Add support for dynamic flows

[markbackman]

This PR extends the FlowManager to support dynamic flows in addition to static flows.

From the README:

• Static Flows: When your conversation structure is known upfront and follows predefined paths. Perfect for customer service scripts, intake forms, or guided experiences.
• Dynamic Flows: When conversation paths need to be determined at runtime based on user input, external data, or business logic. Ideal for personalized experiences or complex decision trees.

For static flows, simplified function registration. All functions are now registered to the LLMs automatically at runtime.

Also, added tests for the new functionality.

This change also simplifies the API:

Static flows:

• Define the flow_config as a JSON blob using compliant LLM formats
• Create the FlowManager instance: flow_manager = FlowManager(task, llm, tts, flow_config=flow_config)
• Optionally, initialize the FlowManager with initial messages: await flow_manager.initialize(messages)

Dynamic flows:

• Define your transition callback, which is responsible for transitioning node to node
• Create the FlowManager instance: flow_manager = FlowManager(task, llm, tts, transition_callback=handle_insurance_transition)
• Optionally, initialize the FlowManager with initial messages: await flow_manager.initialize(messages)

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
pipecat-flows	✅ Ready (Inspect)	Visit Preview	Dec 3, 2024 2:32am