Skip to content

armgit / peach

Public Imported from GitHub
0

Peach: Local-first video metadata scraper and media manager.

fix(catalog): 按站点名录生成 MISSAV 艺人入口

longmeidao2a1215b64102
diff --git a/docs/PRODUCT_BACKLOG.md b/docs/PRODUCT_BACKLOG.md
index 953cd68d9..b40329cf4 100644
--- a/docs/PRODUCT_BACKLOG.md
+++ b/docs/PRODUCT_BACKLOG.md
@@ -90,7 +90,7 @@
6. **口味证据持续刷新**:ledger 已实时记录搜索、播放、高潮、喜欢/理由、不合口味和稍后看;浏览器历史现可用 SQLite 一致性副本增量进入私有源库,并生成不含 URL/标题的 creator/tag candidate 与聚合报告。旧 2026-08-13 原始包已确认不在 Windows 外置盘;仍需在 Mac 开启 iCloud Safari、完成首次导入,并把两端每周刷新装成系统计划任务。AI 结论不得直接改真相字段。
7. **扫描与采集任务的参数标定**:无进展预警的 120 秒与单项动作预算(资料 90 秒、封面 240 秒)目前按最坏请求时长取的保守值;等一轮真实任务记录各阶段实测耗时后标定,同时确定完整问题文件的保留周期。
-## 其他开放需求(66 项)
+## 其他开放需求(67 项)
本区包含尚未开始和部分已实现的需求;每项正文说明剩余工作,验收状态见「已核对的实施状态」。
@@ -199,7 +199,9 @@
74. **关注作者建档的后续**(ADR-0096):名册「创作者」档只列有入库作品的实体,纯在线作者只在「在线」档,两档要合成一份;coomer 与 simpcity 来源的作者和发布账号之间补 `operates_account` 关系;Gumroad、Fantia 的归档 id 形状核实后再还原原站账号;已有 F95 来源按新的名片白名单重新解析一次;绑定后删除别名时,已绑的来源要能拆开。
75. **关注读数和卡片数的口径**(低):关注页与创作者资料页在线视图的「N 项更新」按条目数统计;卡片按已加载的发布组折叠,并按视频或图片过滤,只有外链的条目不出现在两档中。读数改为发布组、当前媒体数量,或并列显示「M 组」的产品口径待确定。
-合计:**73 项开放需求**,其中 7 项已有骨架,66 项其他开放需求。已完成的需求不在这里留痕,去 Git 历史查。
+76. **单影片的在线站点入口**:详情页按同片身份展示 MissAV、Jable 等站点的真实影片页面,多个站点与字幕、破解等版次分别保存。复用已有番号归一、版本判据、来源传输与失败状态;界面语言 `cn` 不等于中文字幕,模板地址仅为候选。保存原 href、最终地址、返回番号、版次与语言证据、核验时间和状态。NASSAV 的五源解析可作参照,但其首个命中下载、仅存流地址的模型不能作为已验证入口。完整功能、依赖、许可证与源码差异见 `peach-data/review/nassav-functional-review.md`;本轮仅审阅,尚未实施。
+
+合计:**74 项开放需求**,其中 7 项已有骨架,67 项其他开放需求。已完成的需求不在这里留痕,去 Git 历史查。
## 待执行的操作(41 项)
diff --git a/docs/REUSE.md b/docs/REUSE.md
index 45656ac93..8dae34555 100644
--- a/docs/REUSE.md
+++ b/docs/REUSE.md
@@ -355,7 +355,7 @@ A 盘垃圾隔离使用短名称 `peach-purge-<32 位 UUID>.peach-quarantine`,
- 操作回执复用 Toast(Sonner 的栈,`frontend/src/react/toaster.tsx`;壳里只调 `toast()`/`actionReceipt()`);按钮以 Spinner 和 `aria-busy` 标明忙态。后台任务显示可恢复进度,断线自动重连。
- 实体的统称由用户在资料页自选:菜单只列这条实体名下已有的写法,选中的提为规范名、换下的留成别名,扁平投影跟着改;先过确认弹层并点名两个写法,成功后发可撤销回执;不收自由文本,撞上另一条实体的规范名只报冲突。
- 艺人入口审计复用 `entry_links` 的 JAV 目录闸与日文名选择、只读 SQLite 和 `review_csv`;`audit_performer_entry_names.py` 覆盖全部 performer,分别记录入口模板、名称选择和中文来源缺口,远端页面未请求时保留「未取得」。MISSAV 使用 `/dm3/cn/actresses/` 与当前日文艺名;译名复核后的原日文名按 `performer-localization` 来源优先。`localize_performer_names.py --mapping-source` 标明 CSV 的真实复核来源,`--only` 限定实体并拒绝不完整的同人合并组;别名、兼容标签、来源留痕和备份沿用已有实现。未新增依赖。
- - MISSAV 的艺名选择排除纯平假名读音,保留同名字形与 `々`;当前汉字艺名读取既有成对译名来源,并要求对应当前规范名及相同姓名前缀。可信假名艺名、不同姓氏或拉丁字母来源字段冲突时不套用原名记录。API 与全库审计共用这一选择函数,派生入口不写 ledger;来源冲突留在复核 CSV,403 验证页记「未取得」。
+ - MISSAV 入口优先读取 `state/missav-actresses.json` 的实际名录 href,保留原字形、数字前缀与括号艺名;字形折叠只用于名称匹配。同名对应不同路径时不显示入口;未采全名录的缺项保留模板候选,不能判为站点未收录。七沢みあ、逢沢みゆ的 h1 与 canonical 已在用户 Chrome 验证,证据在 `peach-data/review/missav-native-slug-browser-evidence.json`;名录原始分页和全库匹配 CSV 在 `review/missav-directory/`。API 与全库审计共用 `entry_links.build`,派生入口不写 ledger,403 验证页记「未取得」。艺名选择沿用现名、读音排除与来源冲突复核。无新增依赖。
- FC2 作品计数复用 `ordered_multipart_items`,`web_catalog.video_work_key` 按账本代次缓存已确认的合集键;资料、名册、同台艺人、标签与补全共用。`/api/items.work_total` 供作品读数,`total` 仍用于文件分页;不确定分段不归组。出演名单展开复用 BoardUI `Button`。跨角色人工合并走 `merge_entity`,`user:identity-merge:<kind>` 的别名与墓碑共同证明旧资料页地址,单纯同名不跨角色跳转。
- 馆藏分组缓存声明 `asset`、`asset_entity`、`entity` 依赖;下载进度和复核留痕不使分组重算,同键计算串行共用。作品计数通过 SQLite `create_function` 注册资产到作品的只读字典查找,沿用既有版次与分卷判据。下载后处理复用已绑定资产、整理备份及逐项外键基线比较(`performer_alias_followup`);单部采集只结算本资产候选,进度在 `state/download-processing/`,共享 CSV 仍受全库处理锁保护。资源查询缓存拒收缺字段和来源失败的旧结果。无新增依赖。
- Application 管理路由与资源,域控制器适配原生宿主。文件检查含本地与网盘,来源等分、共用确认弹层;CloudDrive 分档建议首启与配置页共用。
diff --git a/docs/STATUS.md b/docs/STATUS.md
index 8777b8cac..6a24ec378 100644
--- a/docs/STATUS.md
+++ b/docs/STATUS.md
@@ -10,7 +10,7 @@
- 公司资料、FC2 卖家按来源补空;厂牌去重 6 组。回执 `peach-data/review/`。
- 核查:`peach-data/review/creator-boundary-delivery.md`;StraplessDildo 归厂牌。
- 关注作者 25 位、来源 111 条、身份关联 71 条、别名 30 条(ADR-0096)。
-- 验收:`peach-data/review/portrait-loading-production-verification.json`。
+- 验收:`peach-data/review/entity-cold-missav-verification.json`。
- 身份合并 18 组;版本核查:`peach-data/review/version-evidence-review.md`。
- JAV 头像 817,水印待复核 38;档案 17、候选 253、官方封面 2;[选图规则](adr/0102-profile-avatar-priority.md)。
- 「整理」(ADR-0039)支持预览、执行、回滚;英文视频名分词 169 条。
diff --git a/scripts/audit_performer_entry_names.py b/scripts/audit_performer_entry_names.py
index 9d877a62b..6f947aa8b 100644
--- a/scripts/audit_performer_entry_names.py
+++ b/scripts/audit_performer_entry_names.py
@@ -16,14 +16,14 @@ sys.path.insert(0, str(PROJECT_ROOT))
sys.path.insert(0, str(PROJECT_ROOT / "src"))
from peach import entry_links # noqa: E402
-from peach.config import DATABASE_PATH, DATA_ROOT # noqa: E402
+from peach.config import DATABASE_PATH, DATA_ROOT, STATE_DIR # noqa: E402
from peach.review_csv import write_rows # noqa: E402
from peach.scripting import open_readonly # noqa: E402
from localize_performer_names import RELEASE_SOURCES # noqa: E402
FIELDS = ("entity_id", "current_name", "assets", "jav_directory", "previous_name",
"entry_name", "previous_url", "url", "name_changed", "cn_status", "cn_aliases",
- "aliases", "link_evidence")
+ "aliases", "directory_status", "link_evidence")
KANA = re.compile(r"[\u3040-\u30ff]")
HAN = re.compile(r"[\u3400-\u9fff]")
LATIN = re.compile(r"[A-Za-z]")
@@ -37,7 +37,19 @@ def _name_localization(raw: str):
return metadata.get("name_localization") if isinstance(metadata, dict) else None
-def collect(connection: sqlite3.Connection) -> list[dict]:
+def _directory_evidence(jav: bool, directory, selected: str, name: str) -> tuple[str, str]:
+ """名录匹配与目标页面请求是两个独立的证据状态。"""
+ if not jav:
+ return "不适用", "不适用"
+ path = directory.path(selected, name) if directory else None
+ if path is None:
+ return "未取得", "未取得:未取得名录条目,目标未逐页请求"
+ if not path:
+ return "需复核", "未取得:名录无唯一匹配"
+ return "已匹配", "已取得:名录 href,目标未逐页请求"
+
+
+def collect(connection: sqlite3.Connection, directory=None) -> list[dict]:
"""每位 performer 一行;入口生成正确与远端页面可访问分别留证。"""
connection.row_factory = sqlite3.Row
aliases: dict[int, list[dict]] = defaultdict(list)
@@ -59,6 +71,7 @@ def collect(connection: sqlite3.Connection) -> list[dict]:
jav = entry_links.is_jav_performer(refs[entity_id])
previous = entry_links.japanese_name("", chain) or name
selected = entry_links.japanese_name(name, chain, _name_localization(entity["metadata_json"]))
+ directory_status, link_evidence = _directory_evidence(jav, directory, selected, name)
chinese = sorted({str(row["alias"]) for row in chain
if HAN.search(str(row["alias"])) and not KANA.search(str(row["alias"]))
and not LATIN.search(str(row["alias"]))})
@@ -72,10 +85,13 @@ def collect(connection: sqlite3.Connection) -> list[dict]:
"assets": len(assets[entity_id]), "jav_directory": int(jav),
"previous_name": previous if jav else "", "entry_name": selected if jav else "",
"previous_url": f"https://missav.ws/cn/actresses/{quote(previous, safe='')}" if jav else "",
- "url": f"https://missav.ws/dm3/cn/actresses/{quote(entry_links.missav_name(selected), safe='')}" if jav else "",
+ "url": next((row["url"] for row in entry_links.build(
+ entry_links.defaults(), name, refs[entity_id], chain,
+ _name_localization(entity["metadata_json"]), directory) if row["site"] == "missav"), ""),
"name_changed": int(jav and previous != selected), "cn_status": status,
"cn_aliases": "|".join(chinese), "aliases": "|".join(row["alias"] for row in chain),
- "link_evidence": "未取得:本轮只读本地库,未逐页请求" if jav else "不适用"})
+ "directory_status": directory_status,
+ "link_evidence": link_evidence})
return out
@@ -83,12 +99,18 @@ def main(argv: list[str] | None = None) -> int:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--db", type=Path, default=DATABASE_PATH)
parser.add_argument("--out", type=Path, default=DATA_ROOT / "review" / "performer-entry-name-audit.csv")
+ parser.add_argument("--missav-directory", type=Path, default=STATE_DIR / entry_links.DIRECTORY_FILENAME)
args = parser.parse_args(argv)
+ directory = None
+ if args.missav_directory.is_file():
+ directory = entry_links.directory_index(json.loads(
+ args.missav_directory.read_text(encoding="utf-8")))
with open_readonly(args.db) as connection:
- rows = collect(connection)
+ rows = collect(connection, directory)
write_rows(args.out, FIELDS, rows)
print(json.dumps({"performers": len(rows), "missav_entries": sum(row["jav_directory"] for row in rows),
"entry_names_changed": sum(row["name_changed"] for row in rows),
+ "directory_status": dict(Counter(row["directory_status"] for row in rows)),
"cn_status": dict(Counter(row["cn_status"] for row in rows)),
"review_csv": str(args.out)}, ensure_ascii=False))
return 0
diff --git a/src/peach/entry_links.py b/src/peach/entry_links.py
index 3def4cde1..7432b495c 100644
--- a/src/peach/entry_links.py
+++ b/src/peach/entry_links.py
@@ -14,9 +14,12 @@
日文、比看标签都硬。`stash` 不算(那是本机 Stash 的 id,库里每个人都有),
`r18:performer_name` 也不算(那是名字映射,不是目录页上的 id)。
-**MISSAV 按日文艺名拼。** 入口使用 `/dm3/cn/actresses/`,汉字字形按简中站投影;当前规范名
+**MISSAV 采用名录中的实际 href。** 名录按当前艺名与规范名匹配,路径保留站点原文。
+未取得对应名录条目时,候选入口使用 `/dm3/cn/actresses/` 与来源艺名;当前规范名
含假名且不是纯读音时优先使用,否则从别名里挑日文艺名。中文译名与旧艺名不能代替当前日文艺名:
+页面语言不决定路径字形;站点 canonical 中的艺名与数字前缀由站点处理。
+
- 日文艺名可含假名或纯汉字。汉字现名采用同名字形或已有成对译名的来源证据;
对应的中文名必须仍是当前规范名,不同姓氏与可信假名艺名的冲突保留给复核。
- 别名里的纯平假名读音不作为入口。`r18:performer` 每位都带一条(`しゃくありす`),
@@ -43,8 +46,9 @@ from __future__ import annotations
import json
import re
from dataclasses import dataclass
+from functools import lru_cache
from pathlib import Path
-from urllib.parse import quote
+from urllib.parse import quote, unquote, urlsplit
from filelock import FileLock
@@ -53,6 +57,8 @@ from .kanji import JP_KANJI_TO_SIMPLIFIED, fold_glyphs
#: 入口设置落在数据根的 state 目录下。它是本机显示偏好,不是账本真相。
FILENAME = "entry-links.json"
+DIRECTORY_FILENAME = "missav-actresses.json"
+DIRECTORY_URL = "https://missav.ws/cn/actresses"
#: `entity_external_ref.external_kind`:人物那一档。
EXTERNAL_KIND = "performer"
@@ -112,7 +118,7 @@ _HOST_SHAPE = re.compile(r"[A-Za-z0-9]([A-Za-z0-9-]*[A-Za-z0-9])?"
_ORDINALS = "②③④⑤⑥⑦⑧⑨"
#: 艺名类的别名来源。按子串认:来源串后面还挂着采集批次。
_STAGE_NAME_SOURCES = ("avdb-actor-mapping", "javdb", "wiki", "performer-localization")
-_MISSAV_GLYPHS = {**JP_KANJI_TO_SIMPLIFIED, "條": "条"}
+_NAME_GLYPHS = {**JP_KANJI_TO_SIMPLIFIED, "條": "条"}
def defaults() -> dict:
@@ -235,7 +241,7 @@ def _only_hiragana(written: str) -> bool:
def _glyph_key(written: str) -> str:
- return fold_glyphs(written).translate(str.maketrans(_MISSAV_GLYPHS))
+ return fold_glyphs(written).translate(str.maketrans(_NAME_GLYPHS))
def _han_name(written: str) -> bool:
@@ -306,7 +312,7 @@ def japanese_name(canonical_name: str, aliases, localization=None) -> str:
if localized or (equivalents and not _published_kana_alias(aliases)) \
or (_has_kana(name) and not _only_hiragana(name)):
return max([current, *equivalents],
- key=lambda value: sum(char in _MISSAV_GLYPHS or char == "々" for char in value))
+ key=lambda value: sum(char in _NAME_GLYPHS or char == "々" for char in value))
chosen: tuple[tuple[int, int, int], str] | None = None
for index, row in enumerate(aliases or ()):
written = str((row.get("alias") if isinstance(row, dict) else row) or "").strip()
@@ -327,9 +333,77 @@ def is_jav_performer(refs) -> bool:
return any(provider in JAV_DIRECTORIES for provider in provider_ids(refs))
+@dataclass(frozen=True)
+class MissavDirectory:
+ """目录实际 href 的只读索引;名称折叠只用于匹配,地址保持站点原文。"""
+
+ paths: dict[str, tuple[str, ...]]
+ complete: bool = False
+
+ def path(self, japanese: str, canonical: str) -> str | None:
+ for name in (japanese, canonical):
+ matches = self.paths.get(_glyph_key(name).casefold(), ())
+ if matches:
+ return matches[0] if len(matches) == 1 else ""
+ return "" if self.complete else None
+
+
+def _published_names(name: str) -> list[str]:
+ names = [name.strip()]
+ combined = re.fullmatch(r"([^()()]+)[((]([^()()]+)[))]", name)
+ if combined:
+ names.extend(part.strip() for part in combined.groups())
+ return names
+
+
+def directory_index(data: object) -> MissavDirectory | None:
+ """目录快照的公开名字与实际 href;非法域名、空名与非人物路径不入索引。"""
+ if not isinstance(data, dict) or data.get("source") != DIRECTORY_URL:
+ return None
+ entries = data.get("entries")
+ if not isinstance(entries, list):
+ return None
+ paths: dict[str, dict[str, str]] = {}
+ for row in entries:
+ if not isinstance(row, dict):
+ continue
+ try:
+ target = urlsplit(str(row.get("url") or ""))
+ except ValueError:
+ continue
+ if target.scheme != "https" or target.netloc != "missav.ws" \
+ or not re.fullmatch(r"/(?:dm\d+/)?cn/actresses/[^/?#]+", target.path):
+ continue
+ slug = unquote(target.path.rsplit("/", 1)[-1])
+ names = _published_names(slug) + _published_names(str(row.get("name") or ""))
+ for name in names:
+ if name:
+ paths.setdefault(_glyph_key(name).casefold(), {}).setdefault(slug, target.path)
+ return MissavDirectory({key: tuple(sorted(values.values())) for key, values in paths.items()},
+ complete=data.get("complete") is True)
+
+
+@lru_cache(maxsize=4)
+def _directory_file(path: str, modified: int, size: int) -> MissavDirectory | None:
+ try:
+ return directory_index(json.loads(Path(path).read_text(encoding="utf-8")))
+ except (OSError, ValueError):
+ return None
+
+
+def read_directory(root: Path) -> MissavDirectory | None:
+ """按文件版本读取名录快照;不发网络请求,不写账本。"""
+ path = Path(root) / DIRECTORY_FILENAME
+ try:
+ stat = path.stat()
+ except OSError:
+ return None
+ return _directory_file(str(path.resolve()), stat.st_mtime_ns, stat.st_size)
+
+
def missav_name(japanese: str) -> str:
- """简中入口的汉字字形,假名、叠字符号与其他艺名字符保留。"""
- return japanese.translate(str.maketrans(_MISSAV_GLYPHS))
+ """未取得名录条目时使用的简中候选艺名。"""
+ return japanese.translate(str.maketrans(_NAME_GLYPHS))
def _ordinal(index: int) -> str:
@@ -346,16 +420,20 @@ def _ordinal(index: int) -> str:
return f"({index + 1})"
-def build(settings: dict, canonical_name: str, refs, aliases=(), localization=None) -> list[dict]:
+def build(settings: dict, canonical_name: str, refs, aliases=(), localization=None,
+ directory: MissavDirectory | None = None) -> list[dict]:
"""这条人物实体能直达的地址,按 SITES 的先后排好。缺 id 的站点不出现在结果里。"""
ids = provider_ids(refs)
jav = is_jav_performer(refs)
written = japanese_name(canonical_name, aliases, localization)
+ missav_path = directory.path(written, canonical_name) if directory else None
out: list[dict] = []
for site in SITES:
row = settings.get(site.key) or {}
if site.jav_only and not jav:
continue
+ if site.key == "missav" and missav_path == "":
+ continue
values = ids.get(site.provider, []) if site.provider else ([missav_name(written)] if written else [])
host = _tidy_host(row.get("host")) if site.mirrored else site.host
if not host or _host_problem(site, host):
@@ -363,6 +441,8 @@ def build(settings: dict, canonical_name: str, refs, aliases=(), localization=No
for index, value in enumerate(values):
ordinal = _ordinal(index)
path = site.path.replace("{" + site.placeholder + "}", quote(value, safe=""))
+ if site.key == "missav" and missav_path:
+ path = missav_path
out.append({"site": site.key,
"label": f"{site.label} {ordinal}".strip(),
"ordinal": ordinal, "slot": site.slot, "mark": site.mark,
@@ -372,7 +452,7 @@ def build(settings: dict, canonical_name: str, refs, aliases=(), localization=No
def entry_links(root: Path, canonical_name: str, refs, aliases=(), localization=None) -> list[dict]:
"""读设置并拼出入口。资料页只调这一个。"""
- return build(read(root), canonical_name, refs, aliases, localization)
+ return build(read(root), canonical_name, refs, aliases, localization, read_directory(root))
def javdb_actor_pages(canonical_name: str, refs) -> list[str]:
diff --git a/tests/test_entry_links.py b/tests/test_entry_links.py
index 8f27ae4f9..e2480d2ac 100644
--- a/tests/test_entry_links.py
+++ b/tests/test_entry_links.py
@@ -5,6 +5,7 @@ import json
import tempfile
import unittest
from pathlib import Path
+from urllib.parse import quote
from peach import entry_links
@@ -134,8 +135,20 @@ class EntryLinkAddressTests(unittest.TestCase):
("梨梨衣", "梨々衣")):
with self.subTest(name=name):
self.assertEqual(entry_links.japanese_name(name, [alias(jp)]), jp)
- self.assertEqual(entry_links.missav_name("北條麻妃"), "北条麻妃")
- self.assertEqual(entry_links.missav_name("乃々果花"), "乃々果花")
+
+ def test_missav_preserves_source_stage_name_glyphs(self):
+ from urllib.parse import unquote, urlsplit
+
+ for name, jp in (("七泽美亚", "七沢みあ"), ("逢泽美优", "逢沢みゆ"),
+ ("北条麻妃", "北條麻妃"), ("乃乃果花", "乃々果花"),
+ ("优梨舞奈", "優梨まいな")):
+ with self.subTest(name=name):
+ directory = entry_links.directory_index({"source": entry_links.DIRECTORY_URL,
+ "entries": [{"name": name, "url": "https://missav.ws/dm3/cn/actresses/" + quote(jp)}]})
+ rows = entry_links.build(entry_links.defaults(), name,
+ [ref("javdb", "sample")], [alias(jp)], directory=directory)
+ address = next(row["url"] for row in rows if row["site"] == "missav")
+ self.assertEqual(unquote(urlsplit(address).path), f"/dm3/cn/actresses/{jp}")
def test_paired_current_han_name_wins_over_an_unreviewed_former_name(self):
localized = {"source": "avdb-actor-mapping", "jp": "水谷心音", "zh_cn": "水谷心音"}
@@ -253,6 +266,22 @@ class EntryLinkSettingsFileTests(unittest.TestCase):
self.assertEqual(rows["missav"]["host"], "")
self.assertEqual(rows["missav"]["default_host"], "missav.ws")
+ def test_directory_file_supplies_the_published_missav_path(self):
+ path = self.root / entry_links.DIRECTORY_FILENAME
+ first = "https://missav.ws/dm102/cn/actresses/" + quote("七沢みあ")
+ path.write_text(json.dumps({"source": entry_links.DIRECTORY_URL,
+ "entries": [{"name": "七泽美亚", "url": first}]}), encoding="utf-8")
+ rows = entry_links.entry_links(self.root, "七泽美亚", [ref("javdb", "NPD3")],
+ [alias("七沢みあ")])
+ self.assertEqual(next(row["url"] for row in rows if row["site"] == "missav"), first)
+ second = first.replace("dm102", "dm1002")
+ path.write_text(json.dumps({"source": entry_links.DIRECTORY_URL,
+ "entries": [{"name": "七泽美亚", "url": second}]}), encoding="utf-8")
+ rows = entry_links.entry_links(self.root, "七泽美亚", [ref("javdb", "NPD3")],
+ [alias("七沢みあ")])
+ self.assertEqual(next(row["url"] for row in rows if row["site"] == "missav"), second)
+
+
def test_a_pasted_address_is_trimmed_down_to_its_domain(self):
"""镜像地址多半是整条复制过来的,前缀与末尾的斜杠不值得弹一条错误。"""
entry_links.save(self.root, self.body(missav={"host": "https://missav.ai/"}))
@@ -275,5 +304,64 @@ class EntryLinkSettingsFileTests(unittest.TestCase):
self.assertEqual(entry_links.read(self.root)["javdb"]["host"], "")
+class MissavDirectoryTests(unittest.TestCase):
+ def directory(self, rows, *, complete=False):
+ return entry_links.directory_index({"source": entry_links.DIRECTORY_URL,
+ "entries": rows, "complete": complete})
+
+ def test_site_glyphs_and_compound_stage_names_use_the_published_href(self):
+ for canonical, jp, display, slug in (
+ ("北条麻妃", "北條麻妃", "北条麻妃", "北条麻妃"),
+ ("森泽佳奈", "飯岡かなこ", "森泽佳奈 (饭冈佳奈子)", "森沢かな(飯岡かなこ)")):
+ with self.subTest(canonical=canonical):
+ address = "https://missav.ws/dm193/cn/actresses/" + quote(slug)
+ directory = self.directory([{"name": display, "url": address}])
+ rows = entry_links.build(entry_links.defaults(), canonical, [ref("javdb", "actor")],
+ [alias(jp)], directory=directory)
+ self.assertEqual(next(row["url"] for row in rows if row["site"] == "missav"), address)
+
+ def test_ambiguous_directory_names_have_no_missav_entry(self):
+ directory = self.directory([
+ {"name": "同名艺人", "url": "https://missav.ws/dm1/cn/actresses/A"},
+ {"name": "同名艺人", "url": "https://missav.ws/dm2/cn/actresses/B"},
+ ])
+ rows = entry_links.build(entry_links.defaults(), "同名艺人", [ref("javdb", "actor")],
+ directory=directory)
+ self.assertEqual([row["site"] for row in rows], ["javdb"])
+
+ def test_prefix_variants_of_the_same_published_slug_share_one_identity(self):
+ directory = self.directory([
+ {"name": "A", "url": "https://missav.ws/dm1/cn/actresses/A"},
+ {"name": "A", "url": "https://missav.ws/dm2/cn/actresses/A"},
+ ])
+ self.assertEqual(directory.path("A", "A"), "/dm1/cn/actresses/A")
+
+ def test_an_incomplete_directory_does_not_assert_an_unlisted_name_is_absent(self):
+ directory = self.directory([])
+ self.assertIsNone(directory.path("七沢みあ", "七泽美亚"))
+ rows = entry_links.build(entry_links.defaults(), "七泽美亚", [ref("javdb", "actor")],
+ [alias("七沢みあ")], directory=directory)
+ self.assertTrue(any(row["site"] == "missav" for row in rows))
+
+ def test_a_complete_directory_has_no_entry_for_an_unmatched_name(self):
+ rows = entry_links.build(entry_links.defaults(), "同名艺人", [ref("javdb", "actor")],
+ directory=self.directory([], complete=True))
+ self.assertEqual([row["site"] for row in rows], ["javdb"])
+
+ def test_directory_entries_are_scoped_to_the_public_actress_origin(self):
+ invalid = ("http://missav.ws/cn/actresses/A", "https://missav.ws.evil.test/cn/actresses/A",
+ "https://[email protected]/cn/actresses/A", "https://missav.ws/cn/videos/A",
+ "https://missav.ws/cn/actresses/A/extra")
+ directory = self.directory([{"name": "A", "url": value} for value in invalid])
+ self.assertEqual(directory.paths, {})
+
+ def test_a_mirror_preserves_the_directory_path(self):
+ settings = entry_links.defaults()
+ settings["missav"]["host"] = "missav.ai"
+ directory = self.directory([{"name": "A", "url": "https://missav.ws/dm102/cn/actresses/A"}])
+ rows = entry_links.build(settings, "A", [ref("javdb", "actor")], directory=directory)
+ self.assertEqual(next(row["url"] for row in rows if row["site"] == "missav"),
+ "https://missav.ai/dm102/cn/actresses/A")
+
if __name__ == "__main__":
unittest.main()
diff --git a/tests/test_performer_localization.py b/tests/test_performer_localization.py
index 23d5de4ec..2ef884d28 100644
--- a/tests/test_performer_localization.py
+++ b/tests/test_performer_localization.py
@@ -317,6 +317,21 @@ class PerformerLocalizationTests(unittest.TestCase):
self.assertEqual(rows[14]["cn_status"], "非发行艺人,不判中译")
self.assertEqual(rows[14]["url"], "")
+ def test_entry_audit_uses_the_source_glyphs_for_missav(self):
+ self.con.execute("UPDATE entity SET canonical_name='七泽美亚',metadata_json='{}' WHERE id=18")
+ self.con.execute("DELETE FROM entity_alias WHERE entity_id=18")
+ self.con.execute("INSERT INTO entity_alias VALUES(18,'七沢みあ','七沢みあ','javdb-actor-page',1)")
+ self.con.execute("INSERT INTO entity_external_ref(entity_id,provider,external_kind,external_id) "
+ "VALUES(18,'javdb','performer','sample')")
+ directory = entry_links.directory_index({"source": entry_links.DIRECTORY_URL,
+ "entries": [{"name": "七泽美亚", "url": "https://missav.ws/dm3/cn/actresses/"
+ "%E4%B8%83%E6%B2%A2%E3%81%BF%E3%81%82"}]})
+ rows = {row["entity_id"]: row for row in audit_entries(self.con, directory)}
+ self.assertEqual(rows[18]["entry_name"], "七沢みあ")
+ self.assertEqual(rows[18]["directory_status"], "已匹配")
+ self.assertEqual(rows[18]["url"], "https://missav.ws/dm3/cn/actresses/"
+ "%E4%B8%83%E6%B2%A2%E3%81%BF%E3%81%82")
+
class KanjiTableAgainstOpenCCTests(unittest.TestCase):
"""字形表逐字对 opencc 复核,手写的对应关系不作数。