streamlit pymupdf==1.24.9 PyPDF2 pdfplumber pytesseract pdf2image pypdf2 Pillow faiss-cpu sentence-transformers langchain>=0.2.0 langchain-community>=0.2.0 huggingface_hub together python-docx pypandoc pypandoc-binary numpy pandas pytesseract poppler-utils ghostscript pandoc camelot-py accelerate transformers openai pdf2docx python-dotenv tiktoken