TimeTracker/app/services/task_service.py

"""
Service for task business logic.
"""

from typing import Any, Dict, List, Optional

from app import db
from app.constants import TaskStatus, WebhookEvent
from app.models import Task
from app.repositories import ProjectRepository, TaskRepository
from app.utils.db import safe_commit
from app.utils.event_bus import emit_event


class TaskService:
    """
    Service for task business logic operations.

    This service handles all task-related business logic including:
    - Creating and updating tasks
    - Listing tasks with filtering and pagination
    - Getting task details with related data
    - Task assignment and status management

    All methods use the repository pattern for data access and include
    eager loading to prevent N+1 query problems.

    Example:
        service = TaskService()
        result = service.create_task(
            name="New Task",
            project_id=1,
            created_by=user_id
        )
        if result['success']:
            task = result['task']
    """

    def __init__(self):
        """
        Initialize TaskService with required repositories.
        """
        self.task_repo = TaskRepository()
        self.project_repo = ProjectRepository()

    # Last-ditch fallback — only used if KanbanColumn lookup fails.
    # The real validation happens via KanbanColumn.get_valid_status_keys(project_id=...)
    # so users with custom kanban columns (e.g. on_hold) aren't silently coerced to todo.
    VALID_STATUSES = ("todo", "in_progress", "review", "done", "on_hold", "cancelled")

    def create_task(
        self,
        name: str,
        project_id: int,
        created_by: int,
        description: Optional[str] = None,
        assignee_id: Optional[int] = None,
        priority: str = "medium",
        due_date: Optional[Any] = None,
        estimated_hours: Optional[float] = None,
        color: Optional[str] = None,
        status: Optional[str] = None,
        tags: Optional[str] = None,
    ) -> Dict[str, Any]:
        """
        Create a new task.

        Args:
            name: Task name
            project_id: Project ID
            description: Task description
            assignee_id: User ID of assignee
            priority: Task priority (low, medium, high)
            due_date: Due date
            estimated_hours: Estimated hours
            created_by: User ID of creator
            color: Optional Gantt chart bar color (hex e.g. #3b82f6)
            status: Optional initial status (todo, in_progress, review, done, cancelled)

        Returns:
            dict with 'success', 'message', and 'task' keys
        """
        # Validate project
        project = self.project_repo.get_by_id(project_id)
        if not project:
            return {"success": False, "message": "Invalid project", "error": "invalid_project"}

        # Validate status against the configured kanban columns for this project.
        # Falls back to the hardcoded VALID_STATUSES tuple if KanbanColumn is unavailable
        # (e.g. table not yet seeded during a fresh migration).
        try:
            from app.models import KanbanColumn

            allowed = set(KanbanColumn.get_valid_status_keys(project_id=project_id) or self.VALID_STATUSES)
        except Exception:
            allowed = set(self.VALID_STATUSES)
        task_status = status if status and status in allowed else TaskStatus.TODO.value

        # Create task
        task = self.task_repo.create(
            name=name,
            project_id=project_id,
            description=description,
            assigned_to=assignee_id,
            priority=priority,
            due_date=due_date,
            estimated_hours=estimated_hours,
            status=task_status,
            created_by=created_by,
            tags=tags,
        )
        if color:
            task.color = color

        if not safe_commit("create_task", {"project_id": project_id, "created_by": created_by}):
            return {
                "success": False,
                "message": "Could not create task due to a database error",
                "error": "database_error",
            }

        # Emit domain event
        emit_event(
            WebhookEvent.TASK_CREATED.value, {"task_id": task.id, "project_id": project_id, "created_by": created_by}
        )

        return {"success": True, "message": "Task created successfully", "task": task}

    def get_task_with_details(
        self,
        task_id: int,
        include_time_entries: bool = True,
        include_comments: bool = True,
        include_activities: bool = True,
    ) -> Optional[Task]:
        """
        Get task with all related data using eager loading to prevent N+1 queries.

        Args:
            task_id: The task ID
            include_time_entries: Whether to include time entries
            include_comments: Whether to include comments
            include_activities: Whether to include activities

        Returns:
            Task with eagerly loaded relations, or None if not found
        """
        from sqlalchemy.orm import joinedload

        from app.models import Comment, TaskActivity, TimeEntry

        query = self.task_repo.query().filter_by(id=task_id)

        # Eagerly load project and assignee
        query = query.options(joinedload(Task.project), joinedload(Task.assigned_user), joinedload(Task.creator))

        # Conditionally load relations
        # Note: time_entries is a dynamic relationship (lazy='dynamic') and cannot be eager loaded
        # Time entries must be queried separately using task.time_entries.order_by(...).all()

        if include_comments:
            query = query.options(joinedload(Task.comments).joinedload(Comment.author))

        # Note: activities is a dynamic relationship (lazy='dynamic') and cannot be eager loaded
        # Activities must be queried separately using task.activities.order_by(...).all()

        return query.first()

    def update_task(self, task_id: int, user_id: int, **kwargs) -> Dict[str, Any]:
        """
        Update a task.

        Returns:
            dict with 'success', 'message', and 'task' keys
        """
        task = self.task_repo.get_by_id(task_id)

        if not task:
            return {"success": False, "message": "Task not found", "error": "not_found"}

        # Update fields
        self.task_repo.update(task, **kwargs)

        if not safe_commit("update_task", {"task_id": task_id, "user_id": user_id}):
            return {
                "success": False,
                "message": "Could not update task due to a database error",
                "error": "database_error",
            }

        return {"success": True, "message": "Task updated successfully", "task": task}

    def get_project_tasks(self, project_id: int, status: Optional[str] = None) -> List[Task]:
        """Get tasks for a project"""
        return self.task_repo.get_by_project(project_id=project_id, status=status, include_relations=True)

    def list_tasks(
        self,
        status: Optional[str] = None,
        priority: Optional[str] = None,
        project_id: Optional[int] = None,
        assigned_to: Optional[int] = None,
        search: Optional[str] = None,
        overdue: bool = False,
        tags: Optional[str] = None,
        user_id: Optional[int] = None,
        is_admin: bool = False,
        has_view_all_tasks: bool = False,
        page: int = 1,
        per_page: int = 20,
        project_ids: Optional[list] = None,
        assigned_to_ids: Optional[list] = None,
    ) -> Dict[str, Any]:
        """
        List tasks with filtering and pagination.
        Uses eager loading to prevent N+1 queries.

        Returns:
            dict with 'tasks', 'pagination', and 'total' keys
        """
        import logging
        import time

        from sqlalchemy.orm import joinedload

        from app.utils.timezone import now_in_app_timezone

        logger = logging.getLogger(__name__)
        start_time = time.time()
        step_start = time.time()

        query = self.task_repo.query()
        logger.debug(
            f"[TaskService.list_tasks] Step 1: Initial query creation took {(time.time() - step_start) * 1000:.2f}ms"
        )

        step_start = time.time()
        # Eagerly load relations to prevent N+1
        # Use selectinload for better performance with many tasks (avoids cartesian product)
        from sqlalchemy.orm import selectinload

        query = query.options(selectinload(Task.project), selectinload(Task.assigned_user), selectinload(Task.creator))
        logger.debug(
            f"[TaskService.list_tasks] Step 2: Eager loading setup took {(time.time() - step_start) * 1000:.2f}ms"
        )

        step_start = time.time()
        # Apply filters
        if status:
            query = query.filter(Task.status == status)

        if priority:
            query = query.filter(Task.priority == priority)

        # Support both single ID (backward compatibility) and multi-select
        if project_ids:
            query = query.filter(Task.project_id.in_(project_ids))
        elif project_id:
            query = query.filter(Task.project_id == project_id)

        if assigned_to_ids:
            query = query.filter(Task.assigned_to.in_(assigned_to_ids))
        elif assigned_to:
            query = query.filter(Task.assigned_to == assigned_to)

        if search:
            like = f"%{search}%"
            query = query.filter(db.or_(Task.name.ilike(like), Task.description.ilike(like)))

        # Tags filter: match tasks that have at least one of the specified tags (comma-separated)
        if tags:
            tag_list = [t.strip() for t in tags.split(",") if t.strip()]
            if tag_list:
                tag_conditions = [Task.tags.ilike(f"%{tag}%") for tag in tag_list]
                query = query.filter(db.or_(*tag_conditions))

        # Overdue filter
        if overdue:
            today_local = now_in_app_timezone().date()
            query = query.filter(Task.due_date < today_local, Task.status.in_(["todo", "in_progress", "review"]))

        # Permission filter - users without view_all_tasks permission only see their tasks
        if not has_view_all_tasks and user_id:
            query = query.filter(db.or_(Task.assigned_to == user_id, Task.created_by == user_id))
        logger.debug(
            f"[TaskService.list_tasks] Step 3: Applying filters took {(time.time() - step_start) * 1000:.2f}ms"
        )

        step_start = time.time()
        # Order by priority, due date, created date
        query = query.order_by(Task.priority.desc(), Task.due_date.asc(), Task.created_at.asc())
        logger.debug(f"[TaskService.list_tasks] Step 4: Ordering query took {(time.time() - step_start) * 1000:.2f}ms")

        step_start = time.time()
        # Optimize pagination: fetch one extra item to check for next page without full count
        offset = (page - 1) * per_page
        tasks_with_extra = query.limit(per_page + 1).offset(offset).all()

        # Check if there's a next page
        has_next = len(tasks_with_extra) > per_page
        tasks = tasks_with_extra[:per_page]  # Remove extra item if present

        # For count, use a simpler query without joins (much faster)
        # Only count if we're on first page or we detected a next page
        if page == 1 or has_next:
            count_start = time.time()
            count_query = self.task_repo.query()
            # Apply same filters but without eager loading (faster)
            if status:
                count_query = count_query.filter(Task.status == status)
            if priority:
                count_query = count_query.filter(Task.priority == priority)
            # Support both single ID and multi-select
            if project_ids:
                count_query = count_query.filter(Task.project_id.in_(project_ids))
            elif project_id:
                count_query = count_query.filter(Task.project_id == project_id)
            if assigned_to_ids:
                count_query = count_query.filter(Task.assigned_to.in_(assigned_to_ids))
            elif assigned_to:
                count_query = count_query.filter(Task.assigned_to == assigned_to)
            if search:
                like = f"%{search}%"
                count_query = count_query.filter(db.or_(Task.name.ilike(like), Task.description.ilike(like)))
            if tags:
                tag_list = [t.strip() for t in tags.split(",") if t.strip()]
                if tag_list:
                    tag_conditions = [Task.tags.ilike(f"%{tag}%") for tag in tag_list]
                    count_query = count_query.filter(db.or_(*tag_conditions))
            if overdue:
                today_local = now_in_app_timezone().date()
                count_query = count_query.filter(
                    Task.due_date < today_local, Task.status.in_(["todo", "in_progress", "review"])
                )
            if not has_view_all_tasks and user_id:
                count_query = count_query.filter(db.or_(Task.assigned_to == user_id, Task.created_by == user_id))
            total = count_query.count()
            logger.debug(f"[TaskService.list_tasks] Count query took {(time.time() - count_start) * 1000:.2f}ms")
        else:
            # Estimate: we know there's no next page, so total is at most current page items
            total = (page - 1) * per_page + len(tasks)

        # Create pagination-like object compatible with Flask-SQLAlchemy pagination
        from types import SimpleNamespace

        pagination = SimpleNamespace()
        pagination.items = tasks
        pagination.page = page
        pagination.per_page = per_page
        pagination.total = total
        pagination.pages = (total + per_page - 1) // per_page if total else 1
        pagination.has_next = has_next
        pagination.has_prev = page > 1

        logger.debug(
            f"[TaskService.list_tasks] Step 5: Pagination query execution took {(time.time() - step_start) * 1000:.2f}ms (total: {pagination.total} tasks, page: {page}, per_page: {per_page})"
        )

        step_start = time.time()
        # Pre-calculate total_hours for all tasks in a single query to avoid N+1
        # This prevents the template from triggering individual queries for each task
        tasks = pagination.items
        logger.debug(
            f"[TaskService.list_tasks] Step 6: Getting pagination items took {(time.time() - step_start) * 1000:.2f}ms ({len(tasks)} tasks)"
        )

        if tasks:
            from app.models import KanbanColumn, TimeEntry

            step_start = time.time()
            task_ids = [task.id for task in tasks]
            logger.debug(
                f"[TaskService.list_tasks] Step 7: Extracting task IDs took {(time.time() - step_start) * 1000:.2f}ms"
            )

            step_start = time.time()
            # Calculate total hours for all tasks in one query
            results = (
                db.session.query(TimeEntry.task_id, db.func.sum(TimeEntry.duration_seconds).label("total_seconds"))
                .filter(TimeEntry.task_id.in_(task_ids), TimeEntry.end_time.isnot(None))
                .group_by(TimeEntry.task_id)
                .all()
            )
            total_hours_map = {task_id: total_seconds for task_id, total_seconds in results}
            logger.debug(
                f"[TaskService.list_tasks] Step 8: Calculating total hours query took {(time.time() - step_start) * 1000:.2f}ms ({len(results)} results)"
            )

            step_start = time.time()
            # Pre-load kanban columns to avoid N+1 queries in status_display property
            # Load global columns (project_id is None) since tasks don't have project-specific columns
            kanban_columns = KanbanColumn.get_active_columns(project_id=None)
            status_display_map = {}
            for col in kanban_columns:
                status_display_map[col.key] = col.label
            logger.debug(
                f"[TaskService.list_tasks] Step 9: Loading kanban columns took {(time.time() - step_start) * 1000:.2f}ms ({len(kanban_columns)} columns)"
            )

            # Fallback status map if no columns found
            fallback_status_map = {
                "todo": "To Do",
                "in_progress": "In Progress",
                "review": "Review",
                "done": "Done",
                "cancelled": "Cancelled",
            }

            step_start = time.time()
            # Cache the calculated values on task objects to avoid property queries
            for task in tasks:
                total_seconds = total_hours_map.get(task.id, 0) or 0
                task._cached_total_hours = round(total_seconds / 3600, 2) if total_seconds else 0.0

                # Cache status_display to avoid N+1 queries
                task._cached_status_display = status_display_map.get(
                    task.status, fallback_status_map.get(task.status, task.status.replace("_", " ").title())
                )
            logger.debug(
                f"[TaskService.list_tasks] Step 10: Caching task properties took {(time.time() - step_start) * 1000:.2f}ms"
            )

        total_time = (time.time() - start_time) * 1000
        logger.info(
            f"[TaskService.list_tasks] Total time: {total_time:.2f}ms (tasks: {len(tasks) if tasks else 0}, page: {page}, per_page: {per_page})"
        )

        return {"tasks": tasks, "pagination": pagination, "total": pagination.total}