claude-mpm 4.21.3__py3-none-any.whl → 5.0.2__py3-none-any.whl

claude_mpm/services/agents/agent_selection_service.py

@@ -0,0 +1,484 @@
+"""Agent selection service for minimal and auto-configure deployment modes.
+
+WHY: Provide two agent selection strategies to simplify deployment:
+1. Minimal Configuration: Deploy exactly 6 core agents for basic workflow
+2. Auto-Configure: Detect toolchain and deploy matching specialized agents
+
+Design Decision: Strategy Pattern with Composition
+
+This service composes SingleTierDeploymentService and ToolchainDetector
+rather than inheriting. Each deployment mode is a distinct strategy that
+orchestrates the deployment service differently.
+
+Trade-offs:
+- Flexibility: Easy to add new selection modes (e.g., profile-based)
+- Testability: Can mock deployment service and detector independently
+- Complexity: Slightly more code than monolithic approach
+- Clarity: Clear separation between selection logic and deployment mechanics
+
+Architecture:
+- AgentSelectionService orchestrates deployment strategies
+- SingleTierDeploymentService handles Git sync and file copying
+- ToolchainDetector provides toolchain analysis for auto-configure
+
+Example Workflow:
+    # Minimal mode
+    service.deploy_minimal_configuration()
+    -> Validate 6 core agents available
+    -> Deploy each agent via deployment_service.deploy_agent()
+
+    # Auto-configure mode
+    service.deploy_auto_configure(project_path)
+    -> Detect toolchain (languages, frameworks, tools)
+    -> Recommend agents based on toolchain
+    -> Merge with core agents
+    -> Deploy recommended agents
+"""
+
+import logging
+from pathlib import Path
+from typing import Any, Dict, List, Optional, Set, Tuple
+
+from src.claude_mpm.services.agents.single_tier_deployment_service import (
+    SingleTierDeploymentService,
+)
+from src.claude_mpm.services.agents.toolchain_detector import ToolchainDetector
+
+logger = logging.getLogger(__name__)
+
+
+class AgentSelectionService:
+    """Service for agent selection modes (minimal and auto-configure).
+
+    Provides two deployment strategies:
+
+    1. Minimal Configuration Mode:
+       Deploy exactly 6 core agents needed for basic Claude MPM workflow:
+       - engineer: General implementation
+       - documentation: Documentation creation
+       - qa: Quality assurance testing
+       - research: Codebase analysis
+       - ops: Deployment operations
+       - ticketing: Ticket management
+
+    2. Auto-Configure Mode:
+       Detect project toolchain and deploy matching agents:
+       - Scan project for languages, frameworks, build tools
+       - Map toolchain to specialized agents
+       - Include core agents for complete workflow
+       - Deploy recommended agent set
+
+    Usage:
+        >>> config = AgentSourceConfiguration.load()
+        >>> deployment = SingleTierDeploymentService(config, deployment_dir)
+        >>> service = AgentSelectionService(deployment)
+        >>>
+        >>> # Minimal mode
+        >>> result = service.deploy_minimal_configuration()
+        >>> print(f"Deployed {result['deployed_count']} agents")
+        >>>
+        >>> # Auto-configure mode
+        >>> result = service.deploy_auto_configure(project_path=Path("."))
+        >>> print(f"Detected: {result['toolchain']}")
+        >>> print(f"Deployed: {result['deployed_agents']}")
+    """
+
+    # Minimal configuration agents (6 core agents)
+    MINIMAL_AGENTS = [
+        "engineer",
+        "documentation",
+        "qa",
+        "research",
+        "ops",
+        "ticketing",
+    ]
+
+    def __init__(self, deployment_service: SingleTierDeploymentService):
+        """Initialize agent selection service.
+
+        Args:
+            deployment_service: Deployment service for agent operations
+        """
+        self.deployment_service = deployment_service
+        self.toolchain_detector = ToolchainDetector()
+
+        logger.info("AgentSelectionService initialized")
+
+    def deploy_minimal_configuration(self, dry_run: bool = False) -> Dict[str, Any]:
+        """Deploy minimal configuration (6 core agents).
+
+        Deploys exactly 6 agents required for basic Claude MPM workflow:
+        engineer, documentation, qa, research, ops, ticketing.
+
+        Workflow:
+        1. Get available agents from deployment service
+        2. Validate all 6 core agents are available
+        3. Deploy each core agent individually
+        4. Return deployment report with success/failure details
+
+        Args:
+            dry_run: Show what would be deployed without actually deploying
+
+        Returns:
+            Deployment report:
+            {
+                "status": "success" | "partial" | "error",
+                "mode": "minimal",
+                "deployed_count": 6,
+                "failed_count": 0,
+                "deployed_agents": ["engineer", "qa", ...],
+                "failed_agents": [],
+                "missing_agents": [],
+                "dry_run": false,
+                "details": [
+                    {
+                        "agent_name": "engineer",
+                        "deployed": true,
+                        "source": "bobmatnyc/claude-mpm-agents",
+                        "priority": 100
+                    },
+                    ...
+                ]
+            }
+
+        Error Handling:
+        - Missing agents logged and included in missing_agents list
+        - Failed deployments logged and included in failed_agents list
+        - Returns "partial" status if some agents deployed successfully
+        - Returns "error" status if no agents deployed
+        """
+        logger.info(f"Starting minimal configuration deployment (dry_run={dry_run})")
+
+        # Step 1: Get available agents
+        available_agents = self.deployment_service.list_available_agents()
+        available_agent_ids = {agent["agent_id"].lower() for agent in available_agents}
+
+        logger.debug(
+            f"Available agents: {len(available_agents)} ({sorted(available_agent_ids)})"
+        )
+
+        # Step 2: Validate all minimal agents are available
+        available_minimal, missing_minimal = self._validate_agent_availability(
+            self.MINIMAL_AGENTS, available_agent_ids
+        )
+
+        if missing_minimal:
+            logger.warning(
+                f"Some minimal agents not available: {sorted(missing_minimal)}"
+            )
+
+        # Step 3: Deploy each available minimal agent
+        deployed_agents: List[str] = []
+        failed_agents: List[str] = []
+        deployment_details: List[Dict[str, Any]] = []
+
+        for agent_name in available_minimal:
+            try:
+                result = self.deployment_service.deploy_agent(
+                    agent_name=agent_name, dry_run=dry_run
+                )
+
+                if result.get("deployed", False) or result.get("dry_run", False):
+                    deployed_agents.append(agent_name)
+                    deployment_details.append(result)
+                    logger.info(f"✓ Deployed {agent_name}")
+                else:
+                    failed_agents.append(agent_name)
+                    deployment_details.append(result)
+                    logger.warning(
+                        f"✗ Failed to deploy {agent_name}: {result.get('error', 'Unknown error')}"
+                    )
+
+            except Exception as e:
+                failed_agents.append(agent_name)
+                logger.error(f"✗ Error deploying {agent_name}: {e}")
+                deployment_details.append(
+                    {
+                        "agent_name": agent_name,
+                        "deployed": False,
+                        "error": str(e),
+                    }
+                )
+
+        # Step 4: Build report
+        deployed_count = len(deployed_agents)
+        failed_count = len(failed_agents)
+        missing_count = len(missing_minimal)
+
+        if deployed_count == len(self.MINIMAL_AGENTS) and missing_count == 0:
+            status = "success"
+        elif deployed_count > 0:
+            status = "partial"
+        else:
+            status = "error"
+
+        report = {
+            "status": status,
+            "mode": "minimal",
+            "deployed_count": deployed_count,
+            "failed_count": failed_count,
+            "missing_count": missing_count,
+            "deployed_agents": sorted(deployed_agents),
+            "failed_agents": sorted(failed_agents),
+            "missing_agents": sorted(missing_minimal),
+            "dry_run": dry_run,
+            "details": deployment_details,
+        }
+
+        logger.info(
+            f"Minimal configuration deployment complete: "
+            f"{deployed_count} deployed, {failed_count} failed, "
+            f"{missing_count} missing (status: {status})"
+        )
+
+        return report
+
+    def deploy_auto_configure(
+        self,
+        project_path: Optional[Path] = None,
+        dry_run: bool = False,
+    ) -> Dict[str, Any]:
+        """Auto-detect toolchain and deploy matching agents.
+
+        Workflow:
+        1. Detect toolchain (languages, frameworks, build tools)
+        2. Map toolchain to recommended agents
+        3. Merge recommended agents with core agents
+        4. Validate agent availability
+        5. Deploy recommended agents
+        6. Return deployment report with toolchain analysis
+
+        Args:
+            project_path: Project root path (defaults to current directory)
+            dry_run: Show what would be deployed without actually deploying
+
+        Returns:
+            Deployment report:
+            {
+                "status": "success" | "partial" | "error",
+                "mode": "auto_configure",
+                "deployed_count": 8,
+                "failed_count": 0,
+                "toolchain": {
+                    "languages": ["python"],
+                    "frameworks": ["fastapi"],
+                    "build_tools": ["docker", "make"]
+                },
+                "recommended_agents": ["python-engineer", "ops", "qa", ...],
+                "deployed_agents": ["python-engineer", "ops", "qa", ...],
+                "failed_agents": [],
+                "missing_agents": [],
+                "dry_run": false,
+                "details": [...]
+            }
+
+        Error Handling:
+        - Invalid project path: Returns error status with message
+        - Toolchain detection failure: Falls back to minimal configuration
+        - Missing agents: Logged and included in report
+        - Failed deployments: Logged and included in report
+        """
+        if project_path is None:
+            project_path = Path.cwd()
+
+        logger.info(
+            f"Starting auto-configure deployment (project={project_path}, dry_run={dry_run})"
+        )
+
+        # Step 1: Validate project path
+        if not project_path.exists() or not project_path.is_dir():
+            logger.error(f"Invalid project path: {project_path}")
+            return {
+                "status": "error",
+                "mode": "auto_configure",
+                "error": f"Invalid project path: {project_path}",
+                "deployed_count": 0,
+                "failed_count": 0,
+                "toolchain": {},
+                "recommended_agents": [],
+                "deployed_agents": [],
+                "failed_agents": [],
+                "missing_agents": [],
+                "dry_run": dry_run,
+            }
+
+        # Step 2: Detect toolchain
+        try:
+            toolchain = self.toolchain_detector.detect_toolchain(project_path)
+            logger.info(f"Detected toolchain: {toolchain}")
+        except Exception as e:
+            logger.error(f"Toolchain detection failed: {e}")
+            return {
+                "status": "error",
+                "mode": "auto_configure",
+                "error": f"Toolchain detection failed: {e}",
+                "deployed_count": 0,
+                "failed_count": 0,
+                "toolchain": {},
+                "recommended_agents": [],
+                "deployed_agents": [],
+                "failed_agents": [],
+                "missing_agents": [],
+                "dry_run": dry_run,
+            }
+
+        # Step 3: Recommend agents based on toolchain
+        recommended_agents = self.toolchain_detector.recommend_agents(toolchain)
+        logger.info(
+            f"Recommended {len(recommended_agents)} agents: {recommended_agents}"
+        )
+
+        # Step 4: Get available agents and validate
+        available_agents = self.deployment_service.list_available_agents()
+        available_agent_ids = {agent["agent_id"].lower() for agent in available_agents}
+
+        available_recommended, missing_recommended = self._validate_agent_availability(
+            recommended_agents, available_agent_ids
+        )
+
+        if missing_recommended:
+            logger.warning(
+                f"Some recommended agents not available: {sorted(missing_recommended)}"
+            )
+
+        # Step 5: Deploy each available recommended agent
+        deployed_agents: List[str] = []
+        failed_agents: List[str] = []
+        deployment_details: List[Dict[str, Any]] = []
+
+        for agent_name in available_recommended:
+            try:
+                result = self.deployment_service.deploy_agent(
+                    agent_name=agent_name, dry_run=dry_run
+                )
+
+                if result.get("deployed", False) or result.get("dry_run", False):
+                    deployed_agents.append(agent_name)
+                    deployment_details.append(result)
+                    logger.info(f"✓ Deployed {agent_name}")
+                else:
+                    failed_agents.append(agent_name)
+                    deployment_details.append(result)
+                    logger.warning(
+                        f"✗ Failed to deploy {agent_name}: {result.get('error', 'Unknown error')}"
+                    )
+
+            except Exception as e:
+                failed_agents.append(agent_name)
+                logger.error(f"✗ Error deploying {agent_name}: {e}")
+                deployment_details.append(
+                    {
+                        "agent_name": agent_name,
+                        "deployed": False,
+                        "error": str(e),
+                    }
+                )
+
+        # Step 6: Build report
+        deployed_count = len(deployed_agents)
+        failed_count = len(failed_agents)
+        missing_count = len(missing_recommended)
+
+        if deployed_count > 0 and failed_count == 0 and missing_count == 0:
+            status = "success"
+        elif deployed_count > 0:
+            status = "partial"
+        else:
+            status = "error"
+
+        report = {
+            "status": status,
+            "mode": "auto_configure",
+            "deployed_count": deployed_count,
+            "failed_count": failed_count,
+            "missing_count": missing_count,
+            "toolchain": toolchain,
+            "recommended_agents": sorted(recommended_agents),
+            "deployed_agents": sorted(deployed_agents),
+            "failed_agents": sorted(failed_agents),
+            "missing_agents": sorted(missing_recommended),
+            "dry_run": dry_run,
+            "details": deployment_details,
+        }
+
+        logger.info(
+            f"Auto-configure deployment complete: "
+            f"{deployed_count} deployed, {failed_count} failed, "
+            f"{missing_count} missing (status: {status})"
+        )
+
+        return report
+
+    def _validate_agent_availability(
+        self, required_agents: List[str], available_agents: Set[str]
+    ) -> Tuple[List[str], List[str]]:
+        """Validate which required agents are available.
+
+        Args:
+            required_agents: List of agent names needed
+            available_agents: Set of available agent IDs
+
+        Returns:
+            Tuple of (available_list, missing_list)
+            - available_list: Agents that are available
+            - missing_list: Agents that are not available
+        """
+        available_list: List[str] = []
+        missing_list: List[str] = []
+
+        for agent_name in required_agents:
+            agent_id = agent_name.lower().replace(" ", "-")
+            if agent_id in available_agents:
+                available_list.append(agent_name)
+            else:
+                missing_list.append(agent_name)
+
+        return available_list, missing_list
+
+    def get_available_selection_modes(self) -> List[Dict[str, Any]]:
+        """Get information about available selection modes.
+
+        Returns:
+            List of selection mode metadata:
+            [
+                {
+                    "mode": "minimal",
+                    "description": "Deploy 6 core agents for basic workflow",
+                    "agent_count": 6,
+                    "agents": ["engineer", "documentation", "qa", ...]
+                },
+                {
+                    "mode": "auto_configure",
+                    "description": "Auto-detect toolchain and deploy matching agents",
+                    "agent_count": "varies",
+                    "requires_project_scan": true
+                }
+            ]
+        """
+        return [
+            {
+                "mode": "minimal",
+                "description": "Deploy 6 core agents for basic workflow",
+                "agent_count": len(self.MINIMAL_AGENTS),
+                "agents": self.MINIMAL_AGENTS,
+            },
+            {
+                "mode": "auto_configure",
+                "description": "Auto-detect toolchain and deploy matching agents",
+                "agent_count": "varies",
+                "requires_project_scan": True,
+                "toolchain_support": {
+                    "languages": list(self.toolchain_detector.LANGUAGE_PATTERNS.keys()),
+                    "frameworks": list(
+                        self.toolchain_detector.FRAMEWORK_PATTERNS.keys()
+                    ),
+                    "build_tools": list(
+                        self.toolchain_detector.BUILD_TOOL_PATTERNS.keys()
+                    ),
+                },
+            },
+        ]
+
+    def __repr__(self) -> str:
+        """Return string representation."""
+        return f"AgentSelectionService(deployment_service={self.deployment_service})"