"""M5 离线编排测试：注入 fake store/fetch/extract/chat，验证 done 与 rejected 两条路径。

真实落库的 e2e 见 scripts/run_batch.py（在能连 DB 的云端跑）。
"""
from __future__ import annotations

import json
from pathlib import Path

from creation_knowledge.config import PgConfig, Settings
from creation_knowledge.integrations.crawler import parse_detail_response
from creation_knowledge.models import ExtractedContent
from creation_knowledge.pipeline import run_pipeline

FIXTURES = Path(__file__).parent / "fixtures"
CID = "67e4bdf50000000006028a59"


def _settings() -> Settings:
    return Settings(
        pg=PgConfig(host="h", port=5432, user="u", password="p", database="d"),
        crawler_base_url="http://x", crawler_key="", crawler_timeout=30,
        video_model="m", gemini_api_key="", openrouter_base_url="http://x",
        openrouter_api_key="k", llm_model="m", knowhub_api="http://x",
        ingest_enabled=False,
    )


class FakeStore:
    def __init__(self):
        self.posts: dict = {}
        self.items: list = []

    def upsert_post(self, post):
        self.posts[post.id] = {"stage": "fetched"}

    def set_extracted(self, pid, ex):
        self.posts[pid].update(stage="extracted", extracted=ex)

    def set_screening(self, pid, sc):
        self.posts[pid].update(stage="screened", screening=sc)

    def update_stage(self, pid, stage):
        self.posts[pid]["stage"] = stage

    def save_item(self, pid, item, deco, payload):
        self.items.append({"post_id": pid, "item": item, "deco": deco, "payload": payload})
        return len(self.items)

    def update_item_ingest(self, iid, status, kid):
        self.items[iid - 1].update(ingest_status=status, knowledge_id=kid)


def _fetch(url):
    resp = json.loads((FIXTURES / f"xhs_case_{CID}.json").read_text("utf-8"))
    return parse_detail_response(resp, fallback_content_id=CID)


def _extract(post):
    return ExtractedContent(text=post.body_text or "内容", is_empty=False)


def _chat_pass(system, user):
    if "筛选" in system:
        return {"passed": True, "score": 8, "reason": "ok"}
    if "拆分" in system:
        return {"items": [{"title": "脚本要素", "knowledge_types": ["how"],
                           "what": "null", "why": "null", "how": "逐个填要素",
                           "evidence": ["原句"]}]}
    if "解构" in system:
        return {"stages": ["脚本"], "scopes": [{"scope_type": "form", "value": "操作流程"}],
                "stage_reason": "r", "scope_reason": "r"}
    raise AssertionError(f"unexpected system: {system}")


def _chat_reject(system, user):
    if "筛选" in system:
        return {"passed": False, "score": 2, "reason": "只是作品本身"}
    raise AssertionError("rejected post 不应进入拆分/解构")


def test_pipeline_done_path():
    store = FakeStore()
    results = run_pipeline(
        ["https://www.xiaohongshu.com/explore/" + CID],
        settings=_settings(), ingest_enabled=False, store=store,
        fetch_fn=_fetch, extract_fn=_extract, chat=_chat_pass,
    )
    r = results[0]
    assert r["status"] == "done" and r["items"] == 1
    pid = f"xhs_{CID}"
    assert store.posts[pid]["stage"] == "done"
    assert len(store.items) == 1
    saved = store.items[0]
    assert saved["payload"]["source"]["id"] == pid
    assert saved["payload"]["dim_attributes"] == ["how"]


def test_pipeline_rejected_path():
    store = FakeStore()
    results = run_pipeline(
        [CID], settings=_settings(), ingest_enabled=False, store=store,
        fetch_fn=_fetch, extract_fn=_extract, chat=_chat_reject,
    )
    r = results[0]
    assert r["status"] == "rejected"
    assert store.posts[f"xhs_{CID}"]["stage"] == "rejected"
    assert store.items == []