revert

Turns out unstructured doc/docx requires libre office...reverting, fixing but, and remove .doc support.
BBC-Esq · Jan 7, 2024 · 9866ca0 · 9866ca0
1 parent 7810bfa
commit 9866ca0
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 4 deletions.
diff --git a/src/constants.py b/src/constants.py
@@ -273,8 +273,7 @@
 
 DOCUMENT_LOADERS = {
     ".pdf": "PyMuPDFLoader",
-    ".docx": "UnstructuredWordDocumentLoader",
-    ".doc": "UnstructuredWordDocumentLoader",
+    ".docx": "Docx2txtLoader",
     ".txt": "TextLoader",
     ".enex": "EverNoteLoader",
     ".epub": "UnstructuredEPubLoader",

diff --git a/src/document_processor.py b/src/document_processor.py
@@ -7,7 +7,7 @@
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.document_loaders import (
     PyMuPDFLoader,
-    UnstructuredWordDocumentLoader,
+    Docx2txtLoader,
     TextLoader,
     EverNoteLoader,
     UnstructuredEPubLoader,
@@ -67,7 +67,7 @@ def load_single_document(file_path: Path) -> Document:
         elif file_extension == ".epub":
             loader = UnstructuredEPubLoader(str(file_path), mode="single", strategy="fast")
         elif file_extension == ".docx":
-            loader = UnstructuredWordDocumentLoader(str(file_path), mode="single", strategy="fast")
+            loader = Docx2txtLoader(str(file_path))
         elif file_extension == ".rtf":
             loader = UnstructuredRTFLoader(str(file_path), mode="single", strategy="fast")
         elif file_extension == ".odt":