This directory contains a series of examples demonstrating the streaming capabilities of the A2A (Agent-to-Agent) library. Each example builds on concepts from previous examples to showcase different aspects of streaming functionality.

Before running these examples, make sure you have the required dependencies installed:

# Core requirements
pip install python-a2a

# Additional requirements for UI examples (05_streaming_ui_integration.py)
pip install colorama tqdm flask

# Additional requirements for distributed examples (06_distributed_streaming.py)
pip install aiohttp tqdm

0. basic_streaming.py - Minimal streaming implementation

   • Simplest possible streaming server (only core essentials)
   • Minimal streaming client with proper error handling
   • Proper dictionary chunk handling
   • Ideal starting point for understanding streaming

1. 01_basic_streaming.py - Comprehensive introduction to streaming basics

   • Simple streaming server implementation
   • Basic client for consuming streams
   • Comparison between streamed and non-streamed responses
   • Simulated thinking delays and natural language chunking

2. 02_advanced_streaming.py - Advanced streaming techniques

   • Metrics tracking for streaming performance
   • Different chunking strategies (sentence, word, paragraph)
   • Performance visualization and analytics
   • Stream buffering and handling

3. 03_streaming_llm_integration.py - Integrating with LLM providers

   • Bridging LLM provider streaming APIs with A2A
   • Handling different provider formats (OpenAI, Anthropic, Bedrock)
   • Transformer pipelines for content processing
   • Error handling and fallback mechanisms

4. 04_task_based_streaming.py - Structured task-based streaming

   • Task state transitions (WAITING, COMPLETED, etc.)
   • Progress tracking with step-by-step updates
   • Artifact generation during streaming
   • Partial and final results as structured artifacts

5. 05_streaming_ui_integration.py - Integrating streaming with UIs

   • CLI-based streaming visualization
   • Web interface using Server-Sent Events (SSE)
   • Interactive streaming controls (pause, resume, cancel)
   • Real-time progress visualization

6. 06_distributed_streaming.py - Distributed streaming architecture

   • Multiple streaming servers with load balancing
   • Stream aggregation from multiple sources
   • Fault tolerance and failover strategies
   • Performance monitoring and metrics

Each example is designed to be run independently. Simply use Python to execute the desired example:

python 01_basic_streaming.py

For the more advanced examples, make sure to install the dependencies first:

# For UI integration example
python 05_streaming_ui_integration.py

# For distributed streaming example
python 06_distributed_streaming.py

1. Missing Module Errors: Make sure to install all required dependencies as listed in the Dependencies section.

2. "Message.init() missing required argument 'role'": This error occurs because the Message constructor requires a role parameter. If you're creating your own code based on these examples, always include role=MessageRole.USER or role=MessageRole.AGENT when creating Message objects.

3. Type Errors with Dictionary Chunks: When handling streamed chunks, remember they can sometimes be dictionaries rather than simple strings. Handle this properly by checking types and extracting content when needed.

4. Port Conflicts: If a server won't start due to a port being in use, modify the port number in the code or wait a few moments for the previous server process to fully terminate.

The examples in this directory demonstrate these key streaming concepts:

1. Stream Creation: How to implement streaming on the server side using async generators
2. Stream Consumption: How to consume streams on the client side
3. Chunking Strategies: Different approaches to breaking content into streamable chunks
4. Error Handling: Robust error handling for streaming scenarios
5. Performance Metrics: Tracking and analyzing streaming performance
6. UI Integration: Techniques for rendering streaming content in user interfaces
7. LLM Integration: Connecting A2A streaming with LLM providers
8. Distributed Architecture: Advanced patterns for scalable streaming systems

If you encounter issues with streaming, the core_streaming_test.py script provides a simple test harness to verify that core streaming functionality is working correctly.

For more information on A2A streaming, refer to:

• A2A Example Documentation
• The implementation in python_a2a/server/http.py and python_a2a/client/streaming.py
• Official Documentation for the full API reference