services/project_history.py

from __future__ import annotations

import json
import re
import time
from datetime import datetime
from pathlib import Path
from typing import Any, Callable


class ProjectHistoryStore:
    """Persistent project-package history, separate from normal chat history."""

    def __init__(self, root: Path, validator: Callable[[str], bool], max_items: int = 40):
        self.root = Path(root)
        self.validator = validator
        self.max_items = max(5, int(max_items))

    def _require_user(self, user_id: str) -> str:
        uid = str(user_id or "").strip()
        if not self.validator(uid):
            raise ValueError("invalid user_id")
        return uid

    def _path(self, user_id: str) -> Path:
        uid = self._require_user(user_id)
        return self.root / f"{uid}.project-history.json"

    @staticmethod
    def _cut(value: Any, n: int) -> str:
        return str(value or "")[:n]

    def sanitize_entry(self, item: Any) -> dict[str, Any] | None:
        if not isinstance(item, dict):
            return None
        project = self._cut(item.get("project"), 12000).strip()
        if not project:
            return None

        ident = self._cut(item.get("id"), 120)
        if not ident:
            ident = f"project-{int(time.time()*1000)}"

        title = self._cut(item.get("title"), 240).strip()
        if not title:
            title = re.sub(r"\s+", " ", project)[:42]

        options = item.get("options") if isinstance(item.get("options"), dict) else {}
        clean_options = {
            "include_ocean": bool(options.get("include_ocean", True)),
            "include_tuna": bool(options.get("include_tuna", True)),
            "include_squid": bool(options.get("include_squid", True)),
            "max_package_mb": max(20, min(int(options.get("max_package_mb") or 300), 1200)),
            "selected_ocean_keys": [
                self._cut(x, 160) for x in (options.get("selected_ocean_keys") or [])[:80]
                if str(x or "").strip()
            ],
            "selected_fisheries_names": [
                self._cut(x, 240) for x in (options.get("selected_fisheries_names") or [])[:80]
                if str(x or "").strip()
            ],
        }

        # Plans/results are already JSON from our own APIs. Bound total size by
        # round-tripping and trimming oversized snapshots.
        plan = item.get("plan") if isinstance(item.get("plan"), dict) else None
        estimate = item.get("estimate") if isinstance(item.get("estimate"), dict) else None
        build = item.get("build") if isinstance(item.get("build"), dict) else None

        def bounded(value, limit):
            if value is None:
                return None
            try:
                raw = json.dumps(value, ensure_ascii=False)
                if len(raw) <= limit:
                    return value
            except Exception:
                pass
            return None

        now = int(time.time() * 1000)
        return {
            "id": ident,
            "title": title,
            "project": project,
            "stage": self._cut(item.get("stage") or "analyzed", 40),
            "options": clean_options,
            "plan": bounded(plan, 180000),
            "estimate": bounded(estimate, 120000),
            "build": bounded(build, 120000),
            "created_at": int(item.get("created_at") or now),
            "updated_at": int(item.get("updated_at") or now),
        }

    def read(self, user_id: str) -> list[dict[str, Any]]:
        path = self._path(user_id)
        if not path.exists():
            return []
        try:
            raw = json.loads(path.read_text(encoding="utf-8"))
        except Exception:
            return []
        items = raw.get("items", []) if isinstance(raw, dict) else raw
        out = []
        for item in items if isinstance(items, list) else []:
            clean = self.sanitize_entry(item)
            if clean:
                out.append(clean)
        out.sort(key=lambda x: int(x.get("updated_at") or 0), reverse=True)
        return out[:self.max_items]

    def upsert(self, user_id: str, item: Any) -> list[dict[str, Any]]:
        clean = self.sanitize_entry(item)
        if clean is None:
            raise ValueError("project is required")
        current = self.read(user_id)
        by_id = {str(x.get("id")): x for x in current}
        old = by_id.get(clean["id"])
        if old:
            clean["created_at"] = int(old.get("created_at") or clean["created_at"])
        clean["updated_at"] = int(time.time() * 1000)
        merged = [clean] + [x for x in current if str(x.get("id")) != clean["id"]]
        self._write(user_id, merged[:self.max_items])
        return merged[:self.max_items]

    def delete(self, user_id: str, item_id: str) -> list[dict[str, Any]]:
        ident = self._cut(item_id, 120)
        current = [x for x in self.read(user_id) if str(x.get("id")) != ident]
        self._write(user_id, current)
        return current

    def _write(self, user_id: str, items: list[dict[str, Any]]) -> None:
        path = self._path(user_id)
        path.parent.mkdir(parents=True, exist_ok=True)
        payload = {
            "schema": "squid-project-package-history",
            "schema_version": 1,
            "updated_at": datetime.now().astimezone().isoformat(timespec="seconds"),
            "items": items[:self.max_items],
        }
        tmp = path.with_suffix(path.suffix + ".tmp")
        tmp.write_text(json.dumps(payload, ensure_ascii=False, indent=2), encoding="utf-8")
        tmp.replace(path)