PMFlow/backend/app/services/feishu_signal_pipeline/mixin_persistence_lookup.py

"""Feishu signal pipeline mixin."""
from __future__ import annotations

from .shared import *


class FeishuPipelinePersistenceLookupMixin:
    async def _resolve_user_id_by_name(self, name: Optional[str]) -> Optional[int]:
        key = (name or "").strip()
        if not key:
            return None
        user = (
            await self.db.execute(
                select(User).where(
                    and_(
                        User.tenant_id == self.tenant_id,
                        User.is_active.is_(True),
                        func.lower(User.name) == key.lower(),
                    )
                )
            )
        ).scalar_one_or_none()
        if user:
            return int(user.id)
        fuzzy = (
            await self.db.execute(
                select(User).where(
                    and_(
                        User.tenant_id == self.tenant_id,
                        User.is_active.is_(True),
                        User.name.ilike(f"%{key}%"),
                    )
                ).limit(1)
            )
        ).scalar_one_or_none()
        return int(fuzzy.id) if fuzzy else None

    async def _fallback_user_id(self) -> Optional[int]:
        user = (
            await self.db.execute(
                select(User.id).where(
                    and_(
                        User.tenant_id == self.tenant_id,
                        User.is_active.is_(True),
                    )
                ).limit(1)
            )
        ).scalar_one_or_none()
        return int(user) if user else None

    async def _resolve_chat_context_project(self, *, source_chat_id: str) -> dict[str, Any]:
        chat_id = (source_chat_id or "").strip()
        if not chat_id:
            return {"project_id": None, "ambiguous": False, "candidate_ids": [], "top_ratio": 0.0}

        recent_rows = (
            await self.db.execute(
                select(FeishuItem.matched_project_id)
                .where(
                    and_(
                        FeishuItem.tenant_id == self.tenant_id,
                        FeishuItem.source_chat_id == chat_id,
                        FeishuItem.matched_project_id.is_not(None),
                        FeishuItem.route_status.in_((ROUTE_STATUS_AUTO_DONE, "auto_processed", "confirmed")),
                    )
                )
                .order_by(FeishuItem.id.desc())
                .limit(20)
            )
        ).all()
        project_ids = [int(row[0]) for row in recent_rows if row and row[0] is not None]
        if not project_ids:
            return {"project_id": None, "ambiguous": False, "candidate_ids": [], "top_ratio": 0.0}

        unique_ids = set(project_ids)
        candidate_ids = sorted(unique_ids, key=lambda pid: (-project_ids.count(pid), pid))
        if len(unique_ids) == 1:
            return {"project_id": int(project_ids[0]), "ambiguous": False, "candidate_ids": candidate_ids, "top_ratio": 1.0}

        top_pid = max(unique_ids, key=lambda pid: project_ids.count(pid))
        top_count = project_ids.count(top_pid)
        ratio = top_count / len(project_ids)
        if top_count >= 2 and ratio >= 0.6:
            return {"project_id": int(top_pid), "ambiguous": False, "candidate_ids": candidate_ids, "top_ratio": ratio}
        return {"project_id": None, "ambiguous": True, "candidate_ids": candidate_ids, "top_ratio": ratio}

    async def _infer_project_id_from_chat_context(self, *, source_chat_id: str) -> Optional[int]:
        resolution = await self._resolve_chat_context_project(source_chat_id=source_chat_id)
        return _safe_int(resolution.get("project_id"))

    async def _get_item(self, item_id: int) -> Optional[FeishuItem]:
        return (
            await self.db.execute(
                select(FeishuItem).where(
                    and_(
                        FeishuItem.tenant_id == self.tenant_id,
                        FeishuItem.id == int(item_id),
                    )
                )
            )
        ).scalar_one_or_none()

    async def _get_project_suggestion(self, item_id: int) -> Optional[ProjectSuggestion]:
        return (
            await self.db.execute(
                select(ProjectSuggestion).where(
                    and_(
                        ProjectSuggestion.tenant_id == self.tenant_id,
                        ProjectSuggestion.feishu_item_id == int(item_id),
                    )
                )
            )
        ).scalar_one_or_none()

    async def _get_signal_by_item(self, item_id: int) -> Optional[Signal]:
        return (
            await self.db.execute(
                select(Signal).where(
                    and_(
                        Signal.tenant_id == self.tenant_id,
                        Signal.feishu_item_id == int(item_id),
                    )
                )
            )
        ).scalar_one_or_none()

    def _dedupe_queue_items(self, rows: list[FeishuItem]) -> list[FeishuItem]:
        deduped: list[FeishuItem] = []
        seen_signal_keys: set[str] = set()
        aggregated: dict[str, FeishuItem] = {}
        for row in rows:
            explicit_key = str(row.dedupe_key or "").strip()
            if explicit_key:
                existing = aggregated.get(explicit_key)
                if existing is None:
                    aggregated[explicit_key] = row
                    deduped.append(row)
                else:
                    existing.occurrence_count = max(1, int(existing.occurrence_count or 1)) + max(
                        1, int(row.occurrence_count or 1)
                    )
                continue
            level = str(row.level or "").lower()
            if level != "signal":
                deduped.append(row)
                continue
            message_id = str(row.feishu_message_id or "").strip()
            key = message_id or f"{str(row.raw_message or '')[:120]}::{str(row.source_sender or '')}"
            if key in seen_signal_keys:
                continue
            seen_signal_keys.add(key)
            deduped.append(row)
        return deduped