fix(parser): 原盘判定参考副标题线索与 PCM 音轨
标题写成 `Bluray H.264` 这类压制字样、但副标题写「自购自抓自扫」「BDMV」「原盘」「DIY」 「Blu-ray Disc」的，或带 LPCM / PCM 音轨又没有 x264 / x265 / BDRip / 压制说法的，判为原盘。 只在来源是 BluRay（或没写来源但有分辨率）时补判；「原盘压制」「自压」、WEB-DL、Remux、单个视频文件不受影响。 用「药屋少女的呢喃」真实搜索结果补正例与反例。 Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_01Cxwgnn2pbzQ7rXv3rq7fei
This commit is contained in:
rootandClaude Opus 5.5 committed
1 parent
1845d47d3f
commit
18ee466664
2 files changed
+142
-1
No files matched your search
@@ -45,6 +45,8 @@ type Draft = ParsedRelease & {
|
||||
_fin?: boolean;
|
||||
_disc?: boolean;
|
||||
_remux?: boolean;
|
||||
/** 标题里有 BDRip / BRRip(压制) */
|
||||
_rip?: boolean;
|
||||
/** `{[k=v;…]}` 显式身份块里的季/集/类型,finalize 时覆盖推断值 */
|
||||
_explicit?: { season?: number; seasonEnd?: number; episodes?: EpisodeRange; tv?: boolean };
|
||||
};
|
||||
@@ -746,6 +748,7 @@ function walkTokens(tokens: string[], d: Draft): void {
|
||||
const cnTitle = cn.join(" ").trim();
|
||||
if (enTitle) d.titleEn = enTitle;
|
||||
if (cnTitle) d.titleCn = cnTitle;
|
||||
if (ripWord) d._rip = true;
|
||||
if (d._remux) d.source = "REMUX";
|
||||
else if (d._disc) d.source = "BluRay-Disc";
|
||||
else if (blurayWord && discCodec && !encodeCodec && !ripWord && !d.encoder && sources.size === 1) {
|
||||
@@ -1077,9 +1080,40 @@ export function parseRelease(input: string, opts: ParseOptions = {}): ParsedRele
|
||||
}
|
||||
}
|
||||
|
||||
// 14. 原盘线索:标题的编码字样不足以判定时,参考副标题 / 简介里的明确说法与无损 PCM 音轨
|
||||
refineDisc(d, s, subtitle);
|
||||
|
||||
return finalize(d, opts);
|
||||
}
|
||||
|
||||
/** 明确说是原盘:原盘(不含「原盘压制」)、BDMV、自购自抓 / 自抓自扫、Blu-ray Disc、DIY(原盘 DIY) */
|
||||
const DISC_CUE =
|
||||
/原[盘盤](?![\s:：]*(?:压|壓|rip|转|轉))|BDMV|自[购購]自抓|自抓自[扫掃]|Blu-?ray[\s.]+Disc|(?<![A-Za-z])DIY(?![A-Za-z])/i;
|
||||
/** 明确说是压制或其他来源 */
|
||||
const ENCODE_CUE =
|
||||
/自[压壓]|[压壓]制|重[编編]码|(?<![A-Za-z])(?:[xX]\.?26[45]|BDRip|BRRip|WEB-?DL|WEBRip|HDTV|Remux)(?![A-Za-z])/i;
|
||||
|
||||
/**
|
||||
* 原盘补判(对照 MP 内置规则 BLU 之外的站点惯例):标题写成 `Bluray H.264` 这类压制字样、但副标题写「自购自抓自扫」,
|
||||
* 或带 LPCM / PCM 音轨又没有 x264 / x265 / BDRip / 压制说法的,按原盘。只在来源是 BluRay(或标题没写来源但有分辨率)时补判,
|
||||
* WEB / HDTV / Remux 不动。
|
||||
*/
|
||||
function refineDisc(d: Draft, title: string, subtitle: string | undefined): void {
|
||||
// 单个视频文件(mkv / mp4 …)不会是原盘
|
||||
if (d.isFile || d._rip || d.encoder || d._remux || d.source === "BluRay-Disc") return;
|
||||
if (d.source !== "BluRay" && !(d.source === undefined && d.resolution)) return;
|
||||
const text = subtitle ? `${title} | ${subtitle}` : title;
|
||||
if (ENCODE_CUE.test(text)) return;
|
||||
const pcm = d.source === "BluRay" && d.audio.some((a) => /^L?PCM\b/i.test(a));
|
||||
if (DISC_CUE.test(text)) {
|
||||
d.source = "BluRay-Disc";
|
||||
d.evidence.push("source:disc_hint");
|
||||
} else if (pcm) {
|
||||
d.source = "BluRay-Disc";
|
||||
d.evidence.push("source:disc_pcm");
|
||||
}
|
||||
}
|
||||
|
||||
function applyCnMarkers(d: Draft, cm: CnMarkers, prefix: string): void {
|
||||
const ev = d.evidence;
|
||||
const note = (k: string) => {
|
||||
@@ -1171,7 +1205,7 @@ function finalize(d: Draft, opts: ParseOptions): ParsedRelease {
|
||||
d.title = d.titleCn ?? d.titleEn;
|
||||
d.confidence = confidenceOf(d);
|
||||
|
||||
const { _complete, _fin, _disc, _remux, _explicit, ...out } = d;
|
||||
const { _complete, _fin, _disc, _remux, _rip, _explicit, ...out } = d;
|
||||
return out;
|
||||
}
|
||||
|
||||
|
||||
@@ -627,6 +627,113 @@ describe("MP 差分补迁写法", () => {
|
||||
});
|
||||
});
|
||||
|
||||
// 原盘补判(SEARCH 准确度):站点搜索「药屋少女的呢喃」的真实结果。标题写成压制字样、但副标题说「自购自抓自扫」的是原盘;
|
||||
// LPCM / PCM 音轨且没有 x264 / x265 / BDRip / 压制说法的也是原盘。普通压制、WEB-DL、Remux 不受影响。
|
||||
const DISC_HINTS: Sample[] = [
|
||||
{
|
||||
id: "DISC-sub-self-rip",
|
||||
input: "Kusuriya no Hitorigoto 2025 S02 1080p Bluray H.264 LPCM-@U2",
|
||||
subtitle: "自购自抓自扫 | Team HAYAKU | Animate特典附 | TBR35041D 药屋少女的呢喃 第二季",
|
||||
expect: { source: "BluRay-Disc", isEncode: false },
|
||||
},
|
||||
{
|
||||
id: "DISC-sub-bdmv",
|
||||
input: "The Apothecary Diaries 2025 S02 1080p Blu-ray H.264-GRP",
|
||||
subtitle: "药屋少女的呢喃 第二季 BDMV Vol.1",
|
||||
expect: { source: "BluRay-Disc" },
|
||||
},
|
||||
{
|
||||
id: "DISC-sub-yuanpan",
|
||||
input: "Some Show S01 1080p BluRay H.264 DTS-HD MA 5.1-GRP",
|
||||
subtitle: "某剧 第一季 原盘 中字",
|
||||
expect: { source: "BluRay-Disc" },
|
||||
},
|
||||
{
|
||||
id: "DISC-pcm",
|
||||
input: "Some Movie 2020 1080p BluRay LPCM 2.0-GRP",
|
||||
expect: { source: "BluRay-Disc", isEncode: false },
|
||||
},
|
||||
// 已按编码字样判成原盘的不变
|
||||
{
|
||||
id: "DISC-avc-lpcm",
|
||||
input: "Kusuriya no Hitorigoto S02 2025 Blu-ray 1080p AVC LPCM 2Audio-@U2",
|
||||
subtitle: "Team HAYAKU | Animate全卷特典附 药屋少女的呢喃 第二季",
|
||||
expect: { source: "BluRay-Disc" },
|
||||
},
|
||||
{
|
||||
id: "DISC-bdmv-rip-welcome",
|
||||
input: "The Apothecary Diaries Vol.3 2025 BDMV 1080p AVC LPCM 2.0-Arri-RAWS@U2",
|
||||
subtitle: "U2 Exclusive, RIPs welcomed with credits | 制作RIP请注明出处 | 药屋少女的呢喃/药师少女的独语 S2",
|
||||
expect: { source: "BluRay-Disc" },
|
||||
},
|
||||
// 反例:压制
|
||||
{
|
||||
id: "ENC-self-encode",
|
||||
input: "Kusuriya no Hitorigoto 2023 BDRip 1080p AVC FLAC@U2",
|
||||
subtitle: "药师少女的呢喃/Kusuriya no Hitorigoto/薬屋のひとりごと 1080p 自压(附相关专辑) TV 01-24 Fin+SP",
|
||||
expect: { source: "BluRay", isEncode: true },
|
||||
},
|
||||
{
|
||||
id: "ENC-x265",
|
||||
input: "The Apothecary Diaries S01 2023 1080p BluRay x265 10bit FLAC 2.0-ADE",
|
||||
subtitle: "药屋少女的呢喃 / 药师少女的独语 / The Apothecary Diaries / 薬屋のひとりごと 【简日|繁日|简|繁|日字幕】",
|
||||
expect: { source: "BluRay", isEncode: true },
|
||||
},
|
||||
{
|
||||
id: "ENC-x264",
|
||||
input: "Kusuriya no Hitorigoto S01 Complete 2023 1080p Blu-ray x264 FLAC-jsum@U2",
|
||||
subtitle: "药屋少女的呢喃 / 药师少女的独语 / 薬屋のひとりごと",
|
||||
expect: { source: "BluRay", isEncode: true },
|
||||
},
|
||||
{
|
||||
id: "ENC-av1",
|
||||
input: "Kusuriya no Hitorigoto S02 1080p BluRay AV1 OPUS 2.0-7³ACG@OurBits",
|
||||
subtitle: "药师少女的独语第2季/薬屋のひとりごと 第2期 | 01-24(25-48) 日语+简繁字幕",
|
||||
expect: { source: "BluRay", isEncode: true },
|
||||
},
|
||||
{ id: "ENC-x264-lpcm", input: "Some Movie 2020 1080p BluRay x264 LPCM 2.0-GRP", expect: { source: "BluRay" } },
|
||||
{
|
||||
id: "ENC-from-disc",
|
||||
input: "Some Movie 2020 1080p BluRay H.264 FLAC-GRP",
|
||||
subtitle: "基于原盘压制 简繁字幕",
|
||||
expect: { source: "BluRay" },
|
||||
},
|
||||
{
|
||||
id: "ENC-pcm-encode-note",
|
||||
input: "Some Movie 2020 1080p BluRay H.264 LPCM-GRP",
|
||||
subtitle: "自压 内封字幕",
|
||||
expect: { source: "BluRay" },
|
||||
},
|
||||
// 反例:其他来源不受副标题影响
|
||||
{
|
||||
id: "WEB-sub-yuanpan",
|
||||
input: "Some Movie 2020 1080p WEB-DL H.264 AAC-GRP",
|
||||
subtitle: "原盘 DIY 还没出，先发 WEB",
|
||||
expect: { source: "WEB-DL" },
|
||||
},
|
||||
{
|
||||
id: "REMUX-diy",
|
||||
input: "The Apothecary Diaries S01 2023 1080p BluRay Remux AVC FLAC 2.0-ADE",
|
||||
subtitle: "药屋少女的呢喃 【DIY / 双语简繁字幕】",
|
||||
expect: { source: "REMUX" },
|
||||
},
|
||||
// 标题里写着 2160P 的非发布名(播放器解码信息):按字面解析,不猜
|
||||
{
|
||||
id: "WEB-2160-literal",
|
||||
input:
|
||||
"The Apothecary Diaries 2023 S01E24 2160P TV WEB-DL MPEG-4 AVC1 - Native D3D9 DXVA Decoder VLD AAC LC -Kisssub",
|
||||
subtitle:
|
||||
"药屋少女的呢喃 薬屋のひとりごと (2023) 2160P TVS01 [01-24]（AI增强1080P转2160P） AVC1 @5160kbps AAC @128kbps",
|
||||
expect: { source: "WEB-DL", resolution: "2160p" },
|
||||
},
|
||||
];
|
||||
|
||||
describe("原盘补判(副标题线索 / PCM 音轨)", () => {
|
||||
for (const s of DISC_HINTS) {
|
||||
it(`${s.id} ${s.input}`, () => check(parseRelease(s.input, s.subtitle ? { subtitle: s.subtitle } : {}), s.expect));
|
||||
}
|
||||
});
|
||||
|
||||
describe("parseRelease 样本表", () => {
|
||||
it("样本数量 ≥ 60", () => {
|
||||
expect(TITLES.length).toBeGreaterThanOrEqual(60);
|
||||
|
||||
Reference in new issue
Block a user