polynomial) division of one Hermite series by another
+    results in quotient and remainder terms that are not in the Hermite
+    polynomial basis set.  Thus, to express these results as a Hermite
+    series, it is necessary to "reproject" the results onto the Hermite
+    basis set, which may produce "unintuitive" (but correct) results; see
+    Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermdiv
+    >>> hermdiv([ 52.,  29.,  52.,   7.,   6.], [0, 1, 2])
+    (array([1., 2., 3.]), array([0.]))
+    >>> hermdiv([ 54.,  31.,  52.,   7.,   6.], [0, 1, 2])
+    (array([1., 2., 3.]), array([2., 2.]))
+    >>> hermdiv([ 53.,  30.,  52.,   7.,   6.], [0, 1, 2])
+    (array([1., 2., 3.]), array([1., 1.]))
+
+    """
+    return pu._div(hermmul, c1, c2)
+
+
+def hermpow(c, pow, maxpower=16):
+    """Raise a Hermite series to a power.
+
+    Returns the Hermite series `c` raised to the power `pow`. The
+    argument `c` is a sequence of coefficients ordered from low to high.
+    i.e., [1,2,3] is the series  ``P_0 + 2*P_1 + 3*P_2.``
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of Hermite series coefficients ordered from low to
+        high.
+    pow : integer
+        Power to which the series will be raised
+    maxpower : integer, optional
+        Maximum power allowed. This is mainly to limit growth of the series
+        to unmanageable size. Default is 16
+
+    Returns
+    -------
+    coef : ndarray
+        Hermite series of power.
+
+    See Also
+    --------
+    hermadd, hermsub, hermmulx, hermmul, hermdiv
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermpow
+    >>> hermpow([1, 2, 3], 2)
+    array([81.,  52.,  82.,  12.,   9.])
+
+    """
+    return pu._pow(hermmul, c, pow, maxpower)
+
+
+def hermder(c, m=1, scl=1, axis=0):
+    """
+    Differentiate a Hermite series.
+
+    Returns the Hermite series coefficients `c` differentiated `m` times
+    along `axis`.  At each iteration the result is multiplied by `scl` (the
+    scaling factor is for use in a linear change of variable). The argument
+    `c` is an array of coefficients from low to high degree along each
+    axis, e.g., [1,2,3] represents the series ``1*H_0 + 2*H_1 + 3*H_2``
+    while [[1,2],[1,2]] represents ``1*H_0(x)*H_0(y) + 1*H_1(x)*H_0(y) +
+    2*H_0(x)*H_1(y) + 2*H_1(x)*H_1(y)`` if axis=0 is ``x`` and axis=1 is
+    ``y``.
+
+    Parameters
+    ----------
+    c : array_like
+        Array of Hermite series coefficients. If `c` is multidimensional the
+        different axis correspond to different variables with the degree in
+        each axis given by the corresponding index.
+    m : int, optional
+        Number of derivatives taken, must be non-negative. (Default: 1)
+    scl : scalar, optional
+        Each differentiation is multiplied by `scl`.  The end result is
+        multiplication by ``scl**m``.  This is for use in a linear change of
+        variable. (Default: 1)
+    axis : int, optional
+        Axis over which the derivative is taken. (Default: 0).
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    der : ndarray
+        Hermite series of the derivative.
+
+    See Also
+    --------
+    hermint
+
+    Notes
+    -----
+    In general, the result of differentiating a Hermite series does not
+    resemble the same operation on a power series. Thus the result of this
+    function may be "unintuitive," albeit correct; see Examples section
+    below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermder
+    >>> hermder([ 1. ,  0.5,  0.5,  0.5])
+    array([1., 2., 3.])
+    >>> hermder([-0.5,  1./2.,  1./8.,  1./12.,  1./16.], m=2)
+    array([1., 2., 3.])
+
+    """
+    c = np.array(c, ndmin=1, copy=True)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    cnt = pu._deprecate_as_int(m, "the order of derivation")
+    iaxis = pu._deprecate_as_int(axis, "the axis")
+    if cnt < 0:
+        raise ValueError("The order of derivation must be non-negative")
+    iaxis = normalize_axis_index(iaxis, c.ndim)
+
+    if cnt == 0:
+        return c
+
+    c = np.moveaxis(c, iaxis, 0)
+    n = len(c)
+    if cnt >= n:
+        c = c[:1]*0
+    else:
+        for i in range(cnt):
+            n = n - 1
+            c *= scl
+            der = np.empty((n,) + c.shape[1:], dtype=c.dtype)
+            for j in range(n, 0, -1):
+                der[j - 1] = (2*j)*c[j]
+            c = der
+    c = np.moveaxis(c, 0, iaxis)
+    return c
+
+
+def hermint(c, m=1, k=[], lbnd=0, scl=1, axis=0):
+    """
+    Integrate a Hermite series.
+
+    Returns the Hermite series coefficients `c` integrated `m` times from
+    `lbnd` along `axis`. At each iteration the resulting series is
+    **multiplied** by `scl` and an integration constant, `k`, is added.
+    The scaling factor is for use in a linear change of variable.  ("Buyer
+    beware": note that, depending on what one is doing, one may want `scl`
+    to be the reciprocal of what one might expect; for more information,
+    see the Notes section below.)  The argument `c` is an array of
+    coefficients from low to high degree along each axis, e.g., [1,2,3]
+    represents the series ``H_0 + 2*H_1 + 3*H_2`` while [[1,2],[1,2]]
+    represents ``1*H_0(x)*H_0(y) + 1*H_1(x)*H_0(y) + 2*H_0(x)*H_1(y) +
+    2*H_1(x)*H_1(y)`` if axis=0 is ``x`` and axis=1 is ``y``.
+
+    Parameters
+    ----------
+    c : array_like
+        Array of Hermite series coefficients. If c is multidimensional the
+        different axis correspond to different variables with the degree in
+        each axis given by the corresponding index.
+    m : int, optional
+        Order of integration, must be positive. (Default: 1)
+    k : {[], list, scalar}, optional
+        Integration constant(s).  The value of the first integral at
+        ``lbnd`` is the first value in the list, the value of the second
+        integral at ``lbnd`` is the second value, etc.  If ``k == []`` (the
+        default), all constants are set to zero.  If ``m == 1``, a single
+        scalar can be given instead of a list.
+    lbnd : scalar, optional
+        The lower bound of the integral. (Default: 0)
+    scl : scalar, optional
+        Following each integration the result is *multiplied* by `scl`
+        before the integration constant is added. (Default: 1)
+    axis : int, optional
+        Axis over which the integral is taken. (Default: 0).
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    S : ndarray
+        Hermite series coefficients of the integral.
+
+    Raises
+    ------
+    ValueError
+        If ``m < 0``, ``len(k) > m``, ``np.ndim(lbnd) != 0``, or
+        ``np.ndim(scl) != 0``.
+
+    See Also
+    --------
+    hermder
+
+    Notes
+    -----
+    Note that the result of each integration is *multiplied* by `scl`.
+    Why is this important to note?  Say one is making a linear change of
+    variable :math:`u = ax + b` in an integral relative to `x`.  Then
+    :math:`dx = du/a`, so one will need to set `scl` equal to
+    :math:`1/a` - perhaps not what one would have first thought.
+
+    Also note that, in general, the result of integrating a C-series needs
+    to be "reprojected" onto the C-series basis set.  Thus, typically,
+    the result of this function is "unintuitive," albeit correct; see
+    Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermint
+    >>> hermint([1,2,3]) # integrate once, value 0 at 0.
+    array([1. , 0.5, 0.5, 0.5])
+    >>> hermint([1,2,3], m=2) # integrate twice, value & deriv 0 at 0
+    array([-0.5       ,  0.5       ,  0.125     ,  0.08333333,  0.0625    ]) # may vary
+    >>> hermint([1,2,3], k=1) # integrate once, value 1 at 0.
+    array([2. , 0.5, 0.5, 0.5])
+    >>> hermint([1,2,3], lbnd=-1) # integrate once, value 0 at -1
+    array([-2. ,  0.5,  0.5,  0.5])
+    >>> hermint([1,2,3], m=2, k=[1,2], lbnd=-1)
+    array([ 1.66666667, -0.5       ,  0.125     ,  0.08333333,  0.0625    ]) # may vary
+
+    """
+    c = np.array(c, ndmin=1, copy=True)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    if not np.iterable(k):
+        k = [k]
+    cnt = pu._deprecate_as_int(m, "the order of integration")
+    iaxis = pu._deprecate_as_int(axis, "the axis")
+    if cnt < 0:
+        raise ValueError("The order of integration must be non-negative")
+    if len(k) > cnt:
+        raise ValueError("Too many integration constants")
+    if np.ndim(lbnd) != 0:
+        raise ValueError("lbnd must be a scalar.")
+    if np.ndim(scl) != 0:
+        raise ValueError("scl must be a scalar.")
+    iaxis = normalize_axis_index(iaxis, c.ndim)
+
+    if cnt == 0:
+        return c
+
+    c = np.moveaxis(c, iaxis, 0)
+    k = list(k) + [0]*(cnt - len(k))
+    for i in range(cnt):
+        n = len(c)
+        c *= scl
+        if n == 1 and np.all(c[0] == 0):
+            c[0] += k[i]
+        else:
+            tmp = np.empty((n + 1,) + c.shape[1:], dtype=c.dtype)
+            tmp[0] = c[0]*0
+            tmp[1] = c[0]/2
+            for j in range(1, n):
+                tmp[j + 1] = c[j]/(2*(j + 1))
+            tmp[0] += k[i] - hermval(lbnd, tmp)
+            c = tmp
+    c = np.moveaxis(c, 0, iaxis)
+    return c
+
+
+def hermval(x, c, tensor=True):
+    """
+    Evaluate an Hermite series at points x.
+
+    If `c` is of length `n + 1`, this function returns the value:
+
+    .. math:: p(x) = c_0 * H_0(x) + c_1 * H_1(x) + ... + c_n * H_n(x)
+
+    The parameter `x` is converted to an array only if it is a tuple or a
+    list, otherwise it is treated as a scalar. In either case, either `x`
+    or its elements must support multiplication and addition both with
+    themselves and with the elements of `c`.
+
+    If `c` is a 1-D array, then `p(x)` will have the same shape as `x`.  If
+    `c` is multidimensional, then the shape of the result depends on the
+    value of `tensor`. If `tensor` is true the shape will be c.shape[1:] +
+    x.shape. If `tensor` is false the shape will be c.shape[1:]. Note that
+    scalars have shape (,).
+
+    Trailing zeros in the coefficients will be used in the evaluation, so
+    they should be avoided if efficiency is a concern.
+
+    Parameters
+    ----------
+    x : array_like, compatible object
+        If `x` is a list or tuple, it is converted to an ndarray, otherwise
+        it is left unchanged and treated as a scalar. In either case, `x`
+        or its elements must support addition and multiplication with
+        themselves and with the elements of `c`.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree n are contained in c[n]. If `c` is multidimensional the
+        remaining indices enumerate multiple polynomials. In the two
+        dimensional case the coefficients may be thought of as stored in
+        the columns of `c`.
+    tensor : boolean, optional
+        If True, the shape of the coefficient array is extended with ones
+        on the right, one for each dimension of `x`. Scalars have dimension 0
+        for this action. The result is that every column of coefficients in
+        `c` is evaluated for every element of `x`. If False, `x` is broadcast
+        over the columns of `c` for the evaluation.  This keyword is useful
+        when `c` is multidimensional. The default value is True.
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    values : ndarray, algebra_like
+        The shape of the return value is described above.
+
+    See Also
+    --------
+    hermval2d, hermgrid2d, hermval3d, hermgrid3d
+
+    Notes
+    -----
+    The evaluation uses Clenshaw recursion, aka synthetic division.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermval
+    >>> coef = [1,2,3]
+    >>> hermval(1, coef)
+    11.0
+    >>> hermval([[1,2],[3,4]], coef)
+    array([[ 11.,   51.],
+           [115.,  203.]])
+
+    """
+    c = np.array(c, ndmin=1, copy=False)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    if isinstance(x, (tuple, list)):
+        x = np.asarray(x)
+    if isinstance(x, np.ndarray) and tensor:
+        c = c.reshape(c.shape + (1,)*x.ndim)
+
+    x2 = x*2
+    if len(c) == 1:
+        c0 = c[0]
+        c1 = 0
+    elif len(c) == 2:
+        c0 = c[0]
+        c1 = c[1]
+    else:
+        nd = len(c)
+        c0 = c[-2]
+        c1 = c[-1]
+        for i in range(3, len(c) + 1):
+            tmp = c0
+            nd = nd - 1
+            c0 = c[-i] - c1*(2*(nd - 1))
+            c1 = tmp + c1*x2
+    return c0 + c1*x2
+
+
+def hermval2d(x, y, c):
+    """
+    Evaluate a 2-D Hermite series at points (x, y).
+
+    This function returns the values:
+
+    .. math:: p(x,y) = \\sum_{i,j} c_{i,j} * H_i(x) * H_j(y)
+
+    The parameters `x` and `y` are converted to arrays only if they are
+    tuples or a lists, otherwise they are treated as a scalars and they
+    must have the same shape after conversion. In either case, either `x`
+    and `y` or their elements must support multiplication and addition both
+    with themselves and with the elements of `c`.
+
+    If `c` is a 1-D array a one is implicitly appended to its shape to make
+    it 2-D. The shape of the result will be c.shape[2:] + x.shape.
+
+    Parameters
+    ----------
+    x, y : array_like, compatible objects
+        The two dimensional series is evaluated at the points `(x, y)`,
+        where `x` and `y` must have the same shape. If `x` or `y` is a list
+        or tuple, it is first converted to an ndarray, otherwise it is left
+        unchanged and if it isn't an ndarray it is treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficient of the term
+        of multi-degree i,j is contained in ``c[i,j]``. If `c` has
+        dimension greater than two the remaining indices enumerate multiple
+        sets of coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points formed with
+        pairs of corresponding values from `x` and `y`.
+
+    See Also
+    --------
+    hermval, hermgrid2d, hermval3d, hermgrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._valnd(hermval, c, x, y)
+
+
+def hermgrid2d(x, y, c):
+    """
+    Evaluate a 2-D Hermite series on the Cartesian product of x and y.
+
+    This function returns the values:
+
+    .. math:: p(a,b) = \\sum_{i,j} c_{i,j} * H_i(a) * H_j(b)
+
+    where the points `(a, b)` consist of all pairs formed by taking
+    `a` from `x` and `b` from `y`. The resulting points form a grid with
+    `x` in the first dimension and `y` in the second.
+
+    The parameters `x` and `y` are converted to arrays only if they are
+    tuples or a lists, otherwise they are treated as a scalars. In either
+    case, either `x` and `y` or their elements must support multiplication
+    and addition both with themselves and with the elements of `c`.
+
+    If `c` has fewer than two dimensions, ones are implicitly appended to
+    its shape to make it 2-D. The shape of the result will be c.shape[2:] +
+    x.shape.
+
+    Parameters
+    ----------
+    x, y : array_like, compatible objects
+        The two dimensional series is evaluated at the points in the
+        Cartesian product of `x` and `y`.  If `x` or `y` is a list or
+        tuple, it is first converted to an ndarray, otherwise it is left
+        unchanged and, if it isn't an ndarray, it is treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree i,j are contained in ``c[i,j]``. If `c` has dimension
+        greater than two the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points in the Cartesian
+        product of `x` and `y`.
+
+    See Also
+    --------
+    hermval, hermval2d, hermval3d, hermgrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._gridnd(hermval, c, x, y)
+
+
+def hermval3d(x, y, z, c):
+    """
+    Evaluate a 3-D Hermite series at points (x, y, z).
+
+    This function returns the values:
+
+    .. math:: p(x,y,z) = \\sum_{i,j,k} c_{i,j,k} * H_i(x) * H_j(y) * H_k(z)
+
+    The parameters `x`, `y`, and `z` are converted to arrays only if
+    they are tuples or a lists, otherwise they are treated as a scalars and
+    they must have the same shape after conversion. In either case, either
+    `x`, `y`, and `z` or their elements must support multiplication and
+    addition both with themselves and with the elements of `c`.
+
+    If `c` has fewer than 3 dimensions, ones are implicitly appended to its
+    shape to make it 3-D. The shape of the result will be c.shape[3:] +
+    x.shape.
+
+    Parameters
+    ----------
+    x, y, z : array_like, compatible object
+        The three dimensional series is evaluated at the points
+        `(x, y, z)`, where `x`, `y`, and `z` must have the same shape.  If
+        any of `x`, `y`, or `z` is a list or tuple, it is first converted
+        to an ndarray, otherwise it is left unchanged and if it isn't an
+        ndarray it is  treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficient of the term of
+        multi-degree i,j,k is contained in ``c[i,j,k]``. If `c` has dimension
+        greater than 3 the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the multidimensional polynomial on points formed with
+        triples of corresponding values from `x`, `y`, and `z`.
+
+    See Also
+    --------
+    hermval, hermval2d, hermgrid2d, hermgrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._valnd(hermval, c, x, y, z)
+
+
+def hermgrid3d(x, y, z, c):
+    """
+    Evaluate a 3-D Hermite series on the Cartesian product of x, y, and z.
+
+    This function returns the values:
+
+    .. math:: p(a,b,c) = \\sum_{i,j,k} c_{i,j,k} * H_i(a) * H_j(b) * H_k(c)
+
+    where the points `(a, b, c)` consist of all triples formed by taking
+    `a` from `x`, `b` from `y`, and `c` from `z`. The resulting points form
+    a grid with `x` in the first dimension, `y` in the second, and `z` in
+    the third.
+
+    The parameters `x`, `y`, and `z` are converted to arrays only if they
+    are tuples or a lists, otherwise they are treated as a scalars. In
+    either case, either `x`, `y`, and `z` or their elements must support
+    multiplication and addition both with themselves and with the elements
+    of `c`.
+
+    If `c` has fewer than three dimensions, ones are implicitly appended to
+    its shape to make it 3-D. The shape of the result will be c.shape[3:] +
+    x.shape + y.shape + z.shape.
+
+    Parameters
+    ----------
+    x, y, z : array_like, compatible objects
+        The three dimensional series is evaluated at the points in the
+        Cartesian product of `x`, `y`, and `z`.  If `x`,`y`, or `z` is a
+        list or tuple, it is first converted to an ndarray, otherwise it is
+        left unchanged and, if it isn't an ndarray, it is treated as a
+        scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree i,j are contained in ``c[i,j]``. If `c` has dimension
+        greater than two the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points in the Cartesian
+        product of `x` and `y`.
+
+    See Also
+    --------
+    hermval, hermval2d, hermgrid2d, hermval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._gridnd(hermval, c, x, y, z)
+
+
+def hermvander(x, deg):
+    """Pseudo-Vandermonde matrix of given degree.
+
+    Returns the pseudo-Vandermonde matrix of degree `deg` and sample points
+    `x`. The pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., i] = H_i(x),
+
+    where `0 <= i <= deg`. The leading indices of `V` index the elements of
+    `x` and the last index is the degree of the Hermite polynomial.
+
+    If `c` is a 1-D array of coefficients of length `n + 1` and `V` is the
+    array ``V = hermvander(x, n)``, then ``np.dot(V, c)`` and
+    ``hermval(x, c)`` are the same up to roundoff. This equivalence is
+    useful both for least squares fitting and for the evaluation of a large
+    number of Hermite series of the same degree and sample points.
+
+    Parameters
+    ----------
+    x : array_like
+        Array of points. The dtype is converted to float64 or complex128
+        depending on whether any of the elements are complex. If `x` is
+        scalar it is converted to a 1-D array.
+    deg : int
+        Degree of the resulting matrix.
+
+    Returns
+    -------
+    vander : ndarray
+        The pseudo-Vandermonde matrix. The shape of the returned matrix is
+        ``x.shape + (deg + 1,)``, where The last index is the degree of the
+        corresponding Hermite polynomial.  The dtype will be the same as
+        the converted `x`.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermvander
+    >>> x = np.array([-1, 0, 1])
+    >>> hermvander(x, 3)
+    array([[ 1., -2.,  2.,  4.],
+           [ 1.,  0., -2., -0.],
+           [ 1.,  2.,  2., -4.]])
+
+    """
+    ideg = pu._deprecate_as_int(deg, "deg")
+    if ideg < 0:
+        raise ValueError("deg must be non-negative")
+
+    x = np.array(x, copy=False, ndmin=1) + 0.0
+    dims = (ideg + 1,) + x.shape
+    dtyp = x.dtype
+    v = np.empty(dims, dtype=dtyp)
+    v[0] = x*0 + 1
+    if ideg > 0:
+        x2 = x*2
+        v[1] = x2
+        for i in range(2, ideg + 1):
+            v[i] = (v[i-1]*x2 - v[i-2]*(2*(i - 1)))
+    return np.moveaxis(v, 0, -1)
+
+
+def hermvander2d(x, y, deg):
+    """Pseudo-Vandermonde matrix of given degrees.
+
+    Returns the pseudo-Vandermonde matrix of degrees `deg` and sample
+    points `(x, y)`. The pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., (deg[1] + 1)*i + j] = H_i(x) * H_j(y),
+
+    where `0 <= i <= deg[0]` and `0 <= j <= deg[1]`. The leading indices of
+    `V` index the points `(x, y)` and the last index encodes the degrees of
+    the Hermite polynomials.
+
+    If ``V = hermvander2d(x, y, [xdeg, ydeg])``, then the columns of `V`
+    correspond to the elements of a 2-D coefficient array `c` of shape
+    (xdeg + 1, ydeg + 1) in the order
+
+    .. math:: c_{00}, c_{01}, c_{02} ... , c_{10}, c_{11}, c_{12} ...
+
+    and ``np.dot(V, c.flat)`` and ``hermval2d(x, y, c)`` will be the same
+    up to roundoff. This equivalence is useful both for least squares
+    fitting and for the evaluation of a large number of 2-D Hermite
+    series of the same degrees and sample points.
+
+    Parameters
+    ----------
+    x, y : array_like
+        Arrays of point coordinates, all of the same shape. The dtypes
+        will be converted to either float64 or complex128 depending on
+        whether any of the elements are complex. Scalars are converted to 1-D
+        arrays.
+    deg : list of ints
+        List of maximum degrees of the form [x_deg, y_deg].
+
+    Returns
+    -------
+    vander2d : ndarray
+        The shape of the returned matrix is ``x.shape + (order,)``, where
+        :math:`order = (deg[0]+1)*(deg[1]+1)`.  The dtype will be the same
+        as the converted `x` and `y`.
+
+    See Also
+    --------
+    hermvander, hermvander3d, hermval2d, hermval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._vander_nd_flat((hermvander, hermvander), (x, y), deg)
+
+
+def hermvander3d(x, y, z, deg):
+    """Pseudo-Vandermonde matrix of given degrees.
+
+    Returns the pseudo-Vandermonde matrix of degrees `deg` and sample
+    points `(x, y, z)`. If `l, m, n` are the given degrees in `x, y, z`,
+    then The pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., (m+1)(n+1)i + (n+1)j + k] = H_i(x)*H_j(y)*H_k(z),
+
+    where `0 <= i <= l`, `0 <= j <= m`, and `0 <= j <= n`.  The leading
+    indices of `V` index the points `(x, y, z)` and the last index encodes
+    the degrees of the Hermite polynomials.
+
+    If ``V = hermvander3d(x, y, z, [xdeg, ydeg, zdeg])``, then the columns
+    of `V` correspond to the elements of a 3-D coefficient array `c` of
+    shape (xdeg + 1, ydeg + 1, zdeg + 1) in the order
+
+    .. math:: c_{000}, c_{001}, c_{002},... , c_{010}, c_{011}, c_{012},...
+
+    and  ``np.dot(V, c.flat)`` and ``hermval3d(x, y, z, c)`` will be the
+    same up to roundoff. This equivalence is useful both for least squares
+    fitting and for the evaluation of a large number of 3-D Hermite
+    series of the same degrees and sample points.
+
+    Parameters
+    ----------
+    x, y, z : array_like
+        Arrays of point coordinates, all of the same shape. The dtypes will
+        be converted to either float64 or complex128 depending on whether
+        any of the elements are complex. Scalars are converted to 1-D
+        arrays.
+    deg : list of ints
+        List of maximum degrees of the form [x_deg, y_deg, z_deg].
+
+    Returns
+    -------
+    vander3d : ndarray
+        The shape of the returned matrix is ``x.shape + (order,)``, where
+        :math:`order = (deg[0]+1)*(deg[1]+1)*(deg[2]+1)`.  The dtype will
+        be the same as the converted `x`, `y`, and `z`.
+
+    See Also
+    --------
+    hermvander, hermvander3d, hermval2d, hermval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._vander_nd_flat((hermvander, hermvander, hermvander), (x, y, z), deg)
+
+
+def hermfit(x, y, deg, rcond=None, full=False, w=None):
+    """
+    Least squares fit of Hermite series to data.
+
+    Return the coefficients of a Hermite series of degree `deg` that is the
+    least squares fit to the data values `y` given at points `x`. If `y` is
+    1-D the returned coefficients will also be 1-D. If `y` is 2-D multiple
+    fits are done, one for each column of `y`, and the resulting
+    coefficients are stored in the corresponding columns of a 2-D return.
+    The fitted polynomial(s) are in the form
+
+    .. math::  p(x) = c_0 + c_1 * H_1(x) + ... + c_n * H_n(x),
+
+    where `n` is `deg`.
+
+    Parameters
+    ----------
+    x : array_like, shape (M,)
+        x-coordinates of the M sample points ``(x[i], y[i])``.
+    y : array_like, shape (M,) or (M, K)
+        y-coordinates of the sample points. Several data sets of sample
+        points sharing the same x-coordinates can be fitted at once by
+        passing in a 2D-array that contains one dataset per column.
+    deg : int or 1-D array_like
+        Degree(s) of the fitting polynomials. If `deg` is a single integer
+        all terms up to and including the `deg`'th term are included in the
+        fit. For NumPy versions >= 1.11.0 a list of integers specifying the
+        degrees of the terms to include may be used instead.
+    rcond : float, optional
+        Relative condition number of the fit. Singular values smaller than
+        this relative to the largest singular value will be ignored. The
+        default value is len(x)*eps, where eps is the relative precision of
+        the float type, about 2e-16 in most cases.
+    full : bool, optional
+        Switch determining nature of return value. When it is False (the
+        default) just the coefficients are returned, when True diagnostic
+        information from the singular value decomposition is also returned.
+    w : array_like, shape (`M`,), optional
+        Weights. If not None, the weight ``w[i]`` applies to the unsquared
+        residual ``y[i] - y_hat[i]`` at ``x[i]``. Ideally the weights are
+        chosen so that the errors of the products ``w[i]*y[i]`` all have the
+        same variance.  When using inverse-variance weighting, use
+        ``w[i] = 1/sigma(y[i])``.  The default value is None.
+
+    Returns
+    -------
+    coef : ndarray, shape (M,) or (M, K)
+        Hermite coefficients ordered from low to high. If `y` was 2-D,
+        the coefficients for the data in column k  of `y` are in column
+        `k`.
+
+    [residuals, rank, singular_values, rcond] : list
+        These values are only returned if ``full == True``
+
+        - residuals -- sum of squared residuals of the least squares fit
+        - rank -- the numerical rank of the scaled Vandermonde matrix
+        - singular_values -- singular values of the scaled Vandermonde matrix
+        - rcond -- value of `rcond`.
+
+        For more details, see `numpy.linalg.lstsq`.
+
+    Warns
+    -----
+    RankWarning
+        The rank of the coefficient matrix in the least-squares fit is
+        deficient. The warning is only raised if ``full == False``.  The
+        warnings can be turned off by
+
+        >>> import warnings
+        >>> warnings.simplefilter('ignore', np.RankWarning)
+
+    See Also
+    --------
+    numpy.polynomial.chebyshev.chebfit
+    numpy.polynomial.legendre.legfit
+    numpy.polynomial.laguerre.lagfit
+    numpy.polynomial.polynomial.polyfit
+    numpy.polynomial.hermite_e.hermefit
+    hermval : Evaluates a Hermite series.
+    hermvander : Vandermonde matrix of Hermite series.
+    hermweight : Hermite weight function
+    numpy.linalg.lstsq : Computes a least-squares fit from the matrix.
+    scipy.interpolate.UnivariateSpline : Computes spline fits.
+
+    Notes
+    -----
+    The solution is the coefficients of the Hermite series `p` that
+    minimizes the sum of the weighted squared errors
+
+    .. math:: E = \\sum_j w_j^2 * |y_j - p(x_j)|^2,
+
+    where the :math:`w_j` are the weights. This problem is solved by
+    setting up the (typically) overdetermined matrix equation
+
+    .. math:: V(x) * c = w * y,
+
+    where `V` is the weighted pseudo Vandermonde matrix of `x`, `c` are the
+    coefficients to be solved for, `w` are the weights, `y` are the
+    observed values.  This equation is then solved using the singular value
+    decomposition of `V`.
+
+    If some of the singular values of `V` are so small that they are
+    neglected, then a `RankWarning` will be issued. This means that the
+    coefficient values may be poorly determined. Using a lower order fit
+    will usually get rid of the warning.  The `rcond` parameter can also be
+    set to a value smaller than its default, but the resulting fit may be
+    spurious and have large contributions from roundoff error.
+
+    Fits using Hermite series are probably most useful when the data can be
+    approximated by ``sqrt(w(x)) * p(x)``, where `w(x)` is the Hermite
+    weight. In that case the weight ``sqrt(w(x[i]))`` should be used
+    together with data values ``y[i]/sqrt(w(x[i]))``. The weight function is
+    available as `hermweight`.
+
+    References
+    ----------
+    .. [1] Wikipedia, "Curve fitting",
+           https://en.wikipedia.org/wiki/Curve_fitting
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermfit, hermval
+    >>> x = np.linspace(-10, 10)
+    >>> err = np.random.randn(len(x))/10
+    >>> y = hermval(x, [1, 2, 3]) + err
+    >>> hermfit(x, y, 2)
+    array([1.0218, 1.9986, 2.9999]) # may vary
+
+    """
+    return pu._fit(hermvander, x, y, deg, rcond, full, w)
+
+
+def hermcompanion(c):
+    """Return the scaled companion matrix of c.
+
+    The basis polynomials are scaled so that the companion matrix is
+    symmetric when `c` is an Hermite basis polynomial. This provides
+    better eigenvalue estimates than the unscaled case and for basis
+    polynomials the eigenvalues are guaranteed to be real if
+    `numpy.linalg.eigvalsh` is used to obtain them.
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of Hermite series coefficients ordered from low to high
+        degree.
+
+    Returns
+    -------
+    mat : ndarray
+        Scaled companion matrix of dimensions (deg, deg).
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    if len(c) < 2:
+        raise ValueError('Series must have maximum degree of at least 1.')
+    if len(c) == 2:
+        return np.array([[-.5*c[0]/c[1]]])
+
+    n = len(c) - 1
+    mat = np.zeros((n, n), dtype=c.dtype)
+    scl = np.hstack((1., 1./np.sqrt(2.*np.arange(n - 1, 0, -1))))
+    scl = np.multiply.accumulate(scl)[::-1]
+    top = mat.reshape(-1)[1::n+1]
+    bot = mat.reshape(-1)[n::n+1]
+    top[...] = np.sqrt(.5*np.arange(1, n))
+    bot[...] = top
+    mat[:, -1] -= scl*c[:-1]/(2.0*c[-1])
+    return mat
+
+
+def hermroots(c):
+    """
+    Compute the roots of a Hermite series.
+
+    Return the roots (a.k.a. "zeros") of the polynomial
+
+    .. math:: p(x) = \\sum_i c[i] * H_i(x).
+
+    Parameters
+    ----------
+    c : 1-D array_like
+        1-D array of coefficients.
+
+    Returns
+    -------
+    out : ndarray
+        Array of the roots of the series. If all the roots are real,
+        then `out` is also real, otherwise it is complex.
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyroots
+    numpy.polynomial.legendre.legroots
+    numpy.polynomial.laguerre.lagroots
+    numpy.polynomial.chebyshev.chebroots
+    numpy.polynomial.hermite_e.hermeroots
+
+    Notes
+    -----
+    The root estimates are obtained as the eigenvalues of the companion
+    matrix, Roots far from the origin of the complex plane may have large
+    errors due to the numerical instability of the series for such
+    values. Roots with multiplicity greater than 1 will also show larger
+    errors as the value of the series near such points is relatively
+    insensitive to errors in the roots. Isolated roots near the origin can
+    be improved by a few iterations of Newton's method.
+
+    The Hermite series basis polynomials aren't powers of `x` so the
+    results of this function may seem unintuitive.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite import hermroots, hermfromroots
+    >>> coef = hermfromroots([-1, 0, 1])
+    >>> coef
+    array([0.   ,  0.25 ,  0.   ,  0.125])
+    >>> hermroots(coef)
+    array([-1.00000000e+00, -1.38777878e-17,  1.00000000e+00])
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    if len(c) <= 1:
+        return np.array([], dtype=c.dtype)
+    if len(c) == 2:
+        return np.array([-.5*c[0]/c[1]])
+
+    # rotated companion matrix reduces error
+    m = hermcompanion(c)[::-1,::-1]
+    r = la.eigvals(m)
+    r.sort()
+    return r
+
+
+def _normed_hermite_n(x, n):
+    """
+    Evaluate a normalized Hermite polynomial.
+
+    Compute the value of the normalized Hermite polynomial of degree ``n``
+    at the points ``x``.
+
+
+    Parameters
+    ----------
+    x : ndarray of double.
+        Points at which to evaluate the function
+    n : int
+        Degree of the normalized Hermite function to be evaluated.
+
+    Returns
+    -------
+    values : ndarray
+        The shape of the return value is described above.
+
+    Notes
+    -----
+    .. versionadded:: 1.10.0
+
+    This function is needed for finding the Gauss points and integration
+    weights for high degrees. The values of the standard Hermite functions
+    overflow when n >= 207.
+
+    """
+    if n == 0:
+        return np.full(x.shape, 1/np.sqrt(np.sqrt(np.pi)))
+
+    c0 = 0.
+    c1 = 1./np.sqrt(np.sqrt(np.pi))
+    nd = float(n)
+    for i in range(n - 1):
+        tmp = c0
+        c0 = -c1*np.sqrt((nd - 1.)/nd)
+        c1 = tmp + c1*x*np.sqrt(2./nd)
+        nd = nd - 1.0
+    return c0 + c1*x*np.sqrt(2)
+
+
+def hermgauss(deg):
+    """
+    Gauss-Hermite quadrature.
+
+    Computes the sample points and weights for Gauss-Hermite quadrature.
+    These sample points and weights will correctly integrate polynomials of
+    degree :math:`2*deg - 1` or less over the interval :math:`[-\\inf, \\inf]`
+    with the weight function :math:`f(x) = \\exp(-x^2)`.
+
+    Parameters
+    ----------
+    deg : int
+        Number of sample points and weights. It must be >= 1.
+
+    Returns
+    -------
+    x : ndarray
+        1-D ndarray containing the sample points.
+    y : ndarray
+        1-D ndarray containing the weights.
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    The results have only been tested up to degree 100, higher degrees may
+    be problematic. The weights are determined by using the fact that
+
+    .. math:: w_k = c / (H'_n(x_k) * H_{n-1}(x_k))
+
+    where :math:`c` is a constant independent of :math:`k` and :math:`x_k`
+    is the k'th root of :math:`H_n`, and then scaling the results to get
+    the right value when integrating 1.
+
+    """
+    ideg = pu._deprecate_as_int(deg, "deg")
+    if ideg <= 0:
+        raise ValueError("deg must be a positive integer")
+
+    # first approximation of roots. We use the fact that the companion
+    # matrix is symmetric in this case in order to obtain better zeros.
+    c = np.array([0]*deg + [1], dtype=np.float64)
+    m = hermcompanion(c)
+    x = la.eigvalsh(m)
+
+    # improve roots by one application of Newton
+    dy = _normed_hermite_n(x, ideg)
+    df = _normed_hermite_n(x, ideg - 1) * np.sqrt(2*ideg)
+    x -= dy/df
+
+    # compute the weights. We scale the factor to avoid possible numerical
+    # overflow.
+    fm = _normed_hermite_n(x, ideg - 1)
+    fm /= np.abs(fm).max()
+    w = 1/(fm * fm)
+
+    # for Hermite we can also symmetrize
+    w = (w + w[::-1])/2
+    x = (x - x[::-1])/2
+
+    # scale w to get the right value
+    w *= np.sqrt(np.pi) / w.sum()
+
+    return x, w
+
+
+def hermweight(x):
+    """
+    Weight function of the Hermite polynomials.
+
+    The weight function is :math:`\\exp(-x^2)` and the interval of
+    integration is :math:`[-\\inf, \\inf]`. the Hermite polynomials are
+    orthogonal, but not normalized, with respect to this weight function.
+
+    Parameters
+    ----------
+    x : array_like
+       Values at which the weight function will be computed.
+
+    Returns
+    -------
+    w : ndarray
+       The weight function at `x`.
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    w = np.exp(-x**2)
+    return w
+
+
+#
+# Hermite series class
+#
+
+class Hermite(ABCPolyBase):
+    """An Hermite series class.
+
+    The Hermite class provides the standard Python numerical methods
+    '+', '-', '*', '//', '%', 'divmod', '**', and '()' as well as the
+    attributes and methods listed in the `ABCPolyBase` documentation.
+
+    Parameters
+    ----------
+    coef : array_like
+        Hermite coefficients in order of increasing degree, i.e,
+        ``(1, 2, 3)`` gives ``1*H_0(x) + 2*H_1(X) + 3*H_2(x)``.
+    domain : (2,) array_like, optional
+        Domain to use. The interval ``[domain[0], domain[1]]`` is mapped
+        to the interval ``[window[0], window[1]]`` by shifting and scaling.
+        The default value is [-1, 1].
+    window : (2,) array_like, optional
+        Window, see `domain` for its use. The default value is [-1, 1].
+
+        .. versionadded:: 1.6.0
+    symbol : str, optional
+        Symbol used to represent the independent variable in string
+        representations of the polynomial expression, e.g. for printing.
+        The symbol must be a valid Python identifier. Default value is 'x'.
+
+        .. versionadded:: 1.24
+
+    """
+    # Virtual Functions
+    _add = staticmethod(hermadd)
+    _sub = staticmethod(hermsub)
+    _mul = staticmethod(hermmul)
+    _div = staticmethod(hermdiv)
+    _pow = staticmethod(hermpow)
+    _val = staticmethod(hermval)
+    _int = staticmethod(hermint)
+    _der = staticmethod(hermder)
+    _fit = staticmethod(hermfit)
+    _line = staticmethod(hermline)
+    _roots = staticmethod(hermroots)
+    _fromroots = staticmethod(hermfromroots)
+
+    # Virtual properties
+    domain = np.array(hermdomain)
+    window = np.array(hermdomain)
+    basis_name = 'H'

.venv/lib/python3.11/site-packages/numpy/polynomial/hermite_e.py

@@ -0,0 +1,1695 @@
+"""
+===================================================================
+HermiteE Series, "Probabilists" (:mod:`numpy.polynomial.hermite_e`)
+===================================================================
+
+This module provides a number of objects (mostly functions) useful for
+dealing with Hermite_e series, including a `HermiteE` class that
+encapsulates the usual arithmetic operations.  (General information
+on how this module represents and works with such polynomials is in the
+docstring for its "parent" sub-package, `numpy.polynomial`).
+
+Classes
+-------
+.. autosummary::
+   :toctree: generated/
+
+   HermiteE
+
+Constants
+---------
+.. autosummary::
+   :toctree: generated/
+
+   hermedomain
+   hermezero
+   hermeone
+   hermex
+
+Arithmetic
+----------
+.. autosummary::
+   :toctree: generated/
+
+   hermeadd
+   hermesub
+   hermemulx
+   hermemul
+   hermediv
+   hermepow
+   hermeval
+   hermeval2d
+   hermeval3d
+   hermegrid2d
+   hermegrid3d
+
+Calculus
+--------
+.. autosummary::
+   :toctree: generated/
+
+   hermeder
+   hermeint
+
+Misc Functions
+--------------
+.. autosummary::
+   :toctree: generated/
+
+   hermefromroots
+   hermeroots
+   hermevander
+   hermevander2d
+   hermevander3d
+   hermegauss
+   hermeweight
+   hermecompanion
+   hermefit
+   hermetrim
+   hermeline
+   herme2poly
+   poly2herme
+
+See also
+--------
+`numpy.polynomial`
+
+"""
+import numpy as np
+import numpy.linalg as la
+from numpy.core.multiarray import normalize_axis_index
+
+from . import polyutils as pu
+from ._polybase import ABCPolyBase
+
+__all__ = [
+    'hermezero', 'hermeone', 'hermex', 'hermedomain', 'hermeline',
+    'hermeadd', 'hermesub', 'hermemulx', 'hermemul', 'hermediv',
+    'hermepow', 'hermeval', 'hermeder', 'hermeint', 'herme2poly',
+    'poly2herme', 'hermefromroots', 'hermevander', 'hermefit', 'hermetrim',
+    'hermeroots', 'HermiteE', 'hermeval2d', 'hermeval3d', 'hermegrid2d',
+    'hermegrid3d', 'hermevander2d', 'hermevander3d', 'hermecompanion',
+    'hermegauss', 'hermeweight']
+
+hermetrim = pu.trimcoef
+
+
+def poly2herme(pol):
+    """
+    poly2herme(pol)
+
+    Convert a polynomial to a Hermite series.
+
+    Convert an array representing the coefficients of a polynomial (relative
+    to the "standard" basis) ordered from lowest degree to highest, to an
+    array of the coefficients of the equivalent Hermite series, ordered
+    from lowest to highest degree.
+
+    Parameters
+    ----------
+    pol : array_like
+        1-D array containing the polynomial coefficients
+
+    Returns
+    -------
+    c : ndarray
+        1-D array containing the coefficients of the equivalent Hermite
+        series.
+
+    See Also
+    --------
+    herme2poly
+
+    Notes
+    -----
+    The easy way to do conversions between polynomial basis sets
+    is to use the convert method of a class instance.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import poly2herme
+    >>> poly2herme(np.arange(4))
+    array([  2.,  10.,   2.,   3.])
+
+    """
+    [pol] = pu.as_series([pol])
+    deg = len(pol) - 1
+    res = 0
+    for i in range(deg, -1, -1):
+        res = hermeadd(hermemulx(res), pol[i])
+    return res
+
+
+def herme2poly(c):
+    """
+    Convert a Hermite series to a polynomial.
+
+    Convert an array representing the coefficients of a Hermite series,
+    ordered from lowest degree to highest, to an array of the coefficients
+    of the equivalent polynomial (relative to the "standard" basis) ordered
+    from lowest to highest degree.
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array containing the Hermite series coefficients, ordered
+        from lowest order term to highest.
+
+    Returns
+    -------
+    pol : ndarray
+        1-D array containing the coefficients of the equivalent polynomial
+        (relative to the "standard" basis) ordered from lowest order term
+        to highest.
+
+    See Also
+    --------
+    poly2herme
+
+    Notes
+    -----
+    The easy way to do conversions between polynomial basis sets
+    is to use the convert method of a class instance.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import herme2poly
+    >>> herme2poly([  2.,  10.,   2.,   3.])
+    array([0.,  1.,  2.,  3.])
+
+    """
+    from .polynomial import polyadd, polysub, polymulx
+
+    [c] = pu.as_series([c])
+    n = len(c)
+    if n == 1:
+        return c
+    if n == 2:
+        return c
+    else:
+        c0 = c[-2]
+        c1 = c[-1]
+        # i is the current degree of c1
+        for i in range(n - 1, 1, -1):
+            tmp = c0
+            c0 = polysub(c[i - 2], c1*(i - 1))
+            c1 = polyadd(tmp, polymulx(c1))
+        return polyadd(c0, polymulx(c1))
+
+#
+# These are constant arrays are of integer type so as to be compatible
+# with the widest range of other types, such as Decimal.
+#
+
+# Hermite
+hermedomain = np.array([-1, 1])
+
+# Hermite coefficients representing zero.
+hermezero = np.array([0])
+
+# Hermite coefficients representing one.
+hermeone = np.array([1])
+
+# Hermite coefficients representing the identity x.
+hermex = np.array([0, 1])
+
+
+def hermeline(off, scl):
+    """
+    Hermite series whose graph is a straight line.
+
+    Parameters
+    ----------
+    off, scl : scalars
+        The specified line is given by ``off + scl*x``.
+
+    Returns
+    -------
+    y : ndarray
+        This module's representation of the Hermite series for
+        ``off + scl*x``.
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyline
+    numpy.polynomial.chebyshev.chebline
+    numpy.polynomial.legendre.legline
+    numpy.polynomial.laguerre.lagline
+    numpy.polynomial.hermite.hermline
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeline
+    >>> from numpy.polynomial.hermite_e import hermeline, hermeval
+    >>> hermeval(0,hermeline(3, 2))
+    3.0
+    >>> hermeval(1,hermeline(3, 2))
+    5.0
+
+    """
+    if scl != 0:
+        return np.array([off, scl])
+    else:
+        return np.array([off])
+
+
+def hermefromroots(roots):
+    """
+    Generate a HermiteE series with given roots.
+
+    The function returns the coefficients of the polynomial
+
+    .. math:: p(x) = (x - r_0) * (x - r_1) * ... * (x - r_n),
+
+    in HermiteE form, where the `r_n` are the roots specified in `roots`.
+    If a zero has multiplicity n, then it must appear in `roots` n times.
+    For instance, if 2 is a root of multiplicity three and 3 is a root of
+    multiplicity 2, then `roots` looks something like [2, 2, 2, 3, 3]. The
+    roots can appear in any order.
+
+    If the returned coefficients are `c`, then
+
+    .. math:: p(x) = c_0 + c_1 * He_1(x) + ... +  c_n * He_n(x)
+
+    The coefficient of the last term is not generally 1 for monic
+    polynomials in HermiteE form.
+
+    Parameters
+    ----------
+    roots : array_like
+        Sequence containing the roots.
+
+    Returns
+    -------
+    out : ndarray
+        1-D array of coefficients.  If all roots are real then `out` is a
+        real array, if some of the roots are complex, then `out` is complex
+        even if all the coefficients in the result are real (see Examples
+        below).
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyfromroots
+    numpy.polynomial.legendre.legfromroots
+    numpy.polynomial.laguerre.lagfromroots
+    numpy.polynomial.hermite.hermfromroots
+    numpy.polynomial.chebyshev.chebfromroots
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermefromroots, hermeval
+    >>> coef = hermefromroots((-1, 0, 1))
+    >>> hermeval((-1, 0, 1), coef)
+    array([0., 0., 0.])
+    >>> coef = hermefromroots((-1j, 1j))
+    >>> hermeval((-1j, 1j), coef)
+    array([0.+0.j, 0.+0.j])
+
+    """
+    return pu._fromroots(hermeline, hermemul, roots)
+
+
+def hermeadd(c1, c2):
+    """
+    Add one Hermite series to another.
+
+    Returns the sum of two Hermite series `c1` + `c2`.  The arguments
+    are sequences of coefficients ordered from lowest order term to
+    highest, i.e., [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Array representing the Hermite series of their sum.
+
+    See Also
+    --------
+    hermesub, hermemulx, hermemul, hermediv, hermepow
+
+    Notes
+    -----
+    Unlike multiplication, division, etc., the sum of two Hermite series
+    is a Hermite series (without having to "reproject" the result onto
+    the basis set) so addition, just like that of "standard" polynomials,
+    is simply "component-wise."
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeadd
+    >>> hermeadd([1, 2, 3], [1, 2, 3, 4])
+    array([2.,  4.,  6.,  4.])
+
+    """
+    return pu._add(c1, c2)
+
+
+def hermesub(c1, c2):
+    """
+    Subtract one Hermite series from another.
+
+    Returns the difference of two Hermite series `c1` - `c2`.  The
+    sequences of coefficients are from lowest order term to highest, i.e.,
+    [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Of Hermite series coefficients representing their difference.
+
+    See Also
+    --------
+    hermeadd, hermemulx, hermemul, hermediv, hermepow
+
+    Notes
+    -----
+    Unlike multiplication, division, etc., the difference of two Hermite
+    series is a Hermite series (without having to "reproject" the result
+    onto the basis set) so subtraction, just like that of "standard"
+    polynomials, is simply "component-wise."
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermesub
+    >>> hermesub([1, 2, 3, 4], [1, 2, 3])
+    array([0., 0., 0., 4.])
+
+    """
+    return pu._sub(c1, c2)
+
+
+def hermemulx(c):
+    """Multiply a Hermite series by x.
+
+    Multiply the Hermite series `c` by x, where x is the independent
+    variable.
+
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Array representing the result of the multiplication.
+
+    Notes
+    -----
+    The multiplication uses the recursion relationship for Hermite
+    polynomials in the form
+
+    .. math::
+
+        xP_i(x) = (P_{i + 1}(x) + iP_{i - 1}(x)))
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermemulx
+    >>> hermemulx([1, 2, 3])
+    array([2.,  7.,  2.,  3.])
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    # The zero series needs special treatment
+    if len(c) == 1 and c[0] == 0:
+        return c
+
+    prd = np.empty(len(c) + 1, dtype=c.dtype)
+    prd[0] = c[0]*0
+    prd[1] = c[0]
+    for i in range(1, len(c)):
+        prd[i + 1] = c[i]
+        prd[i - 1] += c[i]*i
+    return prd
+
+
+def hermemul(c1, c2):
+    """
+    Multiply one Hermite series by another.
+
+    Returns the product of two Hermite series `c1` * `c2`.  The arguments
+    are sequences of coefficients, from lowest order "term" to highest,
+    e.g., [1,2,3] represents the series ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    out : ndarray
+        Of Hermite series coefficients representing their product.
+
+    See Also
+    --------
+    hermeadd, hermesub, hermemulx, hermediv, hermepow
+
+    Notes
+    -----
+    In general, the (polynomial) product of two C-series results in terms
+    that are not in the Hermite polynomial basis set.  Thus, to express
+    the product as a Hermite series, it is necessary to "reproject" the
+    product onto said basis set, which may produce "unintuitive" (but
+    correct) results; see Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermemul
+    >>> hermemul([1, 2, 3], [0, 1, 2])
+    array([14.,  15.,  28.,   7.,   6.])
+
+    """
+    # s1, s2 are trimmed copies
+    [c1, c2] = pu.as_series([c1, c2])
+
+    if len(c1) > len(c2):
+        c = c2
+        xs = c1
+    else:
+        c = c1
+        xs = c2
+
+    if len(c) == 1:
+        c0 = c[0]*xs
+        c1 = 0
+    elif len(c) == 2:
+        c0 = c[0]*xs
+        c1 = c[1]*xs
+    else:
+        nd = len(c)
+        c0 = c[-2]*xs
+        c1 = c[-1]*xs
+        for i in range(3, len(c) + 1):
+            tmp = c0
+            nd = nd - 1
+            c0 = hermesub(c[-i]*xs, c1*(nd - 1))
+            c1 = hermeadd(tmp, hermemulx(c1))
+    return hermeadd(c0, hermemulx(c1))
+
+
+def hermediv(c1, c2):
+    """
+    Divide one Hermite series by another.
+
+    Returns the quotient-with-remainder of two Hermite series
+    `c1` / `c2`.  The arguments are sequences of coefficients from lowest
+    order "term" to highest, e.g., [1,2,3] represents the series
+    ``P_0 + 2*P_1 + 3*P_2``.
+
+    Parameters
+    ----------
+    c1, c2 : array_like
+        1-D arrays of Hermite series coefficients ordered from low to
+        high.
+
+    Returns
+    -------
+    [quo, rem] : ndarrays
+        Of Hermite series coefficients representing the quotient and
+        remainder.
+
+    See Also
+    --------
+    hermeadd, hermesub, hermemulx, hermemul, hermepow
+
+    Notes
+    -----
+    In general, the (polynomial) division of one Hermite series by another
+    results in quotient and remainder terms that are not in the Hermite
+    polynomial basis set.  Thus, to express these results as a Hermite
+    series, it is necessary to "reproject" the results onto the Hermite
+    basis set, which may produce "unintuitive" (but correct) results; see
+    Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermediv
+    >>> hermediv([ 14.,  15.,  28.,   7.,   6.], [0, 1, 2])
+    (array([1., 2., 3.]), array([0.]))
+    >>> hermediv([ 15.,  17.,  28.,   7.,   6.], [0, 1, 2])
+    (array([1., 2., 3.]), array([1., 2.]))
+
+    """
+    return pu._div(hermemul, c1, c2)
+
+
+def hermepow(c, pow, maxpower=16):
+    """Raise a Hermite series to a power.
+
+    Returns the Hermite series `c` raised to the power `pow`. The
+    argument `c` is a sequence of coefficients ordered from low to high.
+    i.e., [1,2,3] is the series  ``P_0 + 2*P_1 + 3*P_2.``
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of Hermite series coefficients ordered from low to
+        high.
+    pow : integer
+        Power to which the series will be raised
+    maxpower : integer, optional
+        Maximum power allowed. This is mainly to limit growth of the series
+        to unmanageable size. Default is 16
+
+    Returns
+    -------
+    coef : ndarray
+        Hermite series of power.
+
+    See Also
+    --------
+    hermeadd, hermesub, hermemulx, hermemul, hermediv
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermepow
+    >>> hermepow([1, 2, 3], 2)
+    array([23.,  28.,  46.,  12.,   9.])
+
+    """
+    return pu._pow(hermemul, c, pow, maxpower)
+
+
+def hermeder(c, m=1, scl=1, axis=0):
+    """
+    Differentiate a Hermite_e series.
+
+    Returns the series coefficients `c` differentiated `m` times along
+    `axis`.  At each iteration the result is multiplied by `scl` (the
+    scaling factor is for use in a linear change of variable). The argument
+    `c` is an array of coefficients from low to high degree along each
+    axis, e.g., [1,2,3] represents the series ``1*He_0 + 2*He_1 + 3*He_2``
+    while [[1,2],[1,2]] represents ``1*He_0(x)*He_0(y) + 1*He_1(x)*He_0(y)
+    + 2*He_0(x)*He_1(y) + 2*He_1(x)*He_1(y)`` if axis=0 is ``x`` and axis=1
+    is ``y``.
+
+    Parameters
+    ----------
+    c : array_like
+        Array of Hermite_e series coefficients. If `c` is multidimensional
+        the different axis correspond to different variables with the
+        degree in each axis given by the corresponding index.
+    m : int, optional
+        Number of derivatives taken, must be non-negative. (Default: 1)
+    scl : scalar, optional
+        Each differentiation is multiplied by `scl`.  The end result is
+        multiplication by ``scl**m``.  This is for use in a linear change of
+        variable. (Default: 1)
+    axis : int, optional
+        Axis over which the derivative is taken. (Default: 0).
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    der : ndarray
+        Hermite series of the derivative.
+
+    See Also
+    --------
+    hermeint
+
+    Notes
+    -----
+    In general, the result of differentiating a Hermite series does not
+    resemble the same operation on a power series. Thus the result of this
+    function may be "unintuitive," albeit correct; see Examples section
+    below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeder
+    >>> hermeder([ 1.,  1.,  1.,  1.])
+    array([1.,  2.,  3.])
+    >>> hermeder([-0.25,  1.,  1./2.,  1./3.,  1./4 ], m=2)
+    array([1.,  2.,  3.])
+
+    """
+    c = np.array(c, ndmin=1, copy=True)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    cnt = pu._deprecate_as_int(m, "the order of derivation")
+    iaxis = pu._deprecate_as_int(axis, "the axis")
+    if cnt < 0:
+        raise ValueError("The order of derivation must be non-negative")
+    iaxis = normalize_axis_index(iaxis, c.ndim)
+
+    if cnt == 0:
+        return c
+
+    c = np.moveaxis(c, iaxis, 0)
+    n = len(c)
+    if cnt >= n:
+        return c[:1]*0
+    else:
+        for i in range(cnt):
+            n = n - 1
+            c *= scl
+            der = np.empty((n,) + c.shape[1:], dtype=c.dtype)
+            for j in range(n, 0, -1):
+                der[j - 1] = j*c[j]
+            c = der
+    c = np.moveaxis(c, 0, iaxis)
+    return c
+
+
+def hermeint(c, m=1, k=[], lbnd=0, scl=1, axis=0):
+    """
+    Integrate a Hermite_e series.
+
+    Returns the Hermite_e series coefficients `c` integrated `m` times from
+    `lbnd` along `axis`. At each iteration the resulting series is
+    **multiplied** by `scl` and an integration constant, `k`, is added.
+    The scaling factor is for use in a linear change of variable.  ("Buyer
+    beware": note that, depending on what one is doing, one may want `scl`
+    to be the reciprocal of what one might expect; for more information,
+    see the Notes section below.)  The argument `c` is an array of
+    coefficients from low to high degree along each axis, e.g., [1,2,3]
+    represents the series ``H_0 + 2*H_1 + 3*H_2`` while [[1,2],[1,2]]
+    represents ``1*H_0(x)*H_0(y) + 1*H_1(x)*H_0(y) + 2*H_0(x)*H_1(y) +
+    2*H_1(x)*H_1(y)`` if axis=0 is ``x`` and axis=1 is ``y``.
+
+    Parameters
+    ----------
+    c : array_like
+        Array of Hermite_e series coefficients. If c is multidimensional
+        the different axis correspond to different variables with the
+        degree in each axis given by the corresponding index.
+    m : int, optional
+        Order of integration, must be positive. (Default: 1)
+    k : {[], list, scalar}, optional
+        Integration constant(s).  The value of the first integral at
+        ``lbnd`` is the first value in the list, the value of the second
+        integral at ``lbnd`` is the second value, etc.  If ``k == []`` (the
+        default), all constants are set to zero.  If ``m == 1``, a single
+        scalar can be given instead of a list.
+    lbnd : scalar, optional
+        The lower bound of the integral. (Default: 0)
+    scl : scalar, optional
+        Following each integration the result is *multiplied* by `scl`
+        before the integration constant is added. (Default: 1)
+    axis : int, optional
+        Axis over which the integral is taken. (Default: 0).
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    S : ndarray
+        Hermite_e series coefficients of the integral.
+
+    Raises
+    ------
+    ValueError
+        If ``m < 0``, ``len(k) > m``, ``np.ndim(lbnd) != 0``, or
+        ``np.ndim(scl) != 0``.
+
+    See Also
+    --------
+    hermeder
+
+    Notes
+    -----
+    Note that the result of each integration is *multiplied* by `scl`.
+    Why is this important to note?  Say one is making a linear change of
+    variable :math:`u = ax + b` in an integral relative to `x`.  Then
+    :math:`dx = du/a`, so one will need to set `scl` equal to
+    :math:`1/a` - perhaps not what one would have first thought.
+
+    Also note that, in general, the result of integrating a C-series needs
+    to be "reprojected" onto the C-series basis set.  Thus, typically,
+    the result of this function is "unintuitive," albeit correct; see
+    Examples section below.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeint
+    >>> hermeint([1, 2, 3]) # integrate once, value 0 at 0.
+    array([1., 1., 1., 1.])
+    >>> hermeint([1, 2, 3], m=2) # integrate twice, value & deriv 0 at 0
+    array([-0.25      ,  1.        ,  0.5       ,  0.33333333,  0.25      ]) # may vary
+    >>> hermeint([1, 2, 3], k=1) # integrate once, value 1 at 0.
+    array([2., 1., 1., 1.])
+    >>> hermeint([1, 2, 3], lbnd=-1) # integrate once, value 0 at -1
+    array([-1.,  1.,  1.,  1.])
+    >>> hermeint([1, 2, 3], m=2, k=[1, 2], lbnd=-1)
+    array([ 1.83333333,  0.        ,  0.5       ,  0.33333333,  0.25      ]) # may vary
+
+    """
+    c = np.array(c, ndmin=1, copy=True)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    if not np.iterable(k):
+        k = [k]
+    cnt = pu._deprecate_as_int(m, "the order of integration")
+    iaxis = pu._deprecate_as_int(axis, "the axis")
+    if cnt < 0:
+        raise ValueError("The order of integration must be non-negative")
+    if len(k) > cnt:
+        raise ValueError("Too many integration constants")
+    if np.ndim(lbnd) != 0:
+        raise ValueError("lbnd must be a scalar.")
+    if np.ndim(scl) != 0:
+        raise ValueError("scl must be a scalar.")
+    iaxis = normalize_axis_index(iaxis, c.ndim)
+
+    if cnt == 0:
+        return c
+
+    c = np.moveaxis(c, iaxis, 0)
+    k = list(k) + [0]*(cnt - len(k))
+    for i in range(cnt):
+        n = len(c)
+        c *= scl
+        if n == 1 and np.all(c[0] == 0):
+            c[0] += k[i]
+        else:
+            tmp = np.empty((n + 1,) + c.shape[1:], dtype=c.dtype)
+            tmp[0] = c[0]*0
+            tmp[1] = c[0]
+            for j in range(1, n):
+                tmp[j + 1] = c[j]/(j + 1)
+            tmp[0] += k[i] - hermeval(lbnd, tmp)
+            c = tmp
+    c = np.moveaxis(c, 0, iaxis)
+    return c
+
+
+def hermeval(x, c, tensor=True):
+    """
+    Evaluate an HermiteE series at points x.
+
+    If `c` is of length `n + 1`, this function returns the value:
+
+    .. math:: p(x) = c_0 * He_0(x) + c_1 * He_1(x) + ... + c_n * He_n(x)
+
+    The parameter `x` is converted to an array only if it is a tuple or a
+    list, otherwise it is treated as a scalar. In either case, either `x`
+    or its elements must support multiplication and addition both with
+    themselves and with the elements of `c`.
+
+    If `c` is a 1-D array, then `p(x)` will have the same shape as `x`.  If
+    `c` is multidimensional, then the shape of the result depends on the
+    value of `tensor`. If `tensor` is true the shape will be c.shape[1:] +
+    x.shape. If `tensor` is false the shape will be c.shape[1:]. Note that
+    scalars have shape (,).
+
+    Trailing zeros in the coefficients will be used in the evaluation, so
+    they should be avoided if efficiency is a concern.
+
+    Parameters
+    ----------
+    x : array_like, compatible object
+        If `x` is a list or tuple, it is converted to an ndarray, otherwise
+        it is left unchanged and treated as a scalar. In either case, `x`
+        or its elements must support addition and multiplication with
+        with themselves and with the elements of `c`.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree n are contained in c[n]. If `c` is multidimensional the
+        remaining indices enumerate multiple polynomials. In the two
+        dimensional case the coefficients may be thought of as stored in
+        the columns of `c`.
+    tensor : boolean, optional
+        If True, the shape of the coefficient array is extended with ones
+        on the right, one for each dimension of `x`. Scalars have dimension 0
+        for this action. The result is that every column of coefficients in
+        `c` is evaluated for every element of `x`. If False, `x` is broadcast
+        over the columns of `c` for the evaluation.  This keyword is useful
+        when `c` is multidimensional. The default value is True.
+
+        .. versionadded:: 1.7.0
+
+    Returns
+    -------
+    values : ndarray, algebra_like
+        The shape of the return value is described above.
+
+    See Also
+    --------
+    hermeval2d, hermegrid2d, hermeval3d, hermegrid3d
+
+    Notes
+    -----
+    The evaluation uses Clenshaw recursion, aka synthetic division.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeval
+    >>> coef = [1,2,3]
+    >>> hermeval(1, coef)
+    3.0
+    >>> hermeval([[1,2],[3,4]], coef)
+    array([[ 3., 14.],
+           [31., 54.]])
+
+    """
+    c = np.array(c, ndmin=1, copy=False)
+    if c.dtype.char in '?bBhHiIlLqQpP':
+        c = c.astype(np.double)
+    if isinstance(x, (tuple, list)):
+        x = np.asarray(x)
+    if isinstance(x, np.ndarray) and tensor:
+        c = c.reshape(c.shape + (1,)*x.ndim)
+
+    if len(c) == 1:
+        c0 = c[0]
+        c1 = 0
+    elif len(c) == 2:
+        c0 = c[0]
+        c1 = c[1]
+    else:
+        nd = len(c)
+        c0 = c[-2]
+        c1 = c[-1]
+        for i in range(3, len(c) + 1):
+            tmp = c0
+            nd = nd - 1
+            c0 = c[-i] - c1*(nd - 1)
+            c1 = tmp + c1*x
+    return c0 + c1*x
+
+
+def hermeval2d(x, y, c):
+    """
+    Evaluate a 2-D HermiteE series at points (x, y).
+
+    This function returns the values:
+
+    .. math:: p(x,y) = \\sum_{i,j} c_{i,j} * He_i(x) * He_j(y)
+
+    The parameters `x` and `y` are converted to arrays only if they are
+    tuples or a lists, otherwise they are treated as a scalars and they
+    must have the same shape after conversion. In either case, either `x`
+    and `y` or their elements must support multiplication and addition both
+    with themselves and with the elements of `c`.
+
+    If `c` is a 1-D array a one is implicitly appended to its shape to make
+    it 2-D. The shape of the result will be c.shape[2:] + x.shape.
+
+    Parameters
+    ----------
+    x, y : array_like, compatible objects
+        The two dimensional series is evaluated at the points `(x, y)`,
+        where `x` and `y` must have the same shape. If `x` or `y` is a list
+        or tuple, it is first converted to an ndarray, otherwise it is left
+        unchanged and if it isn't an ndarray it is treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficient of the term
+        of multi-degree i,j is contained in ``c[i,j]``. If `c` has
+        dimension greater than two the remaining indices enumerate multiple
+        sets of coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points formed with
+        pairs of corresponding values from `x` and `y`.
+
+    See Also
+    --------
+    hermeval, hermegrid2d, hermeval3d, hermegrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._valnd(hermeval, c, x, y)
+
+
+def hermegrid2d(x, y, c):
+    """
+    Evaluate a 2-D HermiteE series on the Cartesian product of x and y.
+
+    This function returns the values:
+
+    .. math:: p(a,b) = \\sum_{i,j} c_{i,j} * H_i(a) * H_j(b)
+
+    where the points `(a, b)` consist of all pairs formed by taking
+    `a` from `x` and `b` from `y`. The resulting points form a grid with
+    `x` in the first dimension and `y` in the second.
+
+    The parameters `x` and `y` are converted to arrays only if they are
+    tuples or a lists, otherwise they are treated as a scalars. In either
+    case, either `x` and `y` or their elements must support multiplication
+    and addition both with themselves and with the elements of `c`.
+
+    If `c` has fewer than two dimensions, ones are implicitly appended to
+    its shape to make it 2-D. The shape of the result will be c.shape[2:] +
+    x.shape.
+
+    Parameters
+    ----------
+    x, y : array_like, compatible objects
+        The two dimensional series is evaluated at the points in the
+        Cartesian product of `x` and `y`.  If `x` or `y` is a list or
+        tuple, it is first converted to an ndarray, otherwise it is left
+        unchanged and, if it isn't an ndarray, it is treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree i,j are contained in ``c[i,j]``. If `c` has dimension
+        greater than two the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points in the Cartesian
+        product of `x` and `y`.
+
+    See Also
+    --------
+    hermeval, hermeval2d, hermeval3d, hermegrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._gridnd(hermeval, c, x, y)
+
+
+def hermeval3d(x, y, z, c):
+    """
+    Evaluate a 3-D Hermite_e series at points (x, y, z).
+
+    This function returns the values:
+
+    .. math:: p(x,y,z) = \\sum_{i,j,k} c_{i,j,k} * He_i(x) * He_j(y) * He_k(z)
+
+    The parameters `x`, `y`, and `z` are converted to arrays only if
+    they are tuples or a lists, otherwise they are treated as a scalars and
+    they must have the same shape after conversion. In either case, either
+    `x`, `y`, and `z` or their elements must support multiplication and
+    addition both with themselves and with the elements of `c`.
+
+    If `c` has fewer than 3 dimensions, ones are implicitly appended to its
+    shape to make it 3-D. The shape of the result will be c.shape[3:] +
+    x.shape.
+
+    Parameters
+    ----------
+    x, y, z : array_like, compatible object
+        The three dimensional series is evaluated at the points
+        `(x, y, z)`, where `x`, `y`, and `z` must have the same shape.  If
+        any of `x`, `y`, or `z` is a list or tuple, it is first converted
+        to an ndarray, otherwise it is left unchanged and if it isn't an
+        ndarray it is  treated as a scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficient of the term of
+        multi-degree i,j,k is contained in ``c[i,j,k]``. If `c` has dimension
+        greater than 3 the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the multidimensional polynomial on points formed with
+        triples of corresponding values from `x`, `y`, and `z`.
+
+    See Also
+    --------
+    hermeval, hermeval2d, hermegrid2d, hermegrid3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._valnd(hermeval, c, x, y, z)
+
+
+def hermegrid3d(x, y, z, c):
+    """
+    Evaluate a 3-D HermiteE series on the Cartesian product of x, y, and z.
+
+    This function returns the values:
+
+    .. math:: p(a,b,c) = \\sum_{i,j,k} c_{i,j,k} * He_i(a) * He_j(b) * He_k(c)
+
+    where the points `(a, b, c)` consist of all triples formed by taking
+    `a` from `x`, `b` from `y`, and `c` from `z`. The resulting points form
+    a grid with `x` in the first dimension, `y` in the second, and `z` in
+    the third.
+
+    The parameters `x`, `y`, and `z` are converted to arrays only if they
+    are tuples or a lists, otherwise they are treated as a scalars. In
+    either case, either `x`, `y`, and `z` or their elements must support
+    multiplication and addition both with themselves and with the elements
+    of `c`.
+
+    If `c` has fewer than three dimensions, ones are implicitly appended to
+    its shape to make it 3-D. The shape of the result will be c.shape[3:] +
+    x.shape + y.shape + z.shape.
+
+    Parameters
+    ----------
+    x, y, z : array_like, compatible objects
+        The three dimensional series is evaluated at the points in the
+        Cartesian product of `x`, `y`, and `z`.  If `x`,`y`, or `z` is a
+        list or tuple, it is first converted to an ndarray, otherwise it is
+        left unchanged and, if it isn't an ndarray, it is treated as a
+        scalar.
+    c : array_like
+        Array of coefficients ordered so that the coefficients for terms of
+        degree i,j are contained in ``c[i,j]``. If `c` has dimension
+        greater than two the remaining indices enumerate multiple sets of
+        coefficients.
+
+    Returns
+    -------
+    values : ndarray, compatible object
+        The values of the two dimensional polynomial at points in the Cartesian
+        product of `x` and `y`.
+
+    See Also
+    --------
+    hermeval, hermeval2d, hermegrid2d, hermeval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._gridnd(hermeval, c, x, y, z)
+
+
+def hermevander(x, deg):
+    """Pseudo-Vandermonde matrix of given degree.
+
+    Returns the pseudo-Vandermonde matrix of degree `deg` and sample points
+    `x`. The pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., i] = He_i(x),
+
+    where `0 <= i <= deg`. The leading indices of `V` index the elements of
+    `x` and the last index is the degree of the HermiteE polynomial.
+
+    If `c` is a 1-D array of coefficients of length `n + 1` and `V` is the
+    array ``V = hermevander(x, n)``, then ``np.dot(V, c)`` and
+    ``hermeval(x, c)`` are the same up to roundoff. This equivalence is
+    useful both for least squares fitting and for the evaluation of a large
+    number of HermiteE series of the same degree and sample points.
+
+    Parameters
+    ----------
+    x : array_like
+        Array of points. The dtype is converted to float64 or complex128
+        depending on whether any of the elements are complex. If `x` is
+        scalar it is converted to a 1-D array.
+    deg : int
+        Degree of the resulting matrix.
+
+    Returns
+    -------
+    vander : ndarray
+        The pseudo-Vandermonde matrix. The shape of the returned matrix is
+        ``x.shape + (deg + 1,)``, where The last index is the degree of the
+        corresponding HermiteE polynomial.  The dtype will be the same as
+        the converted `x`.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermevander
+    >>> x = np.array([-1, 0, 1])
+    >>> hermevander(x, 3)
+    array([[ 1., -1.,  0.,  2.],
+           [ 1.,  0., -1., -0.],
+           [ 1.,  1.,  0., -2.]])
+
+    """
+    ideg = pu._deprecate_as_int(deg, "deg")
+    if ideg < 0:
+        raise ValueError("deg must be non-negative")
+
+    x = np.array(x, copy=False, ndmin=1) + 0.0
+    dims = (ideg + 1,) + x.shape
+    dtyp = x.dtype
+    v = np.empty(dims, dtype=dtyp)
+    v[0] = x*0 + 1
+    if ideg > 0:
+        v[1] = x
+        for i in range(2, ideg + 1):
+            v[i] = (v[i-1]*x - v[i-2]*(i - 1))
+    return np.moveaxis(v, 0, -1)
+
+
+def hermevander2d(x, y, deg):
+    """Pseudo-Vandermonde matrix of given degrees.
+
+    Returns the pseudo-Vandermonde matrix of degrees `deg` and sample
+    points `(x, y)`. The pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., (deg[1] + 1)*i + j] = He_i(x) * He_j(y),
+
+    where `0 <= i <= deg[0]` and `0 <= j <= deg[1]`. The leading indices of
+    `V` index the points `(x, y)` and the last index encodes the degrees of
+    the HermiteE polynomials.
+
+    If ``V = hermevander2d(x, y, [xdeg, ydeg])``, then the columns of `V`
+    correspond to the elements of a 2-D coefficient array `c` of shape
+    (xdeg + 1, ydeg + 1) in the order
+
+    .. math:: c_{00}, c_{01}, c_{02} ... , c_{10}, c_{11}, c_{12} ...
+
+    and ``np.dot(V, c.flat)`` and ``hermeval2d(x, y, c)`` will be the same
+    up to roundoff. This equivalence is useful both for least squares
+    fitting and for the evaluation of a large number of 2-D HermiteE
+    series of the same degrees and sample points.
+
+    Parameters
+    ----------
+    x, y : array_like
+        Arrays of point coordinates, all of the same shape. The dtypes
+        will be converted to either float64 or complex128 depending on
+        whether any of the elements are complex. Scalars are converted to
+        1-D arrays.
+    deg : list of ints
+        List of maximum degrees of the form [x_deg, y_deg].
+
+    Returns
+    -------
+    vander2d : ndarray
+        The shape of the returned matrix is ``x.shape + (order,)``, where
+        :math:`order = (deg[0]+1)*(deg[1]+1)`.  The dtype will be the same
+        as the converted `x` and `y`.
+
+    See Also
+    --------
+    hermevander, hermevander3d, hermeval2d, hermeval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._vander_nd_flat((hermevander, hermevander), (x, y), deg)
+
+
+def hermevander3d(x, y, z, deg):
+    """Pseudo-Vandermonde matrix of given degrees.
+
+    Returns the pseudo-Vandermonde matrix of degrees `deg` and sample
+    points `(x, y, z)`. If `l, m, n` are the given degrees in `x, y, z`,
+    then Hehe pseudo-Vandermonde matrix is defined by
+
+    .. math:: V[..., (m+1)(n+1)i + (n+1)j + k] = He_i(x)*He_j(y)*He_k(z),
+
+    where `0 <= i <= l`, `0 <= j <= m`, and `0 <= j <= n`.  The leading
+    indices of `V` index the points `(x, y, z)` and the last index encodes
+    the degrees of the HermiteE polynomials.
+
+    If ``V = hermevander3d(x, y, z, [xdeg, ydeg, zdeg])``, then the columns
+    of `V` correspond to the elements of a 3-D coefficient array `c` of
+    shape (xdeg + 1, ydeg + 1, zdeg + 1) in the order
+
+    .. math:: c_{000}, c_{001}, c_{002},... , c_{010}, c_{011}, c_{012},...
+
+    and  ``np.dot(V, c.flat)`` and ``hermeval3d(x, y, z, c)`` will be the
+    same up to roundoff. This equivalence is useful both for least squares
+    fitting and for the evaluation of a large number of 3-D HermiteE
+    series of the same degrees and sample points.
+
+    Parameters
+    ----------
+    x, y, z : array_like
+        Arrays of point coordinates, all of the same shape. The dtypes will
+        be converted to either float64 or complex128 depending on whether
+        any of the elements are complex. Scalars are converted to 1-D
+        arrays.
+    deg : list of ints
+        List of maximum degrees of the form [x_deg, y_deg, z_deg].
+
+    Returns
+    -------
+    vander3d : ndarray
+        The shape of the returned matrix is ``x.shape + (order,)``, where
+        :math:`order = (deg[0]+1)*(deg[1]+1)*(deg[2]+1)`.  The dtype will
+        be the same as the converted `x`, `y`, and `z`.
+
+    See Also
+    --------
+    hermevander, hermevander3d, hermeval2d, hermeval3d
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    return pu._vander_nd_flat((hermevander, hermevander, hermevander), (x, y, z), deg)
+
+
+def hermefit(x, y, deg, rcond=None, full=False, w=None):
+    """
+    Least squares fit of Hermite series to data.
+
+    Return the coefficients of a HermiteE series of degree `deg` that is
+    the least squares fit to the data values `y` given at points `x`. If
+    `y` is 1-D the returned coefficients will also be 1-D. If `y` is 2-D
+    multiple fits are done, one for each column of `y`, and the resulting
+    coefficients are stored in the corresponding columns of a 2-D return.
+    The fitted polynomial(s) are in the form
+
+    .. math::  p(x) = c_0 + c_1 * He_1(x) + ... + c_n * He_n(x),
+
+    where `n` is `deg`.
+
+    Parameters
+    ----------
+    x : array_like, shape (M,)
+        x-coordinates of the M sample points ``(x[i], y[i])``.
+    y : array_like, shape (M,) or (M, K)
+        y-coordinates of the sample points. Several data sets of sample
+        points sharing the same x-coordinates can be fitted at once by
+        passing in a 2D-array that contains one dataset per column.
+    deg : int or 1-D array_like
+        Degree(s) of the fitting polynomials. If `deg` is a single integer
+        all terms up to and including the `deg`'th term are included in the
+        fit. For NumPy versions >= 1.11.0 a list of integers specifying the
+        degrees of the terms to include may be used instead.
+    rcond : float, optional
+        Relative condition number of the fit. Singular values smaller than
+        this relative to the largest singular value will be ignored. The
+        default value is len(x)*eps, where eps is the relative precision of
+        the float type, about 2e-16 in most cases.
+    full : bool, optional
+        Switch determining nature of return value. When it is False (the
+        default) just the coefficients are returned, when True diagnostic
+        information from the singular value decomposition is also returned.
+    w : array_like, shape (`M`,), optional
+        Weights. If not None, the weight ``w[i]`` applies to the unsquared
+        residual ``y[i] - y_hat[i]`` at ``x[i]``. Ideally the weights are
+        chosen so that the errors of the products ``w[i]*y[i]`` all have the
+        same variance.  When using inverse-variance weighting, use
+        ``w[i] = 1/sigma(y[i])``.  The default value is None.
+
+    Returns
+    -------
+    coef : ndarray, shape (M,) or (M, K)
+        Hermite coefficients ordered from low to high. If `y` was 2-D,
+        the coefficients for the data in column k  of `y` are in column
+        `k`.
+
+    [residuals, rank, singular_values, rcond] : list
+        These values are only returned if ``full == True``
+
+        - residuals -- sum of squared residuals of the least squares fit
+        - rank -- the numerical rank of the scaled Vandermonde matrix
+        - singular_values -- singular values of the scaled Vandermonde matrix
+        - rcond -- value of `rcond`.
+
+        For more details, see `numpy.linalg.lstsq`.
+
+    Warns
+    -----
+    RankWarning
+        The rank of the coefficient matrix in the least-squares fit is
+        deficient. The warning is only raised if ``full = False``.  The
+        warnings can be turned off by
+
+        >>> import warnings
+        >>> warnings.simplefilter('ignore', np.RankWarning)
+
+    See Also
+    --------
+    numpy.polynomial.chebyshev.chebfit
+    numpy.polynomial.legendre.legfit
+    numpy.polynomial.polynomial.polyfit
+    numpy.polynomial.hermite.hermfit
+    numpy.polynomial.laguerre.lagfit
+    hermeval : Evaluates a Hermite series.
+    hermevander : pseudo Vandermonde matrix of Hermite series.
+    hermeweight : HermiteE weight function.
+    numpy.linalg.lstsq : Computes a least-squares fit from the matrix.
+    scipy.interpolate.UnivariateSpline : Computes spline fits.
+
+    Notes
+    -----
+    The solution is the coefficients of the HermiteE series `p` that
+    minimizes the sum of the weighted squared errors
+
+    .. math:: E = \\sum_j w_j^2 * |y_j - p(x_j)|^2,
+
+    where the :math:`w_j` are the weights. This problem is solved by
+    setting up the (typically) overdetermined matrix equation
+
+    .. math:: V(x) * c = w * y,
+
+    where `V` is the pseudo Vandermonde matrix of `x`, the elements of `c`
+    are the coefficients to be solved for, and the elements of `y` are the
+    observed values.  This equation is then solved using the singular value
+    decomposition of `V`.
+
+    If some of the singular values of `V` are so small that they are
+    neglected, then a `RankWarning` will be issued. This means that the
+    coefficient values may be poorly determined. Using a lower order fit
+    will usually get rid of the warning.  The `rcond` parameter can also be
+    set to a value smaller than its default, but the resulting fit may be
+    spurious and have large contributions from roundoff error.
+
+    Fits using HermiteE series are probably most useful when the data can
+    be approximated by ``sqrt(w(x)) * p(x)``, where `w(x)` is the HermiteE
+    weight. In that case the weight ``sqrt(w(x[i]))`` should be used
+    together with data values ``y[i]/sqrt(w(x[i]))``. The weight function is
+    available as `hermeweight`.
+
+    References
+    ----------
+    .. [1] Wikipedia, "Curve fitting",
+           https://en.wikipedia.org/wiki/Curve_fitting
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermefit, hermeval
+    >>> x = np.linspace(-10, 10)
+    >>> np.random.seed(123)
+    >>> err = np.random.randn(len(x))/10
+    >>> y = hermeval(x, [1, 2, 3]) + err
+    >>> hermefit(x, y, 2)
+    array([ 1.01690445,  1.99951418,  2.99948696]) # may vary
+
+    """
+    return pu._fit(hermevander, x, y, deg, rcond, full, w)
+
+
+def hermecompanion(c):
+    """
+    Return the scaled companion matrix of c.
+
+    The basis polynomials are scaled so that the companion matrix is
+    symmetric when `c` is an HermiteE basis polynomial. This provides
+    better eigenvalue estimates than the unscaled case and for basis
+    polynomials the eigenvalues are guaranteed to be real if
+    `numpy.linalg.eigvalsh` is used to obtain them.
+
+    Parameters
+    ----------
+    c : array_like
+        1-D array of HermiteE series coefficients ordered from low to high
+        degree.
+
+    Returns
+    -------
+    mat : ndarray
+        Scaled companion matrix of dimensions (deg, deg).
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    if len(c) < 2:
+        raise ValueError('Series must have maximum degree of at least 1.')
+    if len(c) == 2:
+        return np.array([[-c[0]/c[1]]])
+
+    n = len(c) - 1
+    mat = np.zeros((n, n), dtype=c.dtype)
+    scl = np.hstack((1., 1./np.sqrt(np.arange(n - 1, 0, -1))))
+    scl = np.multiply.accumulate(scl)[::-1]
+    top = mat.reshape(-1)[1::n+1]
+    bot = mat.reshape(-1)[n::n+1]
+    top[...] = np.sqrt(np.arange(1, n))
+    bot[...] = top
+    mat[:, -1] -= scl*c[:-1]/c[-1]
+    return mat
+
+
+def hermeroots(c):
+    """
+    Compute the roots of a HermiteE series.
+
+    Return the roots (a.k.a. "zeros") of the polynomial
+
+    .. math:: p(x) = \\sum_i c[i] * He_i(x).
+
+    Parameters
+    ----------
+    c : 1-D array_like
+        1-D array of coefficients.
+
+    Returns
+    -------
+    out : ndarray
+        Array of the roots of the series. If all the roots are real,
+        then `out` is also real, otherwise it is complex.
+
+    See Also
+    --------
+    numpy.polynomial.polynomial.polyroots
+    numpy.polynomial.legendre.legroots
+    numpy.polynomial.laguerre.lagroots
+    numpy.polynomial.hermite.hermroots
+    numpy.polynomial.chebyshev.chebroots
+
+    Notes
+    -----
+    The root estimates are obtained as the eigenvalues of the companion
+    matrix, Roots far from the origin of the complex plane may have large
+    errors due to the numerical instability of the series for such
+    values. Roots with multiplicity greater than 1 will also show larger
+    errors as the value of the series near such points is relatively
+    insensitive to errors in the roots. Isolated roots near the origin can
+    be improved by a few iterations of Newton's method.
+
+    The HermiteE series basis polynomials aren't powers of `x` so the
+    results of this function may seem unintuitive.
+
+    Examples
+    --------
+    >>> from numpy.polynomial.hermite_e import hermeroots, hermefromroots
+    >>> coef = hermefromroots([-1, 0, 1])
+    >>> coef
+    array([0., 2., 0., 1.])
+    >>> hermeroots(coef)
+    array([-1.,  0.,  1.]) # may vary
+
+    """
+    # c is a trimmed copy
+    [c] = pu.as_series([c])
+    if len(c) <= 1:
+        return np.array([], dtype=c.dtype)
+    if len(c) == 2:
+        return np.array([-c[0]/c[1]])
+
+    # rotated companion matrix reduces error
+    m = hermecompanion(c)[::-1,::-1]
+    r = la.eigvals(m)
+    r.sort()
+    return r
+
+
+def _normed_hermite_e_n(x, n):
+    """
+    Evaluate a normalized HermiteE polynomial.
+
+    Compute the value of the normalized HermiteE polynomial of degree ``n``
+    at the points ``x``.
+
+
+    Parameters
+    ----------
+    x : ndarray of double.
+        Points at which to evaluate the function
+    n : int
+        Degree of the normalized HermiteE function to be evaluated.
+
+    Returns
+    -------
+    values : ndarray
+        The shape of the return value is described above.
+
+    Notes
+    -----
+    .. versionadded:: 1.10.0
+
+    This function is needed for finding the Gauss points and integration
+    weights for high degrees. The values of the standard HermiteE functions
+    overflow when n >= 207.
+
+    """
+    if n == 0:
+        return np.full(x.shape, 1/np.sqrt(np.sqrt(2*np.pi)))
+
+    c0 = 0.
+    c1 = 1./np.sqrt(np.sqrt(2*np.pi))
+    nd = float(n)
+    for i in range(n - 1):
+        tmp = c0
+        c0 = -c1*np.sqrt((nd - 1.)/nd)
+        c1 = tmp + c1*x*np.sqrt(1./nd)
+        nd = nd - 1.0
+    return c0 + c1*x
+
+
+def hermegauss(deg):
+    """
+    Gauss-HermiteE quadrature.
+
+    Computes the sample points and weights for Gauss-HermiteE quadrature.
+    These sample points and weights will correctly integrate polynomials of
+    degree :math:`2*deg - 1` or less over the interval :math:`[-\\inf, \\inf]`
+    with the weight function :math:`f(x) = \\exp(-x^2/2)`.
+
+    Parameters
+    ----------
+    deg : int
+        Number of sample points and weights. It must be >= 1.
+
+    Returns
+    -------
+    x : ndarray
+        1-D ndarray containing the sample points.
+    y : ndarray
+        1-D ndarray containing the weights.
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    The results have only been tested up to degree 100, higher degrees may
+    be problematic. The weights are determined by using the fact that
+
+    .. math:: w_k = c / (He'_n(x_k) * He_{n-1}(x_k))
+
+    where :math:`c` is a constant independent of :math:`k` and :math:`x_k`
+    is the k'th root of :math:`He_n`, and then scaling the results to get
+    the right value when integrating 1.
+
+    """
+    ideg = pu._deprecate_as_int(deg, "deg")
+    if ideg <= 0:
+        raise ValueError("deg must be a positive integer")
+
+    # first approximation of roots. We use the fact that the companion
+    # matrix is symmetric in this case in order to obtain better zeros.
+    c = np.array([0]*deg + [1])
+    m = hermecompanion(c)
+    x = la.eigvalsh(m)
+
+    # improve roots by one application of Newton
+    dy = _normed_hermite_e_n(x, ideg)
+    df = _normed_hermite_e_n(x, ideg - 1) * np.sqrt(ideg)
+    x -= dy/df
+
+    # compute the weights. We scale the factor to avoid possible numerical
+    # overflow.
+    fm = _normed_hermite_e_n(x, ideg - 1)
+    fm /= np.abs(fm).max()
+    w = 1/(fm * fm)
+
+    # for Hermite_e we can also symmetrize
+    w = (w + w[::-1])/2
+    x = (x - x[::-1])/2
+
+    # scale w to get the right value
+    w *= np.sqrt(2*np.pi) / w.sum()
+
+    return x, w
+
+
+def hermeweight(x):
+    """Weight function of the Hermite_e polynomials.
+
+    The weight function is :math:`\\exp(-x^2/2)` and the interval of
+    integration is :math:`[-\\inf, \\inf]`. the HermiteE polynomials are
+    orthogonal, but not normalized, with respect to this weight function.
+
+    Parameters
+    ----------
+    x : array_like
+       Values at which the weight function will be computed.
+
+    Returns
+    -------
+    w : ndarray
+       The weight function at `x`.
+
+    Notes
+    -----
+
+    .. versionadded:: 1.7.0
+
+    """
+    w = np.exp(-.5*x**2)
+    return w
+
+
+#
+# HermiteE series class
+#
+
+class HermiteE(ABCPolyBase):
+    """An HermiteE series class.
+
+    The HermiteE class provides the standard Python numerical methods
+    '+', '-', '*', '//', '%', 'divmod', '**', and '()' as well as the
+    attributes and methods listed in the `ABCPolyBase` documentation.
+
+    Parameters
+    ----------
+    coef : array_like
+        HermiteE coefficients in order of increasing degree, i.e,
+        ``(1, 2, 3)`` gives ``1*He_0(x) + 2*He_1(X) + 3*He_2(x)``.
+    domain : (2,) array_like, optional
+        Domain to use. The interval ``[domain[0], domain[1]]`` is mapped
+        to the interval ``[window[0], window[1]]`` by shifting and scaling.
+        The default value is [-1, 1].
+    window : (2,) array_like, optional
+        Window, see `domain` for its use. The default value is [-1, 1].
+
+        .. versionadded:: 1.6.0
+    symbol : str, optional
+        Symbol used to represent the independent variable in string
+        representations of the polynomial expression, e.g. for printing.
+        The symbol must be a valid Python identifier. Default value is 'x'.
+
+        .. versionadded:: 1.24
+
+    """
+    # Virtual Functions
+    _add = staticmethod(hermeadd)
+    _sub = staticmethod(hermesub)
+    _mul = staticmethod(hermemul)
+    _div = staticmethod(hermediv)
+    _pow = staticmethod(hermepow)
+    _val = staticmethod(hermeval)
+    _int = staticmethod(hermeint)
+    _der = staticmethod(hermeder)
+    _fit = staticmethod(hermefit)
+    _line = staticmethod(hermeline)
+    _roots = staticmethod(hermeroots)
+    _fromroots = staticmethod(hermefromroots)
+
+    # Virtual properties
+    domain = np.array(hermedomain)
+    window = np.array(hermedomain)
+    basis_name = 'He'