ngfchl/guangya_flutter
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

improve media matching toolbar and title parsing

Browse Source
This commit is contained in:
ngfchl
2026-07-18 23:30:10 +08:00
parent e23205166d
commit 9fc69d1832
4 changed files with 504 additions and 260 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/models/media_library.dart
+38 -5
View File
@@ -552,15 +552,33 @@ class ParsedMediaName {
}
String? first(String pattern) =>
RegExp(pattern, caseSensitive: false).firstMatch(normalized)?.group(0);
final yearMatch = RegExp(r'\b(19\d{2}|20\d{2})\b').firstMatch(normalized);
final yearMatches = RegExp(
r'\b(19\d{2}|20\d{2})\b',
).allMatches(normalized).toList();
// Some release names carry an upload year before the real title and
// release year. Prefer the later year in that specific form.
final yearMatch = yearMatches.length > 1 && yearMatches.first.start == 0
? yearMatches.last
: yearMatches.firstOrNull;
final boundary = RegExp(
r'\b(?:19\d{2}|20\d{2}|S\d{1,2}[ ._-]*E\d{1,3}|\d{1,2}x\d{1,3}|2160p|1080p|720p|480p|4k|web[- ]?(?:dl|rip)?|bluray|bdrip|remux|hdtv|dvd|x26[45]|h\.?26[45]|hevc|av1|aac|ac3|eac3|flac|truehd|dts|ddp|atmos|hdr|dv)\b',
r'\b(?:19\d{2}|20\d{2}|S\d{1,2}[ ._-]*E\d{1,3}|\d{1,2}x\d{1,3}|\d{3,4}x\d{3,4}|2160p|1080p|720p|480p|4k|web[- ]?(?:dl|rip)?|bluray|bdrip|remux|hdtv|dvd|bd|x26[45]|h\.?26[45]|hevc|av1|aac|ac3|eac3|flac|truehd|dts|ddp|atmos|hdr|dv|国语|粤语|国粤(?:双语)?|中(?:英|日|韩)?(?:双语|字幕)|中文字幕|简繁(?:字幕)?)\b',
caseSensitive: false,
);
final boundaryMatch = boundary.firstMatch(normalized);
final boundaryMatches = boundary.allMatches(normalized).toList();
final boundaryMatch =
boundaryMatches.length > 1 &&
boundaryMatches.first.start == 0 &&
RegExp(
r'^(?:19\d{2}|20\d{2})$',
).hasMatch(boundaryMatches.first.group(0) ?? '')
? boundaryMatches[1]
: boundaryMatches.firstOrNull;
var title = boundaryMatch != null
? normalized.substring(0, boundaryMatch.start)
: normalized;
if (yearMatches.length > 1 && yearMatches.first.start == 0) {
title = title.replaceFirst(RegExp(r'^\s*\d{4}[ ._-]+'), '');
}
title = _cleanTitle(title);
final genericName =
title.isEmpty ||
@@ -614,7 +632,7 @@ class ParsedMediaName {
season: season ?? parent?.season,
episode: episode,
isEpisode: (season ?? parent?.season) != null && episode != null,
resolution: first(r'\b(?:2160p|1080p|720p|480p|4k)\b'),
resolution: first(r'\b(?:2160p|1080p|720p|480p|4k|\d{3,4}x\d{3,4})\b'),
source: first(
r'\b(?:WEB[- ]?DL|WEBRip|BluRay|BDRip|REMUX|HDTV|DVD|UHD)\b',
),
@@ -629,6 +647,13 @@ class ParsedMediaName {
static String _cleanTitle(String value) {
var title = value.trim();
// Release collectors often prepend their collection date. It has no
// relation to the movie year and degrades a TMDB query substantially.
title = title.replaceFirst(
RegExp(r'^\s*(?:19|20)\d{2}\s*年\s*\d{1,2}\s*月\s*\d{1,2}\s*[日号]?\s*'),
'',
);
// Release groups frequently put all disc/menu/audio/subtitle information
// in a bracket between the Chinese and original titles. Keep the leading
// title and discard the whole release block instead of joining both sides.
@@ -655,15 +680,23 @@ class ParsedMediaName {
// TMDB receives a compact Chinese query first; the original title remains
// available from TMDB after matching.
final chineseEnglishBoundary = RegExp(
r'[\u4e00-\u9fff\)）]\s+(?=[A-Z][A-Za-z])',
r'[\u4e00-\u9fff\)）]\s*(?=[A-Z][A-Za-z])',
).firstMatch(title);
if (chineseEnglishBoundary != null) {
title = title.substring(0, chineseEnglishBoundary.start + 1);
}
title = title
.replaceFirst(RegExp(r'^\s*[\[【(（]\s*\d{1,3}\s*[\]】)）][ ._-]*'), '')
.replaceFirst(RegExp(r'^\s*\d{1,3}[ ._-]+'), '')
.replaceAll(RegExp(r'[\(（](?:港台|港版?|台版?|国配|国语|简繁?|中字?)[\)）]'), ' ')
.replaceAll(
RegExp(
r'(?:[ ._-]+|^)(?:国[粤英日韩][语字]?|国粤(?:双语)?(?:中字|中英(?:字幕|双字)?)?|国语|粤语|(?:中英|中日|中韩)(?:双语|字幕)?|中文字幕|中字|简繁(?:字幕)?|内封(?:特效)?中英(?:双字|字幕)?|CHINESE|CHN)(?=$|[ ._-])',
caseSensitive: false,
),
' ',
)
.replaceAll(RegExp(r'[\[\]【】{}()（）]'), ' ')
.replaceAll(RegExp(r'\s+'), ' ')
.trim();