Versions: * arcade-mcp\==1.0.0rc1 * arcade-mcp-server\==1.0.0rc1 * arcade-core\==2.5.0rc1 * arcade-tdk\==2.6.0rc1 * arcade-serve\==2.2.0rc1 ### Summary Adds first-class MCP support across Arcade, introduces a new MCP server and CLI, unifies the project under the arcade-mcp name, overhauls templates/scaffolding, and improves developer tooling, secrets management, and examples. ### Highlights - **MCP Server & Core** - New MCP server with stdio and HTTP/SSE transports, session management, resumability, and lifecycle handling. - FastAPI-like `MCPApp` for building servers with lazy init; integrated worker+MCP HTTP app option. - Middleware system (logging and error handling), robust exception hierarchy, and Pydantic-based settings. - Async-safe managers for tools, resources, and prompts backed by registries and locks. - Developer-facing, transport-agnostic runtime context interfaces (logs, tools, prompts, resources, sampling, UI, notifications). - Conversion from Arcade ToolDefinition to MCP tool schema; OpenAI JSON tool schema converter. - Parser supports `@app.tool`/`@app.tool(...)` decorators. - **CLI** - New `mcp` command to run MCP servers with stdio or HTTP/SSE. - New `secret` command to set/list/unset tool secrets (supports .env input, preserves original casing for lookups). - `new` command refactored; option to create a full toolkit package with scaffolding. - `chat` command removed. - `serve.py` imports updated to `arcade_serve.fastapi.telemetry`; version retrieval now uses `arcade-mcp`. - `show.py` refactor to use new local catalog utilities. - `display_tool_details` improved: adds “Default” column and handles nested properties. - **Configuration & Discovery** - New `configure.py` to set up Claude Desktop, Cursor, and VS Code to connect to local or Arcade Cloud MCP servers. - Discovery utilities to find/install toolkits, build `ToolCatalog`s, analyze files for tools, load kits from directories (pyproject parsing), and build minimal toolkits. - Better handling of provider API key resolution and evaluation suite loading. - **Templates & Scaffolding** - Reorganized template structure (minimal vs full); moved `.pre-commit-config.yaml`, `.ruff.toml`, license, Makefile, README, tests, and tools layout to correct paths. - Minimal template adds `.env.example` for runtime secret injection. - Template pyproject updated for MCP servers; includes sample server with greeting and secret-reveal tools. - Authorization flow in templates simplified. - **Repo-wide Renaming & Examples** - Migrates references from `arcade-ai` to `arcade-mcp` across READMEs, scripts, and package metadata. - Examples updated (LangChain/LangGraph/AI SDK/TypeScript) and package name changed to `arcade-mcp-sdk`. - **Evals & Core Utilities** - Evals now use OpenAI tooling format (`OpenAIToolList`, `to_openai`); `tool_eval` takes `provider_api_key`. - Core utilities: fixed `does_function_return_value` by dedenting before parse; version bump to `2.5.0rc1` and dependency cleanup. - **Tooling & CI** - `setup-uv-env` action splits toolkit vs contrib dependency installation. - Pre-commit: excludes `libs/arcade-mcp-server/mkdocs.yml` and `libs/tests/` from YAML and Ruff hooks; Ruff per-file ignores (e.g., C901 in `libs/**/*.py`, TRY400 in server docs paths). - Makefile updates for uv env setup, quality checks, tests, builds, and new `shell` target. - Added Makefile to MCP server library to streamline dev workflow. - **Cleanup** - Removed `claude.json` config. - Simplified stdio entrypoint; removed unused imports (`arcade_gmail`, `arcade_search`). ### Breaking Changes - **CLI**: `chat` command removed; use `mcp`, `secret`, and updated `new`. - **Naming**: All users should update references from `arcade-ai` to `arcade-mcp`. - **Templates**: File paths moved; downstream scripts referencing old template locations may need updates. ### Getting Started - Run an MCP server: - `arcade mcp --stdio --toolkits your_toolkit` - `arcade mcp --http --toolkits your_toolkit` - Manage secrets: - `arcade secret set your_toolkit KEY=value` - `arcade secret list your_toolkit` - `arcade secret unset your_toolkit KEY` - Configure clients: - `arcade configure` to set up Claude Desktop, Cursor, and VS Code for local/Arcade Cloud MCP. --------- Co-authored-by: Sam Partee <sam@arcade-ai.com> Co-authored-by: Shub <125150494+shubcodes@users.noreply.github.com>
448 lines
16 KiB
Python
448 lines
16 KiB
Python
from typing import TYPE_CHECKING, Any
|
|
|
|
from arcade_core.schema import ToolDefinition
|
|
from rich.console import Console
|
|
from rich.panel import Panel
|
|
from rich.table import Table
|
|
from rich.text import Text
|
|
|
|
if TYPE_CHECKING:
|
|
from arcade_evals.eval import EvaluationResult
|
|
console = Console()
|
|
|
|
|
|
def display_tools_table(tools: list[ToolDefinition]) -> None:
|
|
"""
|
|
Display a table of tools with their name, description, package, and version.
|
|
"""
|
|
if not tools:
|
|
console.print("No tools found.", style="bold")
|
|
return
|
|
|
|
table = Table(show_header=True, header_style="bold magenta")
|
|
table.add_column("Name")
|
|
table.add_column("Description")
|
|
table.add_column("Package")
|
|
table.add_column("Version")
|
|
|
|
for tool in sorted(tools, key=lambda x: x.toolkit.name):
|
|
table.add_row(
|
|
str(tool.get_fully_qualified_name()),
|
|
tool.description.split("\n")[0] if tool.description else "",
|
|
tool.toolkit.name,
|
|
tool.toolkit.version,
|
|
)
|
|
console.print(f"Found {len(tools)} tools.")
|
|
console.print(table)
|
|
|
|
|
|
def display_tool_details(tool: ToolDefinition, worker: bool = False) -> None:
|
|
"""
|
|
Display detailed information about a specific tool using multiple panels.
|
|
|
|
Args:
|
|
tool: The tool definition to display
|
|
worker: If True, show full worker response structure. If False, show only value structure.
|
|
"""
|
|
# Description Panel
|
|
description_panel = Panel(
|
|
tool.description or "No description available.",
|
|
title=f"Tool: {tool.name}",
|
|
border_style="cyan",
|
|
)
|
|
|
|
# Inputs Panel
|
|
inputs = tool.input.parameters
|
|
if inputs:
|
|
inputs_table = Table(show_header=True, header_style="bold green")
|
|
inputs_table.add_column("Name", style="cyan")
|
|
inputs_table.add_column("Type", style="magenta")
|
|
inputs_table.add_column("Required", style="yellow")
|
|
inputs_table.add_column("Description", style="white")
|
|
inputs_table.add_column("Default", style="blue")
|
|
for param in inputs:
|
|
# Since InputParameter does not have a default field, we use "N/A"
|
|
default_value = "N/A"
|
|
if param.value_schema.enum:
|
|
default_value = f"One of {param.value_schema.enum}"
|
|
inputs_table.add_row(
|
|
param.name,
|
|
param.value_schema.val_type,
|
|
str(param.required),
|
|
param.description or "",
|
|
default_value,
|
|
)
|
|
inputs_panel = Panel(
|
|
inputs_table,
|
|
title="Input Parameters",
|
|
border_style="green",
|
|
)
|
|
else:
|
|
inputs_panel = Panel(
|
|
"No input parameters.",
|
|
title="Input Parameters",
|
|
border_style="green",
|
|
)
|
|
|
|
# Output Panel - Show different levels based on worker flag
|
|
output = tool.output
|
|
if output and output.value_schema:
|
|
output_table = Table(show_header=True, header_style="bold blue")
|
|
output_table.add_column("Field", style="cyan")
|
|
output_table.add_column("Type", style="magenta")
|
|
output_table.add_column("Description", style="white")
|
|
|
|
if worker:
|
|
# Show full worker response structure
|
|
output_table.add_row(
|
|
"[bold]Response Structure[/bold]",
|
|
"",
|
|
"[dim]The tool response follows this structure:[/dim]",
|
|
)
|
|
|
|
# Available modes determine which fields can be present
|
|
modes = output.available_modes
|
|
|
|
if "value" in modes:
|
|
# Show the value field with its schema
|
|
value_type: str = output.value_schema.val_type
|
|
display_type: str = value_type # Separate variable for display string
|
|
if value_type == "array" and output.value_schema.inner_val_type:
|
|
display_type = rf"array\[{output.value_schema.inner_val_type}]"
|
|
elif output.value_schema.enum:
|
|
display_type = f"{value_type} (enum: {', '.join(output.value_schema.enum)})"
|
|
|
|
output_table.add_row(
|
|
" value",
|
|
display_type,
|
|
output.description or "The successful result from the tool",
|
|
)
|
|
|
|
# If the value is a json type with properties, show them
|
|
if (
|
|
output.value_schema.val_type == "json"
|
|
and hasattr(output.value_schema, "properties")
|
|
and output.value_schema.properties
|
|
):
|
|
_add_nested_properties(output_table, output.value_schema.properties, indent=2)
|
|
|
|
if "error" in modes:
|
|
output_table.add_row(
|
|
" error", "object", "[dim]Error details if the tool fails[/dim]"
|
|
)
|
|
output_table.add_row(
|
|
" message", "string", "[dim]User-facing error message[/dim]"
|
|
)
|
|
output_table.add_row(
|
|
" developer_message",
|
|
"string?",
|
|
"[dim]Technical error details (optional)[/dim]",
|
|
)
|
|
|
|
if "null" in modes:
|
|
output_table.add_row(" value", "null", "[dim]Tool can return null/None[/dim]")
|
|
|
|
# Additional fields that may be present
|
|
output_table.add_row("", "", "")
|
|
output_table.add_row(
|
|
"[bold]Additional Fields[/bold]",
|
|
"",
|
|
"[dim]May be present in any response:[/dim]",
|
|
)
|
|
output_table.add_row(
|
|
" logs", "array?", "[dim]Optional warnings or info messages[/dim]"
|
|
)
|
|
output_table.add_row(
|
|
" requires_authorization",
|
|
"object?",
|
|
"[dim]OAuth flow details if auth needed[/dim]",
|
|
)
|
|
else:
|
|
# Show only the value structure (simplified view)
|
|
# Show the value type and description
|
|
display_type = _format_type_string(output.value_schema)
|
|
if output.value_schema.enum:
|
|
display_type = (
|
|
f"{output.value_schema.val_type} (enum: {', '.join(output.value_schema.enum)})"
|
|
)
|
|
|
|
output_table.add_row(
|
|
"[bold]Value[/bold]",
|
|
display_type,
|
|
output.description or "The return value from the tool",
|
|
)
|
|
|
|
# If the value is a json type with properties, show them
|
|
if (
|
|
output.value_schema.val_type == "json"
|
|
and hasattr(output.value_schema, "properties")
|
|
and output.value_schema.properties
|
|
):
|
|
_add_nested_properties(output_table, output.value_schema.properties, indent=1)
|
|
|
|
# Create subtitle with modes info
|
|
modes_text = Text()
|
|
modes_text.append("Response Modes: ", style="bold")
|
|
modes_text.append("One of { ", style="dim")
|
|
for i, mode in enumerate(output.available_modes):
|
|
if i > 0:
|
|
modes_text.append(", ", style="dim")
|
|
if mode == "value":
|
|
modes_text.append(mode, style="green")
|
|
elif mode == "error":
|
|
modes_text.append(mode, style="red")
|
|
elif mode == "null":
|
|
modes_text.append(mode, style="yellow")
|
|
else:
|
|
modes_text.append(mode, style="magenta")
|
|
modes_text.append(" }", style="dim")
|
|
|
|
output_panel = Panel(
|
|
output_table,
|
|
title="Output Schema",
|
|
border_style="blue",
|
|
subtitle=modes_text,
|
|
)
|
|
else:
|
|
# No schema defined
|
|
no_schema_table = Table(show_header=False)
|
|
no_schema_table.add_column()
|
|
|
|
if worker:
|
|
no_schema_table.add_row(
|
|
"[dim]No output schema defined. The tool response will follow this structure:[/dim]"
|
|
)
|
|
no_schema_table.add_row("")
|
|
no_schema_table.add_row("[cyan]Response Structure:[/cyan]")
|
|
no_schema_table.add_row(" • [bold]value[/bold]: null (when successful)")
|
|
no_schema_table.add_row(" • [bold]error[/bold]: object (when failed)")
|
|
no_schema_table.add_row(" • [bold]logs[/bold]: array? (optional warnings/info)")
|
|
else:
|
|
no_schema_table.add_row("[dim]No output schema defined.[/dim]")
|
|
no_schema_table.add_row("")
|
|
no_schema_table.add_row("The tool returns: [bold]null[/bold]")
|
|
|
|
output_panel = Panel(
|
|
no_schema_table,
|
|
title="Output Schema",
|
|
border_style="blue",
|
|
)
|
|
|
|
# Combine all panels vertically
|
|
console.print(description_panel)
|
|
console.print(inputs_panel)
|
|
console.print(output_panel)
|
|
|
|
|
|
def _add_nested_properties(
|
|
table: Table,
|
|
properties: dict[str, Any],
|
|
indent: int = 0,
|
|
is_array_item: bool = False,
|
|
) -> None:
|
|
"""
|
|
Recursively add nested properties to the output table.
|
|
|
|
Args:
|
|
table: The Rich table to add rows to
|
|
properties: Dictionary of property names to ValueSchema objects
|
|
indent: Current indentation level
|
|
is_array_item: Whether these properties are for array items
|
|
"""
|
|
indent_prefix = " " * indent
|
|
|
|
# Show array item indicator if needed
|
|
if is_array_item and indent > 0:
|
|
table.add_row(
|
|
f"{indent_prefix[:-2]}[item]",
|
|
"",
|
|
"[dim]Each item in array:[/dim]",
|
|
)
|
|
|
|
for prop_name, prop_schema in properties.items():
|
|
# Format the type string
|
|
type_str = _format_type_string(prop_schema)
|
|
|
|
# Add the property row with better descriptions
|
|
description = ""
|
|
# For nested properties, we don't have descriptions yet, but we could add them
|
|
if hasattr(prop_schema, "description") and prop_schema.description:
|
|
description = prop_schema.description
|
|
|
|
table.add_row(
|
|
f"{indent_prefix}{prop_name}",
|
|
type_str,
|
|
f"[dim]{description}[/dim]" if description else "",
|
|
)
|
|
|
|
# Recursively add nested properties if this is a json type with properties
|
|
if (
|
|
prop_schema.val_type == "json"
|
|
and hasattr(prop_schema, "properties")
|
|
and prop_schema.properties
|
|
):
|
|
_add_nested_properties(table, prop_schema.properties, indent + 1)
|
|
# Handle arrays with inner properties
|
|
elif (
|
|
prop_schema.val_type == "array"
|
|
and hasattr(prop_schema, "inner_properties")
|
|
and prop_schema.inner_properties
|
|
):
|
|
_add_nested_properties(
|
|
table, prop_schema.inner_properties, indent + 1, is_array_item=True
|
|
)
|
|
|
|
|
|
def _format_type_string(schema: Any) -> str:
|
|
"""Format type string for display."""
|
|
type_str: str = schema.val_type
|
|
|
|
if schema.val_type == "array":
|
|
if hasattr(schema, "inner_properties") and schema.inner_properties:
|
|
type_str = r"array\[object]"
|
|
elif schema.inner_val_type:
|
|
type_str = rf"array\[{schema.inner_val_type}]"
|
|
elif schema.enum:
|
|
type_str = f"{type_str} (enum)"
|
|
|
|
return type_str
|
|
|
|
|
|
def display_tool_messages(tool_messages: list[dict]) -> None:
|
|
for message in tool_messages:
|
|
if message["role"] == "assistant":
|
|
for tool_call in message.get("tool_calls", []):
|
|
console.print(
|
|
f"[bold]Called tool '{tool_call['function']['name']}' with parameters:[/bold] {tool_call['function']['arguments']}",
|
|
style="dim",
|
|
)
|
|
elif message["role"] == "tool":
|
|
console.print(
|
|
f"[bold]'{message['name']}' tool returned:[/bold] {message['content']}",
|
|
style="dim",
|
|
)
|
|
|
|
|
|
def display_eval_results(results: list[list[dict[str, Any]]], show_details: bool = False) -> None:
|
|
"""
|
|
Display evaluation results in a format inspired by pytest's output.
|
|
|
|
Args:
|
|
results: List of dictionaries containing evaluation results for each model.
|
|
show_details: Whether to show detailed results for each case.
|
|
"""
|
|
total_passed = 0
|
|
total_failed = 0
|
|
total_warned = 0
|
|
total_cases = 0
|
|
|
|
for eval_suite in results:
|
|
for model_results in eval_suite:
|
|
model = model_results.get("model", "Unknown Model")
|
|
rubric = model_results.get("rubric", "Unknown Rubric")
|
|
cases = model_results.get("cases", [])
|
|
total_cases += len(cases)
|
|
|
|
console.print(f"[bold]Model:[/bold] [bold magenta]{model}[/bold magenta]")
|
|
if show_details:
|
|
console.print(f"[bold magenta]{rubric}[/bold magenta]")
|
|
|
|
for case in cases:
|
|
evaluation = case["evaluation"]
|
|
status = (
|
|
"[green]PASSED[/green]"
|
|
if evaluation.passed
|
|
else "[yellow]WARNED[/yellow]"
|
|
if evaluation.warning
|
|
else "[red]FAILED[/red]"
|
|
)
|
|
if evaluation.passed:
|
|
total_passed += 1
|
|
elif evaluation.warning:
|
|
total_warned += 1
|
|
else:
|
|
total_failed += 1
|
|
|
|
# Display one-line summary for each case with score as a percentage
|
|
score_percentage = evaluation.score * 100
|
|
console.print(f"{status} {case['name']} -- Score: {score_percentage:.2f}%")
|
|
|
|
if show_details:
|
|
# Show detailed information for each case
|
|
console.print(f"[bold]User Input:[/bold] {case['input']}\n")
|
|
console.print("[bold]Details:[/bold]")
|
|
console.print(_format_evaluation(evaluation))
|
|
console.print("-" * 80)
|
|
|
|
# Summary
|
|
summary = (
|
|
f"[bold]Summary -- [/bold]Total: {total_cases} -- [green]Passed: {total_passed}[/green]"
|
|
)
|
|
if total_warned > 0:
|
|
summary += f" -- [yellow]Warnings: {total_warned}[/yellow]"
|
|
if total_failed > 0:
|
|
summary += f" -- [red]Failed: {total_failed}[/red]"
|
|
console.print(summary + "\n")
|
|
|
|
|
|
def _format_evaluation(evaluation: "EvaluationResult") -> str:
|
|
"""
|
|
Format evaluation results with color-coded matches and scores.
|
|
|
|
Args:
|
|
evaluation: An EvaluationResult object containing the evaluation results.
|
|
|
|
Returns:
|
|
A formatted string representation of the evaluation details.
|
|
"""
|
|
result_lines = []
|
|
if evaluation.failure_reason:
|
|
result_lines.append(f"[bold red]Failure Reason:[/bold red] {evaluation.failure_reason}")
|
|
else:
|
|
for critic_result in evaluation.results:
|
|
is_criticized = critic_result.get("is_criticized", True)
|
|
match_color = (
|
|
"yellow" if not is_criticized else "green" if critic_result["match"] else "red"
|
|
)
|
|
field = critic_result["field"]
|
|
score = critic_result["score"]
|
|
weight = critic_result["weight"]
|
|
expected = critic_result["expected"]
|
|
actual = critic_result["actual"]
|
|
|
|
if is_criticized:
|
|
result_lines.append(
|
|
f"[bold]{field}:[/bold] "
|
|
f"[{match_color}]Match: {critic_result['match']}"
|
|
f"\n Score: {score:.2f}/{weight:.2f}[/{match_color}]"
|
|
f"\n Expected: {expected}"
|
|
f"\n Actual: {actual}"
|
|
)
|
|
else:
|
|
result_lines.append(
|
|
f"[bold]{field}:[/bold] "
|
|
f"[{match_color}]Un-criticized[/{match_color}]"
|
|
f"\n Expected: {expected}"
|
|
f"\n Actual: {actual}"
|
|
)
|
|
return "\n".join(result_lines)
|
|
|
|
|
|
def display_arcade_chat_header(base_url: str, stream: bool) -> None:
|
|
chat_header = Text.assemble(
|
|
"\n",
|
|
(
|
|
"=== Arcade Chat ===",
|
|
"bold magenta underline",
|
|
),
|
|
"\n",
|
|
"\n",
|
|
"Chatting with Arcade Engine at ",
|
|
(
|
|
base_url,
|
|
"bold blue",
|
|
),
|
|
)
|
|
if stream:
|
|
chat_header.append(" (streaming)")
|
|
console.print(chat_header)
|