[feat] MLflow support

[tdadela]

This pull request introduces the following updates to the project's Docker setup:

1. Addition of MLflow Service:

   • Configures a mlflow service in the file. docker-compose.yml
   • Uses a custom to set up MLflow with specific dependencies. Dockerfile.mlflow
   • Exposes port 5000 for accessing the MLflow server UI.

2. Integration with Reagent AI Service:

   • Configures a reagentai service in the file. docker-compose.yml
   • Establishes proper communication between the reagentai and mlflow services using the . mlflow-network
   • Injects the environment variable for service integration. MLFLOW_TRACKING_URI

3. Custom Network Configuration:

   • Defines a shared custom network for seamless inter-service communication. mlflow-network

[Copilot]

Pull Request Overview

This PR adds MLflow support to the project by integrating a dedicated MLflow service into the Docker setup and wiring experiment tracking into both the retrosynthesis module and the LLM agent. Key changes include:

• Adding MLflowTracker usage in retrosynthesis and client code for logging parameters, metrics, and artifacts.
• Integrating MLflow support into the main agent and updating Docker/Docker Compose to expose the MLflow UI.
• Updating project documentation and dependencies to reflect the new MLflow integration.

Reviewed Changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/reagentai/tools/retrosynthesis.py	Adds MLflowTracker initialization and logging around retrosynthesis workflow.
src/reagentai/common/mlflow_tracking.py	Implements the MLflowTracker abstraction for experiment tracking.
src/reagentai/common/client.py	Injects MLflowTracker into LLMClient and logs interactions and model changes.
src/reagentai/agents/main/main_agent.py	Integrates MLflow tracking with the main agent by sharing a tracker instance.
pyproject.toml	Adds the MLflow dependency.
docker-compose.yml	Defines the new MLflow service with appropriate port and network configuration.
README.md	Updates documentation with usage instructions for MLflow including Docker commands.
Dockerfile.mlflow	Provides Docker instructions to run the MLflow server.

Comments suppressed due to low confidence (1)

src/reagentai/common/client.py:119

• [nitpick] The commented-out code for generating the MLflow run name could be removed to improve code clarity if it is no longer needed.

// commented out hash-based run name generation

[tdadela]

Closed, functionality in #14