Update pipelines/utils.py

pipelines/utils.py

@@ -1,7 +1,8 @@
+# pipelines/utils.py
 import io
-import decx
+import docx
 
-def detect_filetype(filename: str, file_btes: bytes) -> str:
+def detect_filetype(filename: str, file_bytes: bytes) -> str:
     fname = (filename or "").lower()
     if fname.endswith(".pdf"):
         return "pdf"
@@ -10,18 +11,18 @@ def detect_filetype(filename: str, file_btes: bytes) -> str:
     if fname.endswith(".docx"):
         return "docx"
     if fname.endswith(".txt"):
-        return "text"
+        return "txt"
     if file_bytes[:4] == b"%PDF":
         return "pdf"
     return "unknown"
 
-def load_text(filetype: str, file_bytes: bytes) -> str:
-    if filestype == "docx":
+def load_doc_text(filetype: str, file_bytes: bytes) -> str:
+    if filetype == "docx":
         f = io.BytesIO(file_bytes)
-        doc = decx.Document(f)
-        return "\n".join(p.text for p in doc.paragraphs)
-    elif filetype == "text":
+        doc = docx.Document(f)
+        return "\n".join([p.text for p in doc.paragraphs])
+    elif filetype == "txt":
         return file_bytes.decode("utf-8", errors="ignore")
     else:
-        #　それ以外は上位でOpenAI側へルーティング
+        # それ以外は上位でOpenAI側へルーティング
         return ""