✨ add xlsx and xls parser (QuivrHQ#997)

backend/parsers/xlsx.py

@@ -0,0 +1,19 @@
+from langchain.document_loaders import UnstructuredExcelLoader
+from models.files import File
+
+from .common import process_file
+
+
+def process_xlsx(
+    file: File,
+    enable_summarization,
+    brain_id,
+    user_openai_api_key,
+):
+    return process_file(
+        file=file,
+        loader_class=UnstructuredExcelLoader,
+        enable_summarization=enable_summarization,
+        brain_id=brain_id,
+        user_openai_api_key=user_openai_api_key,
+    )

backend/requirements.txt

@@ -7,7 +7,7 @@ pypdf==3.9.0
 StrEnum==0.4.15
 supabase==1.0.3
 tiktoken==0.4.0
-unstructured==0.6.5
+unstructured==0.6.7
 fastapi==0.95.2
 python-multipart==0.0.6
 uvicorn==0.22.0
@@ -26,4 +26,5 @@ psycopg2-binary==2.9.6
 sqlalchemy==2.0.19
 html5lib==1.1
 bs4==0.0.1
-newspaper3k
+newspaper3k
+xlrd==1.0.0

backend/utils/processors.py

@@ -11,6 +11,7 @@
 from parsers.pdf import process_pdf
 from parsers.powerpoint import process_powerpoint
 from parsers.txt import process_txt
+from parsers.xlsx import process_xlsx
 
 file_processors = {
     ".txt": process_txt,
@@ -29,6 +30,8 @@
     ".pptx": process_powerpoint,
     ".docx": process_docx,
     ".odt": process_odt,
+    ".xlsx": process_xlsx,
+    ".xls": process_xlsx,
     ".epub": process_epub,
     ".ipynb": process_ipnyb,
 }