Peach: Local-first video metadata scraper and media manager.
feat(follow): 新作接口为没读过资料的新作补读来源资料
longmeidao7fc29bda2e04
diff --git a/src/peach/web_jav_discovery.py b/src/peach/web_jav_discovery.pyindex 18becc792..14a7751cf 100644--- a/src/peach/web_jav_discovery.py+++ b/src/peach/web_jav_discovery.py@@ -128,15 +128,16 @@ def _metadata_response(contract, args, candidates, picked, limit, scope):job_id=args.get("metadata_job_id"))if not job:return {}+ return {**_job_response(job),+ "metadata_coverage": sum(row["metadata_state"] in ("ready", "partial", "stale") for row in picked),+ "metadata_next_cursor": cursor + limit if not picked and cursor + limit < len(candidates) else None}+++def _job_response(job):public_job = {key: value for key, value in job.items() if key not in ("codes", "error")}public_job["message"] = _metadata_message(job)public_job["error"] = public_job["message"] if job["status"] == "failed" else ""- response = {"metadata_job": public_job,- "metadata_coverage": sum(row["metadata_state"] in ("ready", "partial", "stale") for row in picked),- "metadata_next_cursor": cursor + limit if not picked and cursor + limit < len(candidates) else None}- if job["status"] == "running":- response["poll_after_ms"] = 3000- return response+ return {"metadata_job": public_job, **({"poll_after_ms": 3000} if job["status"] == "running" else {})}def _metadata_message(job):@@ -228,7 +229,10 @@ def q_jav_releases(contract, args):候选与「发现」同一批(`_external`),只换成时间顺序、不按口味截断。筛选项只取这批新作里出现过的女优、厂牌与标签;每个维度的计数按「文字与其他维度已生效、本维度不计」的行集算,- 计数为 0 的不列。只读本机缓存的来源资料,不为翻页联网补资料。+ 计数为 0 的不列。++ 筛选结果里还有从没读过来源资料的(`metadata_state` 为 pending),就为它们起一轮资料读取,与「发现」+ 共用同一个有预算的后台任务;读过但没取到的不在此列,不会每次打开都重试。`metadata_job_id` 是轮询。"""try:size = max(1, min(int(args.get("size") or RELEASE_PAGE_SIZE), 100))@@ -287,5 +291,11 @@ def q_jav_releases(contract, args):row["cover_frame"] = contract.cover_frame(row["code"])row["cover_version"] = contract.cover_version(row["code"])row["poster_box"] = contract.poster_box(row["code"])- return {"items": picked, "total": len(shown), "candidate_total": len(rows), "page": page, "size": size,- "facets": {"performers": facets["performer"], "studios": facets["studio"], "tags": facets["tag"]}}+ response = {"items": picked, "total": len(shown), "candidate_total": len(rows), "page": page, "size": size,+ "facets": {"performers": facets["performer"], "studios": facets["studio"], "tags": facets["tag"]}}+ unread = [row["code"] for row in shown if row.get("metadata_state") == "pending"]+ if unread or args.get("metadata_job_id"):+ job = discovery_metadata(contract).ensure(unread, refresh=bool(unread), job_id=args.get("metadata_job_id"))+ if job:+ response.update(_job_response(job))+ return responsediff --git a/tests/test_feeds_web.py b/tests/test_feeds_web.pyindex da9458cce..12a7a3158 100644--- a/tests/test_feeds_web.py+++ b/tests/test_feeds_web.py@@ -1325,14 +1325,17 @@ class JavDiscoveryTest(FeedWebFixture):"HMN-071": {"actors": [{"name": "女优乙"}], "tags": [], "maker_name": "厂牌乙"},"OWN-001": {"actors": [{"name": "馆藏女优"}], "tags": [{"name": "巨乳"}], "maker_name": "馆藏厂牌"}}+ reads = []+def releases(**args):with patch.object(web_jav_discovery, "_cached_movie", side_effect=lambda code, contract: movies.get(code, {})), \- patch("peach.javdb_public.PUBLIC_API.movie") as movie:- result = web_jav_discovery.q_jav_releases(self.contract, args)- movie.assert_not_called()- return result+ patch("peach.jav_discovery_metadata.DiscoveryMetadata.ensure",+ side_effect=lambda codes, **options: reads.append((codes, options))):+ return web_jav_discovery.q_jav_releases(self.contract, args)every = releases()+ self.assertEqual(reads[0], (["SSIS-950", "HMN-071"], {"refresh": True, "job_id": None}),+ "从没读过来源资料的新作当场起一轮读取,筛选项才完整")self.assertEqual(([row["code"] for row in every["items"]], every["total"]), (["SSIS-950", "HMN-071"], 2))self.assertEqual([(ref["name"], ref["count"]) for ref in every["facets"]["performers"]], [("女优乙", 1), ("女优甲", 1)])self.assertEqual([ref["name"] for ref in every["facets"]["studios"]], ["厂牌乙", "厂牌甲"])