64429b19e6
- Added `trillium.py` for searching and creating notes with Trillium's ETAPI. - Implemented `search_notes` and `create_note` functions with appropriate error handling and validation. feat: Add web search functionality using DuckDuckGo - Introduced `web_search.py` for performing web searches without API keys. - Implemented `search_web` function with result handling and validation. feat: Create provider-agnostic function caller for iterative tool calling - Developed `function_caller.py` to manage LLM interactions with tools. - Implemented iterative calling logic with error handling and tool execution. feat: Establish a tool registry for managing available tools - Created `registry.py` to define and manage tool availability and execution. - Integrated feature flags for enabling/disabling tools based on environment variables. feat: Implement event streaming for tool calling processes - Added `stream_events.py` to manage Server-Sent Events (SSE) for tool calling. - Enabled real-time updates during tool execution for enhanced user experience. test: Add tests for tool calling system components - Created `test_tools.py` to validate functionality of code execution, web search, and tool registry. - Implemented asynchronous tests to ensure proper execution and result handling. chore: Add Dockerfile for sandbox environment setup - Created `Dockerfile` to set up a Python environment with necessary dependencies for code execution. chore: Add debug regex script for testing XML parsing - Introduced `debug_regex.py` to validate regex patterns against XML tool calls. chore: Add HTML template for displaying thinking stream events - Created `test_thinking_stream.html` for visualizing tool calling events in a user-friendly format. test: Add tests for OllamaAdapter XML parsing - Developed `test_ollama_parser.py` to validate XML parsing with various test cases, including malformed XML.
80 lines
2.4 KiB
Python
80 lines
2.4 KiB
Python
"""
|
|
Base adapter interface for provider-agnostic tool calling.
|
|
|
|
This module defines the abstract base class that all LLM provider adapters
|
|
must implement to support tool calling in Lyra.
|
|
"""
|
|
|
|
from abc import ABC, abstractmethod
|
|
from typing import Dict, List, Optional
|
|
|
|
|
|
class ToolAdapter(ABC):
|
|
"""Base class for provider-specific tool adapters.
|
|
|
|
Each LLM provider (OpenAI, Ollama, llama.cpp, etc.) has its own
|
|
way of handling tool calls. This adapter pattern allows Lyra to
|
|
support tools across all providers with a unified interface.
|
|
"""
|
|
|
|
@abstractmethod
|
|
async def prepare_request(
|
|
self,
|
|
messages: List[Dict],
|
|
tools: List[Dict],
|
|
tool_choice: Optional[str] = None
|
|
) -> Dict:
|
|
"""Convert Lyra tool definitions to provider-specific format.
|
|
|
|
Args:
|
|
messages: Conversation history in OpenAI format
|
|
tools: List of Lyra tool definitions (provider-agnostic)
|
|
tool_choice: Optional tool forcing ("auto", "required", "none")
|
|
|
|
Returns:
|
|
dict: Provider-specific request payload ready to send to LLM
|
|
"""
|
|
pass
|
|
|
|
@abstractmethod
|
|
async def parse_response(self, response) -> Dict:
|
|
"""Extract tool calls from provider response.
|
|
|
|
Args:
|
|
response: Raw provider response (format varies by provider)
|
|
|
|
Returns:
|
|
dict: Standardized response in Lyra format:
|
|
{
|
|
"content": str, # Assistant's text response
|
|
"tool_calls": [ # List of tool calls or None
|
|
{
|
|
"id": str, # Unique call ID
|
|
"name": str, # Tool name
|
|
"arguments": dict # Tool arguments
|
|
}
|
|
] or None
|
|
}
|
|
"""
|
|
pass
|
|
|
|
@abstractmethod
|
|
def format_tool_result(
|
|
self,
|
|
tool_call_id: str,
|
|
tool_name: str,
|
|
result: Dict
|
|
) -> Dict:
|
|
"""Format tool execution result for next LLM call.
|
|
|
|
Args:
|
|
tool_call_id: ID from the original tool call
|
|
tool_name: Name of the executed tool
|
|
result: Tool execution result dictionary
|
|
|
|
Returns:
|
|
dict: Message object to append to conversation
|
|
(format varies by provider)
|
|
"""
|
|
pass
|