Use python-docx to extract docx files (#2654)

langgenius · Mar 7, 2024 · b163545 · b163545
1 parent c0b82f8
commit b163545
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 10 deletions.
diff --git a/api/core/rag/extractor/word_extractor.py b/api/core/rag/extractor/word_extractor.py
@@ -10,7 +10,7 @@
 
 
 class WordExtractor(BaseExtractor):
-    """Load pdf files.
+    """Load docx files.
 
 
     Args:
@@ -46,14 +46,16 @@ def __del__(self) -> None:
 
     def extract(self) -> list[Document]:
         """Load given path as single page."""
-        import docx2txt
-
-        return [
-            Document(
-                page_content=docx2txt.process(self.file_path),
-                metadata={"source": self.file_path},
-            )
-        ]
+        from docx import Document as docx_Document
+
+        document = docx_Document(self.file_path)
+        doc_texts = [paragraph.text for paragraph in document.paragraphs]
+        content = '\n'.join(doc_texts)
+
+        return [Document(
+            page_content=content,
+            metadata={"source": self.file_path},
+        )]
 
     @staticmethod
     def _is_valid_url(url: str) -> bool:

diff --git a/api/requirements.txt b/api/requirements.txt
@@ -32,7 +32,7 @@ celery==5.2.7
 redis~=4.5.4
 openpyxl==3.1.2
 chardet~=5.1.0
-docx2txt==0.8
+python-docx~=1.1.0
 pypdfium2==4.16.0
 resend~=0.7.0
 pyjwt~=2.8.0