scan_dependencies

Identify and extract project dependency details from pyproject.toml files. Provides JSON output with specifications and metadata, enabling accurate AI-assisted documentation and coding support.

Instructions

Scan project dependencies from pyproject.toml

Args: project_path: Path to project directory (defaults to current directory)

Returns: JSON with dependency specifications and project metadata

Input Schema

TableJSON Schema

Name	Required	Description	Default
`project_path`	No

Implementation Reference

src/autodoc_mcp/main.py:116-201 (handler)

The core handler function for the 'scan_dependencies' MCP tool. It validates the project path, parses dependencies using PyProjectParser, handles errors, tracks metrics, and returns formatted results or error responses. Registered via @mcp.tool decorator.

@mcp.tool
async def scan_dependencies(project_path: str | None = None) -> dict[str, Any]:
    """
    Scan project dependencies from pyproject.toml

    Args:
        project_path: Path to project directory (defaults to current directory)

    Returns:
        JSON with dependency specifications and project metadata
    """
    from .observability import get_metrics_collector, track_request

    async with track_request("scan_dependencies") as metrics:
        if parser is None:
            get_metrics_collector().finish_request(
                metrics.request_id,
                success=False,
                error_type="ServiceNotInitialized",
            )
            return {
                "success": False,
                "error": {
                    "message": "Parser not initialized",
                    "suggestion": "Try again or restart the MCP server",
                    "severity": "critical",
                    "code": "service_not_initialized",
                    "recoverable": False,
                },
            }

        try:
            # Validate project path if provided
            if project_path is not None:
                path = InputValidator.validate_project_path(project_path)
            else:
                path = Path.cwd()
            logger.info("Scanning dependencies", project_path=str(path))

            result = await parser.parse_project(path)

            # Record successful metrics
            get_metrics_collector().finish_request(
                metrics.request_id,
                success=True,
                dependency_count=result.successful_deps,
            )

            # Use ResponseFormatter for consistent error formatting
            return ResponseFormatter.format_scan_response(result)

        except ProjectParsingError as e:
            formatted_error = ErrorFormatter.format_exception(e)
            get_metrics_collector().finish_request(
                metrics.request_id,
                success=False,
                error_type="ProjectParsingError",
            )
            return {
                "success": False,
                "error": {
                    "message": formatted_error.message,
                    "suggestion": formatted_error.suggestion,
                    "severity": formatted_error.severity.value,
                    "code": formatted_error.error_code,
                    "recoverable": formatted_error.recoverable,
                },
            }
        except Exception as e:
            formatted_error = ErrorFormatter.format_exception(e)
            get_metrics_collector().finish_request(
                metrics.request_id,
                success=False,
                error_type=type(e).__name__,
            )
            return {
                "success": False,
                "error": {
                    "message": formatted_error.message,
                    "suggestion": formatted_error.suggestion,
                    "severity": formatted_error.severity.value,
                    "code": formatted_error.error_code,
                    "recoverable": formatted_error.recoverable,
                },
            }

src/autodoc_mcp/main.py:116-116 (registration)
The @mcp.tool decorator registers the scan_dependencies function as an MCP tool in the FastMCP server.
```
@mcp.tool
```

AutoDocs MCP Server