Skip to content

armgit / peach

Public Imported from GitHub
0

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

fix(catalog): 修正艺人外部入口与全库中文名复核

longmeidao9570e363f7e5
diff --git a/docs/REUSE.md b/docs/REUSE.md
index 5f13353ca..dc3f3662c 100644
--- a/docs/REUSE.md
+++ b/docs/REUSE.md
@@ -352,6 +352,7 @@ A 盘垃圾隔离使用短名称 `peach-purge-<32 位 UUID>.peach-quarantine`,
- 口味页顶部给出结论与可点入口:浏览与 Peach 两侧的共同信号、可探索标签、待补证据的下一步动作。
- 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。
- 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/scripts/audit_performer_entry_names.py b/scripts/audit_performer_entry_names.py
new file mode 100644
index 000000000..bfb364f5a
--- /dev/null
+++ b/scripts/audit_performer_entry_names.py
@@ -0,0 +1,90 @@
+#!/usr/bin/env python3
+"""只读审计全库艺人的 MISSAV 入口与中文名证据缺口。"""
+from __future__ import annotations
+
+import argparse
+import json
+import re
+import sqlite3
+import sys
+from collections import Counter, defaultdict
+from pathlib import Path
+from urllib.parse import quote
+
+PROJECT_ROOT = Path(__file__).resolve().parents[1]
+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.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")
+KANA = re.compile(r"[\u3040-\u30ff]")
+HAN = re.compile(r"[\u3400-\u9fff]")
+LATIN = re.compile(r"[A-Za-z]")
+
+
+def collect(connection: sqlite3.Connection) -> 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)
+ refs: dict[int, list[dict]] = defaultdict(list)
+ sources: dict[int, set[str]] = defaultdict(set)
+ assets: dict[int, set[int]] = defaultdict(set)
+ for row in connection.execute("SELECT entity_id,alias,source FROM entity_alias ORDER BY entity_id,alias"):
+ aliases[int(row["entity_id"])].append(dict(row))
+ for row in connection.execute("SELECT entity_id,provider,external_kind,external_id FROM entity_external_ref"):
+ refs[int(row["entity_id"])].append(dict(row))
+ for row in connection.execute("SELECT entity_id,asset_id,source FROM asset_entity"):
+ sources[int(row["entity_id"])].add(str(row["source"]))
+ assets[int(row["entity_id"])].add(int(row["asset_id"]))
+ out = []
+ for entity in connection.execute("SELECT id,canonical_name FROM entity WHERE kind='performer' ORDER BY id"):
+ entity_id = int(entity["id"])
+ name = str(entity["canonical_name"])
+ chain = aliases[entity_id]
+ jav = entry_links.is_jav_performer(refs[entity_id])
+ previous = entry_links.japanese_name("", chain) or name
+ selected = entry_links.japanese_name(name, chain)
+ 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"]))})
+ release = bool(RELEASE_SOURCES.intersection(sources[entity_id])) or jav
+ status = ("非发行艺人,不判中译" if not release else
+ "假名规范名,有汉字别名待复核" if KANA.search(name) and chinese else
+ "假名规范名,中文来源未取得" if KANA.search(name) else
+ "拉丁规范名,需区分艺名与转写" if LATIN.search(name) else
+ "无假名规范名,通用中译覆盖未验证")
+ out.append({"entity_id": entity_id, "current_name": name,
+ "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 "",
+ "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"})
+ return out
+
+
+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")
+ args = parser.parse_args(argv)
+ with open_readonly(args.db) as connection:
+ rows = collect(connection)
+ 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),
+ "cn_status": dict(Counter(row["cn_status"] for row in rows)),
+ "review_csv": str(args.out)}, ensure_ascii=False))
+ return 0
+
+
+if __name__ == "__main__":
+ raise SystemExit(main())
diff --git a/scripts/localize_performer_names.py b/scripts/localize_performer_names.py
index 7ab25e70d..bd2310f5c 100644
--- a/scripts/localize_performer_names.py
+++ b/scripts/localize_performer_names.py
@@ -2,7 +2,7 @@
# -*- coding: utf-8 -*-
r"""用经维护的中日姓名映射本地化番号体系女优名称,并把残留的日本字形归一为简体。
-译名只处理带发行来源的实体,包括 Javinizer 的发行来源;账号型 performer 不翻译。
+译名只处理带发行来源或明确 JAV 目录身份的实体,包括 Javinizer 的发行来源;账号型 performer 不翻译。
匹配按精确证据分层:既有人工复核日文名、当前日文名、既有别名,再到复核表里的旧名。
一条名字命中多个映射时不猜。多个 ledger 实体落到同一映射条目时,按同一人的旧名合并。
@@ -36,6 +36,7 @@ if str(SRC_DIR) not in sys.path:
from peach import javdb
from peach.config import DATA_ROOT, GENERATED_DIR
from peach.entities import merge_entity, normalize_entity_name
+from peach.entry_links import JAV_DIRECTORIES
from peach.kanji import JP_KANJI_TO_SIMPLIFIED, simplify_kanji # noqa: F401 - 供测试与脚本共用
from peach.review_csv import read_rows, write_rows
from peach.scripting import add_ledger_write_args, counts_of, open_for_write, verify_after_write
@@ -110,7 +111,7 @@ def read_mapping(path: Path) -> list[ActorMapping]:
def read_mapping_csv(path: Path, accept: Sequence[str] = ("ok",)) -> list[ActorMapping]:
- """读 `harvest_javdb_cn_names.py` 产出的映射,只收 `accept` 里那几种判定的行。
+ """读资料页复核映射,只收 `accept` 里那几种判定的行。
默认只收 `ok`。站上已经改了名的(`改艺名`、`旧名`)要不要跟着改,是人的决定,
所以得在命令行里点名——判定名照抄 CSV 里那一列的字面值。
@@ -177,6 +178,17 @@ def _target_name(mapping: ActorMapping) -> tuple[str, str]:
return mapping.zh_cn or mapping.jp, "zh_cn" if mapping.zh_cn else "jp-fallback"
+def release_performer_ids(connection: sqlite3.Connection) -> set[int]:
+ """发行来源或 JAV 目录明确收录的实体编号,供译名范围判定使用。"""
+ source_marks = ",".join("?" for _ in RELEASE_SOURCES)
+ directory_marks = ",".join("?" for _ in JAV_DIRECTORIES)
+ return {int(row[0]) for row in connection.execute(
+ f"SELECT entity_id FROM asset_entity WHERE source IN ({source_marks}) UNION "
+ "SELECT entity_id FROM entity_external_ref WHERE external_kind='performer' "
+ f"AND trim(external_id)!='' AND provider IN ({directory_marks})",
+ (*RELEASE_SOURCES, *JAV_DIRECTORIES))}
+
+
def collect(
connection: sqlite3.Connection,
mappings: list[ActorMapping],
@@ -218,12 +230,12 @@ def collect(
FROM entity e WHERE e.kind='performer' ORDER BY e.id
"""
).fetchall()
+ release_ids = release_performer_ids(connection)
rows: list[dict[str, object]] = []
resolved_groups: dict[int, list[dict[str, object]]] = defaultdict(list)
for entity in entities:
entity_id = int(entity["id"])
sources = str(entity["sources"] or "")
- source_set = set(filter(None, sources.split(",")))
base: dict[str, object] = {
"entity_id": entity_id,
"current_name": str(entity["canonical_name"]),
@@ -240,7 +252,7 @@ def collect(
"_aliases": set(aliases.get(entity_id, [])),
"_mapping_index": None,
}
- if not RELEASE_SOURCES.intersection(source_set):
+ if entity_id not in release_ids:
base.update(resolution="non-release performer", action="keep-non-release")
rows.append(base)
continue
@@ -543,6 +555,10 @@ def build_parser() -> argparse.ArgumentParser:
# javdb 资料页取回的映射(`harvest_javdb_cn_names.py`)。它覆盖的是假名规范名,
# XML 那份覆盖的是已收录艺人,两份不冲突也不必同时给。
parser.add_argument("--mapping-csv", type=Path)
+ parser.add_argument("--mapping-source", default="",
+ help="CSV 的复核来源前缀;留空表示 javdb:performer-cn")
+ parser.add_argument("--only", type=int, action="append", default=[],
+ help="只处理指定 entity id,可重复;同一映射合并组必须完整点名")
# 站上已经改了名的要不要跟着改,是人的决定,所以判定要在命令行里点名。
parser.add_argument(
"--accept", default="ok",
@@ -564,11 +580,23 @@ def run(args: argparse.Namespace) -> int:
mappings = read_mapping(args.mapping_xml)
elif args.mapping_csv:
accept = [part.strip() for part in str(args.accept).split(",") if part.strip()]
- mappings, prefix = read_mapping_csv(args.mapping_csv, accept), JAVDB_ALIAS_SOURCE
+ mappings = read_mapping_csv(args.mapping_csv, accept)
+ prefix = args.mapping_source.strip() or JAVDB_ALIAS_SOURCE
identity_review = read_identity_review(args.identity_review)
connection = open_for_write(args)
try:
rows = collect(connection, mappings, identity_review, args.mapping_revision)
+ if args.only:
+ selected = set(args.only)
+ groups = defaultdict(set)
+ for row in rows:
+ if row["mapping_key"]:
+ groups[str(row["mapping_key"])].add(int(row["entity_id"]))
+ rows = [row for row in rows if int(row["entity_id"]) in selected]
+ for row in rows:
+ if row["action"] in {"merge-drop", "merge-and-localize"} \
+ and not groups[str(row["mapping_key"])].issubset(selected):
+ row.update(action="conflict", resolution="partial-merge-group-selection")
write_csv(args.review_csv, rows)
print(f"已审计 performer {len(rows)} 位;复核 CSV:{args.review_csv}")
print(" 动作分布:", dict(Counter(str(row["action"]) for row in rows)))
@@ -597,6 +625,8 @@ def main(argv: list[str] | None = None) -> int:
args = parser.parse_args(argv)
if args.mapping_xml and args.mapping_csv:
parser.error("两份映射只能给一份:混着跑的话,别名来源记的是哪一份就说不清了")
+ if args.mapping_source and not args.mapping_csv:
+ parser.error("--mapping-source 必须和 --mapping-csv 一起使用")
if (args.mapping_xml or args.mapping_csv) and not args.mapping_revision:
parser.error("给了映射就必须给 --mapping-revision,别名要记得住来源")
args.mapping_revision = args.mapping_revision or KANJI_ONLY_REVISION
diff --git a/src/peach/entry_links.py b/src/peach/entry_links.py
index 70b9a97c4..1428c285f 100644
--- a/src/peach/entry_links.py
+++ b/src/peach/entry_links.py
@@ -14,9 +14,8 @@
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。** 它的路径就是名字,而账本里 794 位 performer 有 601 位的规范名
-是中文译名,拿译名拼出来的是一个不存在的页面(实测 `/cn/actresses/释爱丽丝` 返回 404,
-`/cn/actresses/釈アリス` 返回 200)。所以这一站从别名里挑日文艺名:
+**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
+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:
- 含假名才算日文写法。汉字分不出中日——`佐々木さき` 认得出,纯汉字的艺名认不出,
那种只能退回规范名。
@@ -50,6 +49,7 @@ from urllib.parse import quote
from filelock import FileLock
from .fsutil import atomic_write_text
+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"
@@ -100,7 +100,7 @@ SITES: tuple[EntrySite, ...] = (
"javdb.com", "/actors/{javdb_id}",
SLOT_MARK, "mark-javdb", jav_only=True, mirrored=True),
EntrySite("missav", "MISSAV", "MISSAV", "name", "",
- "missav.ws", "/cn/actresses/{name}",
+ "missav.ws", "/dm3/cn/actresses/{name}",
SLOT_MARK, "", jav_only=True, mirrored=True),
)
_BY_KEY = {site.key: site for site in SITES}
@@ -111,7 +111,7 @@ _HOST_SHAPE = re.compile(r"[A-Za-z0-9]([A-Za-z0-9-]*[A-Za-z0-9])?"
#: 第二枚起的序号。用完退回括号数字,标签宁可长一点也不能两枚长得一模一样。
_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")
+_STAGE_NAME_SOURCES = ("avdb-actor-mapping", "javdb", "wiki", "performer-localization")
def defaults() -> dict:
@@ -236,6 +236,13 @@ def _only_hiragana(written: str) -> bool:
def japanese_name(canonical_name: str, aliases) -> str:
"""这位的日文艺名。挑不出来时退回规范名,判据见模块开头。"""
name = str(canonical_name or "").strip()
+ if _has_kana(name) and not _only_hiragana(name):
+ equivalents = [name]
+ for row in aliases or ():
+ written = str((row.get("alias") if isinstance(row, dict) else row) or "").strip()
+ if written and fold_glyphs(written) == fold_glyphs(name):
+ equivalents.append(written)
+ return max(equivalents, key=lambda value: sum(char in JP_KANJI_TO_SIMPLIFIED 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()
@@ -255,6 +262,11 @@ def is_jav_performer(refs) -> bool:
return any(provider in JAV_DIRECTORIES for provider in provider_ids(refs))
+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(JP_KANJI_TO_SIMPLIFIED))
+
+
def _ordinal(index: int) -> str:
"""同一个站点第二枚起的序号,第一枚没有。
@@ -279,7 +291,7 @@ def build(settings: dict, canonical_name: str, refs, aliases=()) -> list[dict]:
row = settings.get(site.key) or {}
if site.jav_only and not jav:
continue
- values = ids.get(site.provider, []) if site.provider else ([written] if written else [])
+ 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):
host = site.host
diff --git a/tests/test_entry_links.py b/tests/test_entry_links.py
index 6ff60b527..b8dd6e748 100644
--- a/tests/test_entry_links.py
+++ b/tests/test_entry_links.py
@@ -76,7 +76,7 @@ class EntryLinkAddressTests(unittest.TestCase):
def test_missav_uses_the_japanese_stage_name(self):
"""规范名是中文译名时,MISSAV 那条按别名里的日文艺名拼。
- 实测 `/cn/actresses/釈アリス` 返回 200,中文译名与纯平假名读音都是 404。
+ 中文译名与纯平假名读音不作为混合文字艺名的入口。
"""
rows = entry_links.build(entry_links.defaults(), "释爱丽丝",
[ref("r18dev", "1096544")], [
@@ -86,7 +86,26 @@ class EntryLinkAddressTests(unittest.TestCase):
])
self.assertEqual(
rows[0]["url"],
- "https://missav.ws/cn/actresses/%E9%87%88%E3%82%A2%E3%83%AA%E3%82%B9")
+ "https://missav.ws/dm3/cn/actresses/%E9%87%88%E3%82%A2%E3%83%AA%E3%82%B9")
+
+ def test_current_japanese_name_wins_over_a_former_name_alias(self):
+ rows = entry_links.build(entry_links.defaults(), "逢沢みゆ",
+ [ref("javdb", "zK98J")], [
+ alias("あいざわみゆ", "r18dev:performer-profile"),
+ alias("坂本マリア", "auto:performer-alias@2048"),
+ ])
+ self.assertEqual(next(row["url"] for row in rows if row["site"] == "missav"),
+ "https://missav.ws/dm3/cn/actresses/"
+ "%E9%80%A2%E6%B3%BD%E3%81%BF%E3%82%86")
+
+ def test_plain_hiragana_canonical_reading_uses_a_stage_name_alias(self):
+ self.assertEqual(entry_links.japanese_name("あいざわみゆ", [alias("逢沢みゆ")]),
+ "逢沢みゆ")
+
+ def test_glyph_equivalent_alias_preserves_the_japanese_spelling(self):
+ self.assertEqual(entry_links.japanese_name("蓝川みれい", [
+ alias("藍川みれい", "avdb-actor-mapping"), alias("坂本マリア"),
+ ]), "藍川みれい")
def test_a_reading_in_plain_hiragana_loses_to_the_written_stage_name(self):
rows = entry_links.build(entry_links.defaults(), "天川空",
@@ -111,7 +130,7 @@ class EntryLinkAddressTests(unittest.TestCase):
[alias("Karen Yuzuriha", "r18:performer")])
self.assertEqual(
rows[0]["url"],
- "https://missav.ws/cn/actresses/%E6%A5%AA%20%E3%82%AB%E3%83%AC%E3%83%B3")
+ "https://missav.ws/dm3/cn/actresses/%E6%A5%AA%20%E3%82%AB%E3%83%AC%E3%83%B3")
def test_ids_from_another_entity_kind_do_not_count(self):
# 事务所那 55 条 minnano-av 引用记的是 `production`,不是这个人在站上的号;
diff --git a/tests/test_fastapi_api.py b/tests/test_fastapi_api.py
index 546d4544e..a8d8fc70f 100644
--- a/tests/test_fastapi_api.py
+++ b/tests/test_fastapi_api.py
@@ -1246,7 +1246,7 @@ class FastApiContractTests(unittest.IsolatedAsyncioTestCase):
self.assertEqual([row["site"] for row in entries], ["javdb", "missav"])
self.assertEqual([row["slot"] for row in entries], ["mark", "mark"])
self.assertEqual(entries[0]["url"], "https://javdb.com/actors/NPD3")
- self.assertTrue(entries[1]["url"].startswith("https://missav.ws/cn/actresses/%"))
+ self.assertTrue(entries[1]["url"].startswith("https://missav.ws/dm3/cn/actresses/%"))
async def test_review_queue_is_readable_and_decisions_are_persisted(self):
response = await self.client.get("/api/review?t=secret")
diff --git a/tests/test_performer_localization.py b/tests/test_performer_localization.py
index 43666227e..23d5de4ec 100644
--- a/tests/test_performer_localization.py
+++ b/tests/test_performer_localization.py
@@ -1,5 +1,6 @@
"""番号体系女优中译:来源匹配、同人合并、日本字形归一与兼容标签同步。"""
import csv
+import json
import sqlite3
import tempfile
import unittest
@@ -8,6 +9,8 @@ from pathlib import Path
import opencc
from peach.migrations import upgrade
+from peach import entry_links
+from scripts.audit_performer_entry_names import collect as audit_entries
from scripts.localize_performer_names import (
JP_KANJI_TO_SIMPLIFIED, KANJI_ALIAS_SOURCE, apply_rows, collect, main,
read_identity_review, read_mapping, report_conflicts, simplify_kanji,
@@ -254,6 +257,66 @@ class PerformerLocalizationTests(unittest.TestCase):
main(["--db", str(self.db), "--mapping-xml", str(self.mapping),
"--review-csv", str(out)])
+ def test_review_csv_keeps_its_source_and_current_japanese_name(self):
+ root = Path(self.tmp.name).resolve()
+ mapping_csv = root / "mapping.csv"
+ with mapping_csv.open("w", encoding="utf-8", newline="") as handle:
+ writer = csv.DictWriter(handle, fieldnames=("current_name", "adopt", "verdict"))
+ writer.writeheader()
+ writer.writerow({"current_name": "飯岡かなこ", "adopt": "饭冈加奈子", "verdict": "ok"})
+ self.con.execute("INSERT INTO entity_alias VALUES(18,'坂本マリア','坂本マリア','auto:performer-alias',1)")
+ self.con.execute("UPDATE asset_entity SET source='javinizer:avbase:performer' WHERE entity_id=18")
+ self.con.execute("INSERT INTO entity_external_ref(entity_id,provider,external_kind,external_id) "
+ "VALUES(18,'javdb','performer','sample')")
+ self.con.commit()
+ self.assertEqual(main([
+ "--db", str(self.db), "--mapping-csv", str(mapping_csv),
+ "--mapping-source", "review:performer-localization", "--mapping-revision", "review-1",
+ "--only", "18",
+ "--identity-review", str(root / "absent.csv"), "--review-csv", str(root / "plan.csv"),
+ "--apply", "--backup", str(root / "backup.db"),
+ ]), 0)
+ name, metadata_json = self.con.execute("SELECT canonical_name,metadata_json FROM entity WHERE id=18").fetchone()
+ self.assertEqual(name, "饭冈加奈子")
+ self.assertEqual(self.con.execute("SELECT canonical_name FROM entity WHERE id=16").fetchone()[0],
+ "斎藤満里奈")
+ self.assertEqual(json.loads(metadata_json)["name_localization"]["source"],
+ "review:performer-localization")
+ aliases = [{"alias": row[0], "source": row[1]} for row in self.con.execute(
+ "SELECT alias,source FROM entity_alias WHERE entity_id=18")]
+ self.assertEqual(entry_links.japanese_name(name, aliases), "飯岡かなこ")
+
+ def test_mapping_source_requires_a_csv(self):
+ with self.assertRaises(SystemExit):
+ main(["--mapping-source", "review:performer-localization"])
+
+ def test_selected_member_of_a_merge_group_requires_the_whole_group(self):
+ root = Path(self.tmp.name).resolve()
+ out = root / "selected-plan.csv"
+ self.assertEqual(main([
+ "--db", str(self.db), "--mapping-xml", str(self.mapping),
+ "--mapping-revision", "r1", "--identity-review", str(self.review),
+ "--only", "11", "--review-csv", str(out),
+ ]), 1)
+ with out.open(encoding="utf-8-sig", newline="") as handle:
+ rows = list(csv.DictReader(handle))
+ self.assertEqual(len(rows), 1)
+ self.assertEqual(rows[0]["action"], "conflict")
+ self.assertEqual(rows[0]["resolution"], "partial-merge-group-selection")
+
+ def test_full_entry_audit_includes_release_and_account_performers(self):
+ self.con.execute("INSERT INTO entity_alias VALUES(18,'坂本マリア','坂本マリア','auto:performer-alias',1)")
+ self.con.execute("INSERT INTO entity_external_ref(entity_id,provider,external_kind,external_id) "
+ "VALUES(18,'javdb','performer','sample')")
+ rows = {row["entity_id"]: row for row in audit_entries(self.con)}
+ self.assertEqual(len(rows), self.con.execute("SELECT count(*) FROM entity WHERE kind='performer'").fetchone()[0])
+ self.assertEqual(rows[18]["entry_name"], "飯岡かなこ")
+ self.assertEqual(rows[18]["name_changed"], 1)
+ self.assertIn("/dm3/cn/actresses/", rows[18]["url"])
+ self.assertEqual(rows[18]["cn_status"], "假名规范名,中文来源未取得")
+ self.assertEqual(rows[14]["cn_status"], "非发行艺人,不判中译")
+ self.assertEqual(rows[14]["url"], "")
+
class KanjiTableAgainstOpenCCTests(unittest.TestCase):
"""字形表逐字对 opencc 复核,手写的对应关系不作数。