Integrate Observability

[bonk1t]

Observability

This PR introduces comprehensive observability support for Agency Swarm through Langchain callbacks, allowing you to track, monitor, and analyze the behavior and performance of your agents.

To use tracking with Langchain callbacks, you need to install the langchain package:

pip install langchain

Langfuse

Langfuse is an observability platform that provides detailed execution tracking with features like tracing, metrics, and debugging tools.

To use Langfuse with Agency Swarm:

1. Install the langfuse package:

pip install langfuse

2. Set the required environment variables:

export LANGFUSE_SECRET_KEY=<your-secret-key>
export LANGFUSE_PUBLIC_KEY=<your-public-key>

You can get your keys from the Langfuse dashboard.

3. Initialize tracking in your code:

from agency_swarm import init_tracking
init_tracking("langfuse")

Configuration options can be passed to the Langfuse callback handler:

init_tracking(
    "langfuse",
    debug=True,
    host="custom-host",
    user_id="user-123",
)

For additional parameters, see the Langfuse documentation.

Local Tracking

This PR also introduces a lightweight local tracker that logs agent activities to a SQLite database.

To use the local tracker:

from agency_swarm import init_tracking
init_tracking("local")

This creates a SQLite database in the current working directory. For a custom database location:

from agency_swarm import init_tracking
init_tracking("local", db_path="path/to/your/database.db")

Additional Features

• Support for token counting and cost tracking
• Streaming response monitoring
• Tool usage tracking
• Extensive refactoring for maintainability