AI Document Analyzer

An intelligent document processing service that extracts text from PDF and DOCX files, analyzes content using Large Language Models (LLMs) via OpenRouter, and provides structured metadata extraction.

Features

• Document Upload: Support for PDF and DOCX files.
• Secure Storage: Files are stored securely in Minio/S3.
• Text Extraction: Automatic extraction of text content from uploaded documents.
• AI Analysis:
  • Concise Summaries.
  • Document Type Detection (Invoice, CV, Report, etc.).
  • Metadata Extraction (Date, Sender, Amount, etc.).
• REST API: Built with FastAPI for high performance and easy integration.

Tech Stack

• Backend: FastAPI
• Database: PostgreSQL
• Storage: Minio (S3 Compatible)
• AI/LLM: OpenRouter (GPT-4o-mini or compatible)
• ORM: SQLAlchemy
• Dependencies: boto3, pdfplumber, python-docx, python-multipart

Prerequisites

• Python 3.9+
• PostgreSQL
• Minio Server (or S3 access)
• OpenRouter API Key

Installation

1. Clone the repository

   git clone https://github.com/danielzfega/document-analyzer
   cd document-analyzer

2. Create a virtual environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

4. Configuration Create a .env file in the root directory (ensure it is UTF-8 encoded):

   DATABASE_URL=postgresql://user:password@localhost:5432/document_analyzer
   
   MINIO_ENDPOINT=http://localhost:9000
   MINIO_ACCESS_KEY=minioadmin
   MINIO_SECRET_KEY=minioadmin123
   MINIO_BUCKET_NAME=document-analyzer
   
   OPENROUTER_API_KEY=your_openrouter_api_key_here

5. Run the Application

   uvicorn main:app --reload

API Usage

1. Upload Document

POST /documents/upload

• Body: multipart/form-data with file field.
• Response:

  {
      "id": "1",
      "file_name": "resume.pdf"
  }

2. Analyze Document

POST /documents/{id}/analyze

• Response:

  {
      "message": "Analysis complete"
  }

3. Get Document Details

GET /documents/{id}

• Response:

  {
      "id": "1",
      "file_name": "resume.pdf",
      "text": "Extracted text content...",
      "summary": "This is a resume for...",
      "detected_type": "CV",
      "attributes": {
          "name": "John Doe",
          "email": "john@example.com"
      }
  }

Testing

Run the test suite using pytest:

pytest

Note: Ensure your .env is configured for testing (e.g., using a test database or SQLite).