Peach: Local-first video metadata scraper and media manager.
fix(browser): MGS 的 SOD 店铺前缀归一为裸番号
longmeidao5de4c36edbf2
diff --git a/src/peach/catalog_rules.py b/src/peach/catalog_rules.pyindex 49d85b271..a1ef86753 100644--- a/src/peach/catalog_rules.py+++ b/src/peach/catalog_rules.py@@ -289,6 +289,10 @@ _DATED_COMPACT = re.compile(rf"^{_MMDDYY}\d{{2,4}}$")_CODE_MAKER_PREFIXED = re.compile(r"^(\d{3})([A-Z]{2,8})-(\d{2,5})$")#: 数字前缀单独保留;`h_` 是 DMM content_id 的 label 标记。_RELEASE_ID = re.compile(r"^(\d{1,4})?([A-Z]{2,8})0*(\d{1,5})([A-Z]{0,2})$")+#: MGStage 给 SOD 作品加的店铺前缀:商品页 `107STARS-325` 与 DMM、javdb 的 `STARS-325` 是同一部,+#: 账本与资源站都只用裸番号。2026-11-01 逐组核过 MGStage 商品页与 javdb 的标题、出演、日期,+#: 账本里这三个字母段只挂 SOD Create。+_STORE_PREFIXES = {("107", "STARS"), ("107", "START"), ("107", "STZY")}# 仅登记已核验的写法关系;查询扩展表不是身份等价表。## MGStage 前缀那 9 行 2026-09-25 逐组核过:MGStage 商品页与 r18dev 快照的标题、时长、@@ -297,7 +301,7 @@ _RELEASE_ID = re.compile(r"^(\d{1,4})?([A-Z]{2,8})0*(\d{1,5})([A-Z]{0,2})$")_RELEASE_PREFIX_ALIASES = {("259", "LUXU"), ("7", "BAZX"), ("49", "HA"),("230", "ORETD"), ("278", "GYAN"), ("336", "KBI"), ("345", "SIMM"), ("413", "INST"),- ("428", "SUKE"), ("435", "MFC"), ("476", "MLA"), ("762", "FKOS"),+ ("428", "SUKE"), ("435", "MFC"), ("476", "MLA"), ("762", "FKOS"), *_STORE_PREFIXES,}_DMM_LABEL_PREFIX = re.compile(r"^H_(?=\d)")_MEDIA_EXTENSION = re.compile(@@ -668,6 +672,15 @@ def code_query_variants(code: str | None) -> tuple[str, ...]:return (primary,)+def without_store_prefix(code: str | None) -> str:+ """去掉发行平台的店铺前缀,得到账本与资源站使用的写法:`107STZY-019` → `STZY-019`。"""+ primary = normalise_code_key(code)+ prefixed = _CODE_MAKER_PREFIXED.fullmatch(primary)+ if prefixed and prefixed.group(1, 2) in _STORE_PREFIXES:+ return f"{prefixed.group(2)}-{prefixed.group(3)}"+ return primary++def release_identity(code: str | None) -> str:"""把番号或来源返回的 id 收敛成可比对的作品身份。diff --git a/src/peach/jav_browser.py b/src/peach/jav_browser.pyindex 8d44c0e97..e5160e4b6 100644--- a/src/peach/jav_browser.py+++ b/src/peach/jav_browser.py@@ -13,7 +13,9 @@ from functools import lru_cachefrom pathlib import Pathfrom urllib.parse import parse_qsl, urlencode, urlsplit-from .catalog_rules import is_jav_code, normalise_code_key, release_code_from_product, same_release_code+from .catalog_rules import (+ is_jav_code, normalise_code_key, release_code_from_product, same_release_code, without_store_prefix,+)from .config import PROJECT_ROOTfrom .downloads import (CLOUDDRIVE_CREDENTIAL, INGESTED, POSTPROCESS_LABELS, REMOTE_RUNNING, RESUBMITTABLE_STATES,@@ -75,7 +77,8 @@ def code_key(value: object) -> str:code = release_code_from_product(raw) if len(raw) <= 80 else Noneif not code or not is_jav_code(code) or normalise_code_key(raw) != normalise_code_key(code):raise ValueError("请提供一个明确的 JAV 番号")- return normalise_code_key(code)+ # 店铺商品号(`107STZY-019`)在账本与资源站都是裸番号,查询、搜索与入库统一用它。+ return without_store_prefix(code)def source_page(value: object) -> str:@@ -158,12 +161,17 @@ class JavBrowser:payload = json.loads(self.path.read_text(encoding="utf-8"))if not isinstance(payload, dict) or len(payload) > 200:raise ValueError("浏览器添加队列无法读取")+ rows = {}for key, row in payload.items():- if not isinstance(row, dict) or code_key(key) != key or row.get("state") not in MESSAGES:+ if not isinstance(row, dict) or normalise_code_key(key) != key or row.get("state") not in MESSAGES:raise ValueError("浏览器添加队列无法读取")if row["state"] == "searching":row["state"] = "queued"- return payload+ # 店铺前缀归一之前记下的行(`107STZY-019`)改记到裸番号下,裸番号已有记录时以它为准。+ canonical = code_key(key)+ if canonical == key or canonical not in rows:+ rows[canonical] = row+ return rowsdef _save(self):atomic_write_text(self.path, json.dumps(self.rows, ensure_ascii=False, indent=2) + "\n")diff --git a/tests/test_downloads_jav_browser.py b/tests/test_downloads_jav_browser.pyindex 34e5e1633..87d05ba5d 100644--- a/tests/test_downloads_jav_browser.py+++ b/tests/test_downloads_jav_browser.py@@ -647,6 +647,37 @@ class BrowserTests(unittest.TestCase):self.service.search.assert_not_called()self.remote.submit.assert_not_called()+ def test_mgstage_store_prefix_finds_the_owned_plain_release(self):+ source = "https://www.mgstage.com/product/product_detail/107STZY-019/"+ with self.app.state.web_contract.database.write_transaction() as connection:+ connection.execute("INSERT INTO asset(path,name,location,medium,code) VALUES(?,?,?,?,?)",+ ("R:\\Media\\STZY-019.mp4", "STZY-019.mp4", "local", "video", "STZY-019"))+ result = self.client.get("/api/jav-browser/identify", headers=self.script_headers,+ params={"source_url": source, "code": "107STZY-019"})+ self.assertEqual(result.status_code, 200, result.text)+ self.assertEqual((result.json()["code"], result.json()["state"]), ("STZY-019", "owned"))+ self.assertEqual(self.add("107STZY-019").json()["state"], "owned")+ self.assertFalse(self.service.process_one())+ self.service.search.assert_not_called()++ def test_mgstage_store_prefix_searches_and_submits_the_plain_release(self):+ self.service.search.return_value = {"items": [{"name": "STZY-019 1080p", "uri": MAGNET}]}+ self.assertEqual(self.add("107STZY-019").json()["code"], "STZY-019")+ self.service.process_one()+ self.assertEqual(self.service.rows["STZY-019"]["state"], "submitted")+ self.service.search.assert_called_once_with(self.app.state.web_contract, {"code": "STZY-019", "goal": "quality"})+ self.assertEqual(self.downloads.store.get(self.service.rows["STZY-019"]["task_id"]).code, "STZY-019")++ def test_queue_rows_saved_under_a_store_prefix_move_to_the_plain_release(self):+ self.service.path.parent.mkdir(parents=True, exist_ok=True)+ self.service.path.write_text(json.dumps({+ "107STZY-019": {"state": "no_match", "provider": "115", "source_url": ""},+ "107STARS-325": {"state": "no_match", "provider": "115", "source_url": ""},+ "STARS-325": {"state": "queued", "provider": "115", "source_url": ""}}), encoding="utf-8")+ restored = JavBrowser(self.app.state.web_contract, search=self.service.search)+ self.assertEqual({key: row["state"] for key, row in restored.rows.items()},+ {"STZY-019": "no_match", "STARS-325": "queued"})+def test_explicit_product_suffix_is_preserved_through_the_download_submission(self):code = "104DANDY-851C"self.service.search.return_value = {"items": [diff --git a/tests/test_jav_code_domain.py b/tests/test_jav_code_domain.pyindex ecfbdd2fe..7edfac1df 100644--- a/tests/test_jav_code_domain.py+++ b/tests/test_jav_code_domain.py@@ -51,6 +51,7 @@ from peach.catalog_rules import (scrapes_as_jav,tokyo_hot_code,western_release_identity,+ without_store_prefix,)from peach import scriptingfrom peach.migrations import upgrade@@ -123,6 +124,13 @@ class NormalisationTests(unittest.TestCase):self.assertFalse(is_jav_asset("HHD800", None, None, ("performer",)))self.assertTrue(is_jav_asset("PBD390", "MOODYZ"))+ def test_store_prefix_is_dropped_only_for_verified_labels(self):+ self.assertEqual(without_store_prefix("107stzy019"), "STZY-019")+ self.assertTrue(same_release_code("107STARS-325", "STARS-325"))+ # 素人系前缀是作品身份的一部分,未核验的前缀仍是另一部作品。+ self.assertEqual(without_store_prefix("259LUXU-1475"), "259LUXU-1475")+ self.assertFalse(same_release_code("348NTR-001", "NTR-001"))+class ScrapeGateTests(unittest.TestCase):"""`scrapes_as_jav` 只拦「创作者作品的文件名被读成番号」,不替代发行证据。"""