Robust PDF and DOCX Text Parsing in FastAPI Backend

[Sumedh-6504]

Description

Implement robust binary file parsing in the Python FastAPI backend (backend/parser.py) so that users can upload actual PDF and DOCX files, not just raw text.

User Story

As a LexGuard user,
I want to drop a .pdf or .docx contract into the upload pane,
So that I don't have to manually open, copy, and paste text.

Technical Tasks

• Add standard Python parsing libraries pdfplumber and python-docx to backend/requirements.txt.
• Update /parse route in backend/main.py to handle binary streams under UploadFile.
• Implement secure text extraction in backend/parser.py:
  • Handle multi-page layouts, headers, and footers.
  • Reject corrupted or scanned-only PDFs (return a user-friendly error).
• Sanitize extracted text to strip unnecessary control codes or empty lines before passing to the Detective agent.