This repository contains the backend of the Cogito project, which consists of several main components for handling analytical models, LLM model integration, and the main API server.

cogito-backend/
├── analytical-setup/           # Analytical model setup (Python)
│   └── fallacy_detector_model/ # Model
│   ├── .gitignore
│   └── app.py                  # Entry point
│
├── llama-setup/               # Integration with LLM (Node.js/Express)
│   ├── node_modules/
│   ├── src/                    # Source code for LLM wrapper/handler
│   ├── .env
│   ├── .gitignore
│   ├── package.json
│   └── package-lock.json
│
├── server-api/                # Main backend (Node.js/Express)
│   ├── migrations/             # Contains database migration scripts
│   ├── node_modules/
│   ├── scripts/                # Additional scripts
│   ├── src/                    # Main application code
│   ├── .env
│   ├── .env.test
│   ├── .gitignore
│   ├── package.json
│   └── package-lock.json

• Python >= 3.8
• Libraries such as flask, transformers, etc. (More details in ./fallacy_detector_model/requirements.py)

• Node.js >= 18.x
• PostgreSQL
• npm, node-pg-migrate, and other

Before running the project, make sure the model is placed in the following directory:

• LLM Model for llama-setup/models: Download on Google Drive
• Fallacy Detection Model for analytical-setup/fallacy_detector_model: Download on Google Drive

Please extract the model file to the appropriate directory after downloading it.

cd analytical-setup
pip install -r ./fallacy_detector_model/requirements.py
python app.py

cd llama-setup
npm install
npm start

cd server-api
npm install
npm run migrate up
npm start

• Ensure that the .env file is available in each component (llama-setup and server-api).
• For testing, you can use the .env.test file in server-api.
• The structure may change depending on the final implementation.

MIT License

Translated with DeepL.com (free version)