fix(bugs)

anyqa/build_index.py

@@ -41,7 +41,9 @@ def parse_data():
         for chunk in doc_chunks:
             chunk.metadata["name"] = source["name"]
             chunk.metadata["domain"] = source["domain"]
-            chunk.metadata["url"] = source.get("url", None)
+            url = source.get("url", None)
+            if url:
+                chunk.metadata["url"] = source.get("url", None)
             chunk.metadata["page_number"] = chunk.metadata["page"]
             chunk.metadata["short_name"] = chunk.metadata["name"]
             docs.append(chunk)

anyqa/config.py

@@ -2,10 +2,13 @@ import os
 
 # can be used to add metadata to the index, for instance URL
 metadata_by_file_path = {
-    "data/Daoism/Tao_Te_Ching.pdf": { "url": "https://www.with.org/tao_te_ching_en.pdf" },
-    "data/Confucianism/Analects of Confucius.pdf": { "url": "https://chinatxt.sitehost.iu.edu/Analects_of_Confucius_(Eno-2015).pdf" },
+    "data/Daoism/Tao_Te_Ching.pdf": {"url": "https://www.with.org/tao_te_ching_en.pdf"},
+    "data/Confucianism/Analects of Confucius.pdf": {
+        "url": "https://chinatxt.sitehost.iu.edu/Analects_of_Confucius_(Eno-2015).pdf"
+    },
 }
 
+
 def get_domains():
     domains = []
     for root, dirs, files in os.walk("data"):
@@ -26,7 +29,7 @@ def get_sources():
                         "domain": parse_domain(file_path),
                         "name": parse_name(file_path),
                         "file_path": file_path,
-                        **metadata_by_file_path.get(file_path, {})
+                        **metadata_by_file_path.get(file_path, {}),
                     }
                 )
 
@@ -34,8 +37,11 @@ def get_sources():
 
 
 def parse_name(source: str) -> str:
-    return source.split("/")[-1].split(".")[0].replace("_", " ")
+    filename = os.path.basename(source)
+    name, _ = os.path.splitext(filename)
+    return name.replace("_", " ")
 
 
 def parse_domain(source: str) -> str:
-    return source.split("/")[1].replace("_", " ")
+    domain = source.split(os.sep)[1]
+    return domain.replace("_", " ")