Skip to content

USER_GUIDE.md

Latest commit

 

History

History
69 lines (54 loc) · 3.24 KB

USER_GUIDE.md

File metadata and controls

69 lines (54 loc) · 3.24 KB

📘 User Guide: Enterprise Document Analyzer

Introduction

The Enterprise Document Analyzer is a premium, AI-powered platform designed to extract, index, and query information from complex PDF documents. Whether dealing with digital text, scanned invoices, or handwritten notes, the system uses a hybrid approach combining traditional OCR and Google's Gemini Vision to ensure high-accuracy data retrieval.

🧭 Navigation

The application uses a streamlined 3-view interface. You can switch between views using the sidebar or context buttons.

1. Landing Page (Ingestion)

The starting point for all new workflows. Landing Page

Features:

  • Drag & Drop Upload: Support for PDF files.
  • Previewer: Instant visual confirmation of the file.
  • Extraction Modes:
    • OCR (Default): Best for standard digital PDFs or clean scans. Uses Docling and Tesseract.
    • GEMINI: Advanced AI Vision mode. Best for handwriting, poor scans, or complex layouts where OCR fails.
  • Ingest Button: Starts the pipeline. This processes the document, extracts tables/images, creates vector chunks, and indexes them.

2. Dashboard

The main command center. Dashboard

A. 💬 Chat Interface

Interact with your indexed knowledge base using natural language.

  • History: Switch between multiple conversation threads on the left.
  • Chat Window: Ask questions like "What is the total revenue in Q3?" or "Summarize the safety protocols."
  • Citations:
    • Quick Sources: Small pills showing page numbers.
    • Visual References: The right-hand panel shows the exact text chunk and a visual snapshot of the source page with the relevant section highlighted.

B. 👁 Inspector

A deep-dive tool to verify extraction quality. Inspector

  • Images Tab: View all extracted figures, diagrams, and scanned pages. Click to zoom.
  • Tables Tab: View extracted tables converted into interactive DataFrames.
  • Chunks Tab: Audit the text chunks created during ingestion.
  • Full Doc: View the original PDF in an embedded frame.

C. 💾 Database Management

Manage your knowledge repository. Database

  • Document List: See all indexed files.
  • Activate: Switch the "Current Context" to a specific document for the Inspector.
  • Delete: Remove a document and its vectors from the system.
  • Nuclear Reset: A "Danger Zone" option to wipe the entire database clean.

🧠 Core Capabilities

Hybrid Extraction Strategy

Feature OCR Mode Gemini Vision Mode
Speed Fast ⚡ Slower (API dependant) 🐢
Cost Low (Local compute) Higher (Token costs)
Best For Contracts, Reports, Clean Forms Handwritten notes, Faded receipts, Charts
Table Extraction Heuristic/Structural Visual Understanding

Intelligent RAG (Retrieval Augmented Generation)

  1. Query Analysis: The system understands your question.
  2. Vector Search: It retrieves the top 5 most relevant text chunks from ChromaDB.
  3. Synthesis: Google Gemini summarizes the chunks into a coherent answer.
  4. Citation: Every answer is backed by verifiable source evidence.