arcade-mcp/libs/arcade-evals/arcade_evals/_evalsuite/_tracks.py
jottakka 98fad93d21
Adding MCP Servers supports to Arcade Evals (#689)
# MCP Server Tool Evaluation Support

## Overview
Add support for evaluating tools from remote MCP servers without
requiring Python callables. Enables direct evaluation of any
MCP-compatible tool server.

## What's New

### Core Features
- **`MCPToolRegistry`**: Evaluate tools from a single MCP server
- **`CompositeMCPRegistry`**: Evaluate tools from multiple MCP servers
simultaneously
- **Automatic loaders**: `load_from_stdio()` and `load_from_http()` to
fetch tools from running servers
- **Automatic namespacing**: Tools prefixed with server name (e.g.,
`server_tool_name`)
- **Smart name resolution**: Use short names if unique, full names if
ambiguous
- **OpenAI strict mode**: Automatic schema conversion prevents parameter
hallucinations

### Usage

**Automatic Loading:**
```python
from arcade_evals import load_from_stdio, MCPToolRegistry

# Load tools automatically from MCP server
tools = load_from_stdio(["npx", "-y", "@modelcontextprotocol/server-github"])
registry = MCPToolRegistry(tools)
```

**Single MCP Server:**
```python
from arcade_evals import MCPToolRegistry, ExpectedToolCall

registry = MCPToolRegistry(mcp_tools)
suite = EvalSuite(catalog=registry)

suite.add_case(
    expected_tool_calls=[
        ExpectedToolCall(tool_name="tool_name", args={...})
    ]
)
```

**Multiple MCP Servers:**
```python
from arcade_evals import CompositeMCPRegistry, load_from_stdio

# Load from multiple servers
github_tools = load_from_stdio(["npx", "-y", "@modelcontextprotocol/server-github"])
slack_tools = load_from_stdio(["npx", "-y", "@modelcontextprotocol/server-slack"])

composite = CompositeMCPRegistry(
    tool_lists={
        "github": github_tools,
        "slack": slack_tools,
    }
)

suite = EvalSuite(catalog=composite)

suite.add_case(
    expected_tool_calls=[
        ExpectedToolCall(tool_name="github_list_issues", args={...})
    ]
)
```

## Implementation

### Files Changed
- **`libs/arcade-evals/arcade_evals/registry.py`** (NEW): Registry
abstractions and implementations
- **`libs/arcade-evals/arcade_evals/loaders.py`** (NEW): Automatic tool
loading from MCP servers
- **`libs/arcade-evals/arcade_evals/eval.py`** (MODIFIED): Enhanced
`ExpectedToolCall` and evaluation logic
- **`libs/arcade-evals/arcade_evals/__init__.py`** (MODIFIED): Exported
new registries and loaders

### Key Technical Details
- Added `BaseToolRegistry` interface for abstraction
- `MCPToolRegistry` handles single server tools
- `CompositeMCPRegistry` manages multiple servers with collision
detection
- `load_from_stdio()` and `load_from_http()` for automatic tool
discovery
- Fixed name normalization bug: MCP tools use underscores (not dots)
- Optimized tool copying: 2.5x faster via shallow copy

## Testing
-  41 tests passing (25 new tests added)
-  `test_eval_mcp_registry.py`: MCPToolRegistry functionality
-  `test_eval_composite_mcp.py`: CompositeMCPRegistry with multiple
servers
-  Verified backward compatibility with Python tools

## Backward Compatibility
 **100% backward compatible** - No breaking changes


## Breaking Changes
**None**


<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> Adds end-to-end eval UX: examples, a robust CLI runner, and rich
outputs.
> 
> - **New examples**: `eval_arcade_gateway.py`,
`eval_stdio_mcp_server.py`, `eval_http_mcp_server.py`,
`eval_comprehensive_comparison.py` with timeouts, error handling, and
track-based comparisons; detailed `README.md`
> - **CLI runner**: `arcade_cli/evals_runner.py` to execute
evals/capture in parallel with progress, error isolation, failed-only
filtering, context inclusion, and multi-provider/model support
> - **Output formatters**: `arcade_cli/formatters/` (txt, md, html,
json) for evals and capture; comparative and multi-model HTML with tabs
and context rendering
> - **Display refactor**: `display.py` now supports writing multiple
formats, failed-only disclaimers, include-context, and improved console
summaries
> 
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
ff8acf9c34a6b61462a019a1ee9df081006517d0. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

---------

Co-authored-by: Francisco Liberal <francisco@arcade.dev>
Co-authored-by: Mateo Torres <torresmateo@gmail.com>
2026-01-07 20:26:23 -03:00

97 lines
2.7 KiB
Python

"""Track management for comparative evaluations.
A track represents an isolated tool registry with a unique name.
This enables running the same evaluation cases against different
tool sources (e.g., different MCP servers) for comparison.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from arcade_evals._evalsuite._tool_registry import EvalSuiteToolRegistry
class TrackManager:
"""Manages named tracks, each with its own isolated tool registry.
Tracks enable comparative evaluations where the same cases are run
against different tool sources.
Example:
manager = TrackManager()
manager.create_track("Google Weather", registry1)
manager.create_track("OpenWeather", registry2)
for track_name in manager.get_track_names():
registry = manager.get_registry(track_name)
# Run cases against this registry
"""
def __init__(self) -> None:
self._tracks: dict[str, EvalSuiteToolRegistry] = {}
def create_track(self, name: str, registry: EvalSuiteToolRegistry) -> str:
"""Create a new track with an isolated registry.
Args:
name: Unique track name.
registry: The tool registry for this track.
Returns:
The track name (for use as track ID).
Raises:
ValueError: If track name already exists.
"""
if name in self._tracks:
raise ValueError(f"Track '{name}' already exists. Use a unique track name.")
self._tracks[name] = registry
return name
def get_registry(self, track_name: str) -> EvalSuiteToolRegistry | None:
"""Get the registry for a track.
Args:
track_name: The track name.
Returns:
The registry if found, None otherwise.
"""
return self._tracks.get(track_name)
def get_track_names(self) -> list[str]:
"""Get all registered track names.
Returns:
List of track names in registration order.
"""
return list(self._tracks.keys())
def has_track(self, name: str) -> bool:
"""Check if a track exists.
Args:
name: The track name.
Returns:
True if track exists, False otherwise.
"""
return name in self._tracks
def track_count(self) -> int:
"""Get number of registered tracks.
Returns:
Number of tracks.
"""
return len(self._tracks)
def get_all_registries(self) -> dict[str, EvalSuiteToolRegistry]:
"""Get all registries by track name.
Returns:
Dictionary mapping track names to registries.
"""
return dict(self._tracks)