pymupdf unstructured pytesseract pdf2image hnswlib pillow_heif pytesseract pikepdf google-cloud-aiplatform google-api-python-client langchain chromadb==0.4.14 pdfplumber pdfminer.six pypdf docx python-docx langchain_google_genai aspose-words plum-dispatch==1.7.4 opencv-python unstructured_pytesseract unstructured_inference langchainhub docx2txt python-pptx groq streamlit openai openpyxl pandas dropbox pyjwt passlib fastapi fastapi-jwt-auth python-decouple firebase-admin httpx-oauth typing nest-asyncio pyngrok langchain-openai langchain-cohere==0.1.1 markdown pydantic[email]