DocQuery

Overview

DocQuery is an AI-powered document platform to extract, analyze, and retrieve information from documents efficiently.

Technology Stack

Backend

• Python FastAPI
• Uvicorn (Application Server)
• Tesseract OCR
• LangChain and Ollama for LLM inference
• RecursiveCharacterTextSplitter for document chunking
• Hugging Face Embedding Model: all-MiniLM-L6-v2
• ChromaDB for vector storage

Frontend

• Flutter

---

Development Environment Setup

1. Create and Activate a Virtual Environment

Create a virtual environment:

python -m venv .venv

Windows

.venv\Scripts\activate.bat

Linux and macOS

source .venv/bin/activate

2. Install Backend Dependencies

Install the required Python packages:

pip install -r requirements.txt

3. Install Tesseract OCR

DocQuery uses Tesseract OCR to extract text from images and scanned documents.

Option A: Automatic Detection (Recommended)

If Tesseract is installed and available in your system PATH, the application will detect it automatically.

Installation Instructions

macOS

brew install tesseract

Linux

sudo apt update -y
sudo apt install tesseract-ocr

Windows

Install Tesseract using the official installer:

https://github.com/UB-Mannheim/tesseract/wiki

After installation, ensure that the Tesseract installation directory is added to your system PATH.

Option B: Manual Configuration

If you prefer to use a specific Tesseract installation, create a .env file inside the backend/ directory and specify the executable path:

tesseract_path="C:\Path\To\Your\tesseract.exe"

Verify Installation

Confirm that Tesseract is installed correctly by running:

tesseract --version

---

API Documentation

For API usage examples and integration details, refer to Documentation

---

License

This project is licensed under the MIT License. See the LICENSE file for details.