initialization

.env

@@ -0,0 +1,3 @@
+API_KEY=123456
+TYPHOON_OCR_API_KEY=sk-D1sMp4fivWSKNhnydFOvqFwzdxqK7OIsVoLHn6rN3komSG3L
+TYPHOON_BASE_URL=https://api.opentyphoon.ai/v1

app.py

@@ -0,0 +1,122 @@
+import os
+from fastapi import FastAPI, HTTPException, Header, UploadFile, File
+from fastapi.middleware.cors import CORSMiddleware
+import gradio as gr
+from typhoon_ocr import ocr_document
+from pdf2image import convert_from_bytes
+from PIL import Image
+import re
+from dotenv import load_dotenv
+
+# --- Load environment variables from .env ---
+load_dotenv()
+
+# --- Config ---
+API_KEY = os.getenv("API_KEY")
+TYPHOON_API_KEY = os.getenv("TYPHOON_OCR_API_KEY")
+TYPHOON_BASE_URL = os.getenv("TYPHOON_BASE_URL", "https://api.opentyphoon.ai/v1")
+
+# --- FastAPI App ---
+app = FastAPI()
+
+# CORS (optional for public usage)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+def extract_fields_regex(text: str) -> dict:
+    # Preprocess text
+    text = re.sub(r"<.*?>", "", text)                  # Strip tags
+    text = re.sub(r"\n+", "\n", text)                  # Collapse newlines
+    text = re.sub(r"\s{2,}", " ", text)                # Collapse multiple spaces
+    text = re.sub(r"\t+", " ", text)
+
+    patterns = {
+        "เลขที่ผู้เสียภาษี": r"(?:TAX\s*ID|เลขที่ผู้เสียภาษี)[\s:\-\.]*([\d]{10,13})",
+        
+        # Updated pattern for correct tax invoice number
+        "เลขที่ใบกำกับภาษี": r"(?:TAX\s*INV\.?|เลขที่ใบกำกับภาษี|ใบกำกับ)[\s:\-\.]*([\d]{8,20})",
+        "จำนวนเงิน": r"(?:AMOUNT\s*THB|จำนวนเงิน|รวมเงิน)[\s:\-\.]*([\d,]+\.\d{2})",
+        "ราคาต่อลิตร": r"(?:Baht\/Litr\.?|Bath\/Ltr\.?|ราคาต่อลิตร|ราคา\/ลิตร|ราคาน้ำมัน)[\s:\-\.]*([\d,]+\.\d{2})",
+        "ลิตร": r"(?:Ltr\.?|Ltrs?\.?|ลิตร)[\s:\-\.]*([\d,]+\.\d{3})",
+        "ภาษีมูลค่าเพิ่ม": r"(?:VAT|ภาษีมูลค่าเพิ่ม)[\s:\-\.]*([\d,]+\.\d{2})",
+        "ยอดรวม": r"(?:TOTAL\s*THB|ยอดรวม|รวมทั้งสิ้น|รวมเงินทั้งสิ้น)[\s:\-\.]*([\d,]+\.\d{2})",
+        "วันที่": r"(?:DATE|วันที่|ออกใบกำกับวันที่)[\s:\-\.]*([\d]{2}/[\d]{2}/[\d]{2,4})",
+    }
+
+    results = {}
+    for field, pattern in patterns.items():
+        match = re.search(pattern, text, re.IGNORECASE)
+        results[field] = match.group(1).strip() if match else None
+    
+    # Optional fallback if regex fails
+    # if not results["เลขที่ใบกำกับภาษี"]:
+    #     match = re.search(r"TAX\s*INV\.?\s*</td>\s*<td>\s*([\d\-]+)", text, re.IGNORECASE)
+    # if match:
+    #     results["เลขที่ใบกำกับภาษี"] = match.group(1).strip()
+
+    return results
+
+
+
+def pdf_to_image(file_bytes: bytes) -> Image.Image:
+    images = convert_from_bytes(file_bytes)
+    return images[0]  # First page only
+
+# --- API Endpoint ---
+@app.post("/api/ocr_receipt")
+async def ocr_receipt(
+    file: UploadFile = File(...),
+    x_api_key: str | None = Header(None),
+):
+    if API_KEY and x_api_key != API_KEY:
+        raise HTTPException(status_code=401, detail="Invalid API key")
+
+    content = await file.read()
+
+    try:
+        # Handle PDF and image
+        if file.filename.lower().endswith(".pdf"):
+            image = pdf_to_image(content)
+            raw_output = ocr_document(image, task_type="structure")
+        else:
+            raw_output = ocr_document(content, task_type="structure")
+
+        text = raw_output if isinstance(raw_output, str) else raw_output.get("text", "")
+        extracted = extract_fields_regex(text)
+
+        return {
+            "raw_ocr": text,
+            "extracted_fields": extracted,
+        }
+
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+# --- Gradio UI ---
+def gradio_interface(image_path: str | Image.Image):
+    if isinstance(image_path, str) and image_path.lower().endswith(".pdf"):
+        with open(image_path, "rb") as f:
+            image = pdf_to_image(f.read())
+    else:
+        image = image_path
+
+    raw = ocr_document(image, task_type="structure")
+    text = raw if isinstance(raw, str) else raw.get("text", "")
+    extracted = extract_fields_regex(text)
+    return text, extracted
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🧾 Thai Receipt OCR")
+    with gr.Row():
+        img = gr.Image(type="filepath", label="Upload receipt image or PDF")
+        out_text = gr.Textbox(label="OCR Text", lines=10)
+        out_fields = gr.JSON(label="Extracted Fields")
+    btn = gr.Button("Run OCR")
+    btn.click(fn=gradio_interface, inputs=img, outputs=[out_text, out_fields])
+
+# --- Mount Gradio on FastAPI ---
+app = gr.mount_gradio_app(app, demo, path="/ui")

dockerfile

@@ -0,0 +1,19 @@
+FROM python:3.12-slim
+
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+
+WORKDIR /app
+
+# System dependencies
+RUN apt-get update && apt-get install -y \
+    poppler-utils \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt .
+RUN pip install --upgrade pip && pip install -r requirements.txt
+
+COPY . .
+
+EXPOSE 7860
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

requirements.txt

@@ -0,0 +1,8 @@
+typhoon-ocr
+fastapi
+gradio
+uvicorn
+python-multipart
+pdf2image
+Pillow
+python-dotenv