🐳Docker Setup

Kallia provides comprehensive Docker support for both development and production deployments. This guide covers Docker setup, configuration, and deployment strategies based on the actual Docker configurations.

Overview

Kallia consists of two main services:

• Core API Service: FastAPI-based document processing service

• Playground: Chainlit-powered interactive chat interface

Each service can be deployed independently using Docker and Docker Compose.

Prerequisites

• Docker 20.10 or higher

• Docker Compose 2.0 or higher

• Ollama running locally (for AI model inference)

Quick Start

1. Core API Service

Navigate to the core directory and start the service:

cd kallia/core
docker-compose up -d

The API will be available at http://localhost:8000

2. Playground Service

Navigate to the playground directory and start the service:

cd kallia/playground
docker-compose up -d

The playground will be available at http://localhost:8000

Core API Service

Docker Compose Configuration

The core service uses the following docker-compose.yml:

name: kallia
services:
  core:
    image: overheatsystem/kallia:0.1.6
    container_name: core
    ports:
      - "8000:80"
    environment:
      - KALLIA_PROVIDER_API_KEY=ollama
      - KALLIA_PROVIDER_BASE_URL=http://localhost:11434/v1
      - KALLIA_PROVIDER_MODEL=qwen2.5vl:32b

Dockerfile

The core service uses this Dockerfile:

FROM python:3.11-slim
WORKDIR /code
COPY ./requirements.txt /code/requirements.txt
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
COPY ./kallia_core /code/kallia_core
CMD ["fastapi", "run", "kallia_core/main.py", "--port", "80"]

Environment Configuration

Create a .env file in the core directory:

# Copy the example file
cp .env.example .env

The .env.example file contains:

KALLIA_PROVIDER_API_KEY=ollama
KALLIA_PROVIDER_BASE_URL=http://localhost:11434/v1
KALLIA_PROVIDER_MODEL=qwen2.5vl:32b

Playground Service

Docker Compose Configuration

The playground service uses the following docker-compose.yml:

name: kallia
services:
  playground:
    build: .
    container_name: playground
    ports:
      - "8000:80"
    environment:
      - KALLIA_PROVIDER_API_KEY=ollama
      - KALLIA_PROVIDER_BASE_URL=http://localhost:11434/v1
      - KALLIA_PROVIDER_MODEL=qwen2.5vl:32b

Dockerfile

The playground service uses this Dockerfile:

FROM python:3.11-slim
WORKDIR /code
COPY ./requirements.txt /code/requirements.txt
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
COPY ./kallia_playground /code/kallia_playground
CMD ["chainlit", "run", "kallia_playground/main.py", "--host", "0.0.0.0", "--port", "80"]

Environment Configuration

Create a .env file in the playground directory:

# Copy the example file
cp .env.example .env

The .env.example file contains:

KALLIA_PROVIDER_API_KEY=ollama
KALLIA_PROVIDER_BASE_URL=http://localhost:11434/v1
KALLIA_PROVIDER_MODEL=qwen2.5vl:32b

Environment Variables

Core Configuration

• KALLIA_PROVIDER_API_KEY: API key for the provider (set to "ollama" for local Ollama)

• KALLIA_PROVIDER_BASE_URL: Base URL for the AI provider API

• KALLIA_PROVIDER_MODEL: Model name to use for processing

Custom Configuration

You can modify the environment variables to use different AI providers:

# For OpenAI
KALLIA_PROVIDER_API_KEY=your-openai-api-key
KALLIA_PROVIDER_BASE_URL=https://api.openai.com/v1
KALLIA_PROVIDER_MODEL=gpt-4-vision-preview

# For Azure OpenAI
KALLIA_PROVIDER_API_KEY=your-azure-api-key
KALLIA_PROVIDER_BASE_URL=https://your-resource.openai.azure.com/
KALLIA_PROVIDER_MODEL=gpt-4-vision

Docker Hub Images

Pre-built images are available on Docker Hub:

• Core Service: overheatsystem/kallia:0.1.6

• Latest: overheatsystem/kallia:latest

# Pull the latest image
docker pull overheatsystem/kallia:latest

# Run with Docker Hub image
docker run -d \
  --name kallia-core \
  -p 8000:80 \
  -e KALLIA_PROVIDER_API_KEY=ollama \
  -e KALLIA_PROVIDER_BASE_URL=http://localhost:11434/v1 \
  -e KALLIA_PROVIDER_MODEL=qwen2.5vl:32b \
  overheatsystem/kallia:latest

Network Configuration

Docker Network

Both services use the default Docker network. For custom networking:

networks:
  kallia-network:
    driver: bridge

services:
  core:
    networks:
      - kallia-network

Connecting Services

If running both core and playground together:

version: "3.8"
services:
  core:
    image: overheatsystem/kallia:0.1.6
    ports:
      - "8000:80"
    networks:
      - kallia-network

  playground:
    build: ./playground
    ports:
      - "8001:80"
    networks:
      - kallia-network
    depends_on:
      - core

networks:
  kallia-network:
    driver: bridge

Security Considerations

Environment Security

• Use .env files for sensitive configuration

• Never commit .env files to version control

• Use Docker secrets for production deployments

• Regularly update base images and dependencies

Network Security

• Use custom networks to isolate services

• Implement reverse proxy for production

• Enable HTTPS for external access

• Restrict container capabilities

Next Steps

• Explore configuration options for advanced setup

• Learn about REST API endpoints

• Check use cases for practical implementations

• Review core library for direct integration

Support

For Docker-related issues:

• GitHub Issues: https://github.com/kallia-project/kallia/issues

• Docker Hub: https://hub.docker.com/r/overheatsystem/kallia

• Email: [e-mail]