version: '3.4'

• This specifies the version of Docker Compose you are using. Ensure it matches or is compatible with your Docker setup.

services:
    openwebui:
        image: ghcr.io/open-webui/open-webui:main

• Uses the official OpenWebUI image which provides a web-based interface for interacting with AI models.
• restart: unless-stopped ensures the service restarts if interrupted (e.g., by Ctrl+C).
• depends_on: - ollamadeepseek specifies that OpenWebUI depends on the Ollama DeepSeek service to function properly. If this service fails to start, OpenWebUI will not start either.
• Ports: Maps port 8333 (used by OpenWebUI) to port 8080 (used by your application).
• Volumes: mounts your project's open-webui directory to the host's /app/backend/data location. This is where OpenWebUI stores its configuration files and data.

    ollamadeepseek: 
        build:
            context: .
            dockerfile: Dockerfile

• This service runs the Ollama DeepSeek server, which allows you to interact with an Ollama model through a web interface.
• restart: unless-stopped ensures the service restarts if interrupted.

        environment:
            OLLAMA_BASE_URL: ollamadeepseek

• Sets the base URL for Ollama services (e.g., for accessing tokens or configuration files).

        ports:
          - './open-webui:/app/backend/data'

• OpenWebUI Port Mapping: Maps port 8333 on your local machine to port 8080 of the OpenWebUI service. This allows data and requests to flow between your application (running on port 8080) and OpenWebUI.
• Volumes: mounts your project's open-webui directory to the host's /app/backend/data location.

        volumes:
          - './open-webui:/app/backend/data'

• OpenWebUI Volumes: Specifies that files in your project's open-webui directory should be mounted to the host's /app/backend/data location. This is where OpenWebUI stores its configuration and data.

        environment:
            OLLAMA_HOST: 0.0.0.0
            OLLAMA_PORT: 8080

• OLLAMA_HOST: Specifies the hostname used by the Ollama service.
• OLLAMA_PORT: Specifies the port used by the Ollama service.

1. OpenWebUI runs first and maps its ports to your application's ports.
2. Ollama DeepSeek runs next, using the environment variables (e.g., OLLAMA_HOST and OLLAMA_PORT) to connect to Ollama services.
3. OpenWebUI communicates with Ollama through JSON-RPC requests on port 8080.

• OpenWebUI: Runs and provides a web interface for interacting with your model (e.g., entering prompts in the chat interface).
• Ollama DeepSeek: Allows you to query an external Ollama server using the OpenWebUI interface.
• Volumes: Ensure that configuration files and data are persistently available between services.

1. Ensure all services are running before starting OpenWebUI.
2. Verify that ports 8333 (OpenWebUI) and 8080 (Ollama DeepSeek) are correctly mapped in your host machine.
3. Check that environment variables like OLLAMA_HOST and OLLAMA_PORT match those configured in Ollama services.

This configuration provides a simple way to integrate OpenWebUI with an Ollama DeepSeek service using Docker Compose, allowing you to work with AI models through a web interface.