This repository hosts the backend for the ReCallMe project, a Flask-based API providing cognitive assistance features. The system is designed to help dementia patients by integrating face identification, task management, and communication functionalities through API endpoints. This backend can be extended with LLM (Language Model) capabilities for enhanced interactions.

• Overview
• Features
• Installation
• Configuration
• File Structure
• Usage
• Endpoints
• Contributing
• License

The ReCallMe Backend Service uses Flask with the Blueprint pattern to provide an extensible API structure. It supports various features, such as face recognition and task management, making it an ideal solution for dementia patient care and cognitive assistance. This service integrates with third-party AI models and is deployable to Google Cloud Platform.

• Face Identification: Recognizes faces to assist with memory reinforcement.
• LLM-based Interaction: Powered by language models for context-based conversation.
• Task Management: Helps users manage and keep track of daily tasks.
• Caregiver Notifications: Notifies caregivers based on specific triggers or requests.
• Modular and Extensible: Utilizes Flask Blueprints for easy extensibility.

1. Clone the Repository:

   git clone https://github.com/morozovdd/ReCallMe-backend-svc.git
   cd ReCallMe-backend-svc

2. Set up Virtual Environment:

   python3 -m venv venv
   source venv/bin/activate

3. Install Dependencies: Use pyproject.toml to install all required libraries.

   pip install -r requirements.txt

Environment variables must be set for specific functionalities:

• GEMINI_URL: URL for Gemini service integration.
• ANTHROPIC_API_KEY: API key for access to Anthropic’s AI models.

These variables can be configured in the .env file for local development.

Here's an overview of the repository's structure:

├── webserver
│   ├── endpoints.py
│   ├── face_prompt_llm.py
│   ├── face_identification.py
│   ├── api_models.py
│   ├── __init__.py
├── extensions.py
├── app.py
└── Dockerfile

flask run

The application will be available at http://127.0.0.1:5000.

To deploy using Docker, build and run the container:

docker build -t recallme-backend .
docker run -p 5000:5000 recallme-backend

Utilize the provided deploy.yml GitHub Actions workflow, configuring your GCP credentials and Artifact Registry.

The following are key endpoints provided by the backend. Each endpoint’s implementation can be found in webserver/endpoints.py:

• /identify_face - POST: Processes image input to identify and verify known faces.
• /generate_response - POST: Uses LLM to generate context-based responses based on prompts.
• /tasks - GET/POST/DELETE: Manage user tasks, including creation, retrieval, and deletion.

Swagger documentation is available to explore the endpoints interactively. Access it via /swagger when the app is running.

1. Fork the repository.
2. Create a new branch.
3. Commit your changes and push them to the branch.
4. Open a pull request to contribute.

This project is licensed under the MIT License.

For further details, please visit our GitHub repository.

https://github.com/morozovdd/ReCallMe-backend-svc.git