feat(live): add language-aware flexible channel search (#23)

Add backend language detection, flexible live channel search, iptv-org catalog caching, and protected built-in EPG source registration.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
blomios
2026-10-06 11:42:06 +02:00
parent a6759b116d
commit 2a1c745a26
11 changed files with 314 additions and 19 deletions

View File

@ -0,0 +1,20 @@
const BUILTIN_EPG_SOURCES = Object.freeze([
Object.freeze({ builtinKey: 'epgpw-fr', type: 'epg', name: 'EPG.PW France', url: 'https://epg.pw/xmltv/epg_FR.xml.gz', provider: 'EPG.PW', language: 'fr', country: 'FR', priority: 100 }),
Object.freeze({ builtinKey: 'epgpw-us', type: 'epg', name: 'EPG.PW United States', url: 'https://epg.pw/xmltv/epg_US.xml.gz', provider: 'EPG.PW', language: 'en', country: 'US', priority: 90 })
]);
function protectedFields(source) {
return source?.builtinKey ? { readonly: true, removable: false } : { readonly: false, removable: true };
}
function ensureBuiltinSources(database) {
database.sources ||= [];
for (const definition of BUILTIN_EPG_SOURCES) {
const existing = database.sources.find(source => source.builtinKey === definition.builtinKey);
if (existing) Object.assign(existing, definition, protectedFields(definition));
else database.sources.push({ id: database.nextId++, ...definition, ...protectedFields(definition), enabled: false, created_at: new Date().toISOString(), updated_at: new Date().toISOString() });
}
return database;
}
module.exports = { BUILTIN_EPG_SOURCES, ensureBuiltinSources, protectedFields };

View File

@ -0,0 +1,50 @@
const LANGUAGES = Object.freeze({
fr: { code: 'fr', short: 'FR', label: 'French' },
en: { code: 'en', short: 'EN', label: 'English' },
es: { code: 'es', short: 'ES', label: 'Spanish' },
de: { code: 'de', short: 'DE', label: 'German' },
it: { code: 'it', short: 'IT', label: 'Italian' },
pt: { code: 'pt', short: 'PT', label: 'Portuguese' }
});
const CODE_ALIASES = Object.freeze({
fr: 'fr', fra: 'fr', fre: 'fr', french: 'fr', francais: 'fr', france: 'fr',
en: 'en', eng: 'en', english: 'en', anglais: 'en', uk: 'en', us: 'en',
es: 'es', spa: 'es', spanish: 'es', espanol: 'es', spain: 'es',
de: 'de', deu: 'de', ger: 'de', german: 'de', deutsch: 'de',
it: 'it', ita: 'it', italian: 'it', italiano: 'it',
pt: 'pt', por: 'pt', portuguese: 'pt', portugues: 'pt'
});
function normalize(value) {
return String(value || '').normalize('NFKD').replace(/[\u0300-\u036f]/g, '').toLowerCase();
}
function detectLanguage(item) {
const data = item.data && typeof item.data === 'object' ? item.data : {};
const values = [item.name, item.category_name, data.groupTitle, data.tvgId, data.tvgName].filter(Boolean);
const scores = {};
const evidence = [];
const prefix = String(item.name || '').trim().match(/^(?:\|\s*([a-z-]{2,8})\s*\||\[\s*([a-z-]{2,8})\s*\]|([a-z-]{2,8})\s*\|)/i);
if (prefix) add(prefix[1] || prefix[2] || prefix[3], 70, 'name-prefix');
values.forEach((value, index) => {
const tokens = normalize(value).split(/[^a-z]+/).filter(Boolean);
for (const token of tokens) add(token, index === 0 ? 45 : 25, index === 0 ? 'name-token' : 'metadata');
});
function add(token, weight, type) {
const code = CODE_ALIASES[normalize(token)];
if (!code) return;
scores[code] = (scores[code] || 0) + weight;
evidence.push({ code, weight, type, value: String(token) });
}
const ranked = Object.entries(scores).sort((a, b) => b[1] - a[1]);
if (!ranked.length) return { code: 'und', short: 'UND', label: 'Unknown', confidence: 0, scores, evidence };
const total = ranked.reduce((sum, [, score]) => sum + score, 0);
const confidence = Math.round(ranked[0][1] / total * 100);
if (confidence < 55 || (ranked[1] && ranked[0][1] - ranked[1][1] < 15)) {
return { code: 'und', short: 'UND', label: 'Unknown', confidence: 0, scores, evidence };
}
return { ...LANGUAGES[ranked[0][0]], confidence, scores, evidence };
}
module.exports = { CODE_ALIASES, LANGUAGES, detectLanguage, normalize };

View File

@ -0,0 +1,19 @@
function normalizeSearch(value) {
return String(value || '')
.normalize('NFKD').replace(/[\u0300-\u036f]/g, '')
.toLowerCase().replace(/[^a-z0-9]+/g, ' ').trim()
.split(/\s+/).filter(Boolean)
.map(token => token.length > 3 && token.endsWith('s') ? token.slice(0, -1) : token);
}
function matchesSearch(channel, query) {
const wanted = normalizeSearch(query);
if (!wanted.length) return true;
const corpus = normalizeSearch([
channel.displayName, channel.standardName, channel.fingerprint,
...(channel.variants || []).flatMap(variant => [variant.name, variant.tvgId, variant.categoryName])
].filter(Boolean).join(' '));
return wanted.every(token => corpus.some(candidate => candidate.includes(token)));
}
module.exports = { matchesSearch, normalizeSearch };

View File

@ -0,0 +1,65 @@
const API_ROOT = 'https://iptv-org.github.io/api';
function normalize(value) {
return String(value || '').normalize('NFKD').replace(/[\u0300-\u036f]/g, '')
.toLowerCase().replace(/[^a-z0-9]+/g, ' ').trim();
}
function parseJson(value, fallback = []) {
try { return JSON.parse(value); } catch { return fallback; }
}
async function refresh(database, fetchImpl = globalThis.fetch) {
if (typeof fetchImpl !== 'function') throw new Error('A fetch implementation is required');
const [channelsResponse, feedsResponse] = await Promise.all([
fetchImpl(`${API_ROOT}/channels.json`), fetchImpl(`${API_ROOT}/feeds.json`)
]);
if (!channelsResponse.ok || !feedsResponse.ok) throw new Error('iptv-org API refresh failed');
const channels = await channelsResponse.json();
const feeds = await feedsResponse.json();
const feedLanguages = new Map();
for (const feed of feeds) {
const list = feedLanguages.get(feed.channel) || new Set();
for (const language of feed.languages || []) list.add(String(language).toLowerCase());
feedLanguages.set(feed.channel, list);
}
const write = database.transaction(() => {
database.prepare('DELETE FROM iptv_org_channels').run();
const insert = database.prepare(`
INSERT INTO iptv_org_channels
(id, name, normalized_name, alt_names, country, languages, data, updated_at)
VALUES (?, ?, ?, ?, ?, ?, ?, ?)
`);
const now = Date.now();
for (const channel of channels) {
insert.run(channel.id, channel.name, normalize(channel.name), JSON.stringify(channel.alt_names || []),
channel.country || null, JSON.stringify([...feedLanguages.get(channel.id) || []]), JSON.stringify(channel), now);
}
database.prepare(`
INSERT INTO iptv_org_cache_state (resource, refreshed_at, item_count) VALUES ('catalog', ?, ?)
ON CONFLICT(resource) DO UPDATE SET refreshed_at = excluded.refreshed_at, item_count = excluded.item_count
`).run(now, channels.length);
});
write();
return { refreshedAt: Date.now(), itemCount: channels.length };
}
function findMatch(database, channel) {
const names = new Set([channel.displayName, channel.fingerprint, ...(channel.variants || []).map(variant => variant.name)].map(normalize).filter(Boolean));
if (!names.size) return null;
const direct = database.prepare(`SELECT * FROM iptv_org_channels WHERE normalized_name IN (${[...names].map(() => '?').join(',')})`).all(...names);
const candidates = direct.length ? direct : database.prepare('SELECT * FROM iptv_org_channels').all().filter(row =>
parseJson(row.alt_names).some(name => names.has(normalize(name)))
);
const language = channel.language?.code;
const compatible = candidates.find(row => language === 'und' || parseJson(row.languages).includes(language)) || candidates[0];
if (!compatible) return null;
return {
channelId: compatible.id,
name: compatible.name,
country: compatible.country,
confidence: names.has(compatible.normalized_name) ? 100 : 90
};
}
module.exports = { API_ROOT, findMatch, normalize, refresh };

View File

@ -1,5 +1,8 @@
const crypto = require('crypto');
const { pairKey, scoreCandidate, identity } = require('./liveGroupingScoring');
const { detectLanguage } = require('./channelLanguage');
const { matchesSearch } = require('./flexibleChannelSearch');
const iptvOrgCatalog = require('./iptvOrgCatalogService');
const QUALITY_TOKENS = new Set([
'4k', 'uhd', 'fhd', 'fullhd', 'hd', 'sd', 'hevc', 'h265', 'h264',
@ -76,6 +79,11 @@ function projectCatalog(items, options = {}) {
qualityLabel: quality.label,
qualityRank: quality.rank
};
const language = detectLanguage({ ...projected, data: parseData(projected.data) });
projected.languageCode = language.code;
projected.languageConfidence = language.confidence;
projected.languageScores = language.scores;
projected.languageEvidence = language.evidence;
const itemIdentity = identity(fingerprint);
const candidateKey = itemIdentity.family && itemIdentity.number
? `${itemIdentity.family}:${itemIdentity.number}`
@ -241,7 +249,9 @@ function rebuild(database = null, options = {}) {
database.prepare('DELETE FROM logical_channels').run();
database.prepare(`
UPDATE playlist_items SET logical_channel_id = NULL, grouping_fingerprint = NULL,
grouping_input_hash = NULL, quality_label = NULL, quality_rank = NULL, variant_rank = NULL
grouping_input_hash = NULL, quality_label = NULL, quality_rank = NULL, variant_rank = NULL,
language_code = NULL, language_confidence = NULL, language_scores = NULL,
language_evidence = NULL, language_detector = NULL, language_input_hash = NULL
WHERE type = 'live'
`).run();
const insertLogical = database.prepare(`
@ -251,7 +261,9 @@ function rebuild(database = null, options = {}) {
`);
const updateVariant = database.prepare(`
UPDATE playlist_items SET logical_channel_id = ?, grouping_fingerprint = ?,
grouping_input_hash = ?, quality_label = ?, quality_rank = ?, variant_rank = ?
grouping_input_hash = ?, quality_label = ?, quality_rank = ?, variant_rank = ?,
language_code = ?, language_confidence = ?, language_scores = ?,
language_evidence = ?, language_detector = 'algorithm', language_input_hash = ?
WHERE id = ? AND type = 'live'
`);
const now = Date.now();
@ -259,7 +271,10 @@ function rebuild(database = null, options = {}) {
insertLogical.run(group.id, group.fingerprint, group.displayName, group.preferredVariantId, group.variants.length, now, now);
for (const variant of group.variants) {
updateVariant.run(group.id, group.fingerprint, variant.groupingInputHash,
variant.qualityLabel, variant.qualityRank, variant.variantRank, variant.id);
variant.qualityLabel, variant.qualityRank, variant.variantRank,
variant.languageCode, variant.languageConfidence,
JSON.stringify(variant.languageScores), JSON.stringify(variant.languageEvidence),
stableHash(`${variant.name || ''}\0${variant.category_name || ''}\0${variant.tvg_id || ''}`), variant.id);
}
}
database.prepare(`
@ -330,7 +345,7 @@ function parseData(value) {
try { return value ? JSON.parse(value) : {}; } catch { return {}; }
}
function getChannels({ sourceId = null, includeHidden = false, userId = null }, database = null) {
function getChannels({ sourceId = null, includeHidden = false, userId = null, q = '', languages = [] }, database = null) {
database ||= defaultDatabase();
const params = [];
let sourceFilter = '';
@ -350,7 +365,16 @@ function getChannels({ sourceId = null, includeHidden = false, userId = null },
WHERE p.type = 'live' ${sourceFilter}
ORDER BY l.display_name, p.variant_rank
`).all(userId ?? -1, ...params);
return serializeChannels(rows, includeHidden);
return serializeChannels(rows, includeHidden).map(channel => {
const match = iptvOrgCatalog.findMatch(database, channel);
if (match) {
channel.standardName = match.name;
channel.standard = { provider: 'iptv-org', channelId: match.channelId, country: match.country, confidence: match.confidence };
}
return channel;
}).filter(channel =>
(!languages.length || languages.includes(channel.language.code)) && matchesSearch(channel, q)
);
}
function serializeChannels(rows, includeHidden = false) {
@ -376,22 +400,33 @@ function serializeChannels(rows, includeHidden = false) {
isHidden: hidden,
isFavorite: Boolean(row.user_favorite)
};
if (!groups.has(row.logical_channel_id)) {
groups.set(row.logical_channel_id, {
id: row.logical_channel_id,
const languageCode = row.language_code || 'und';
const languageKey = `${row.logical_channel_id}:${languageCode}`;
if (!groups.has(languageKey)) {
groups.set(languageKey, {
id: languageKey,
logicalChannelId: row.logical_channel_id,
fingerprint: row.grouping_fingerprint,
displayName: row.display_name,
standardName: row.display_name,
standard: { provider: 'local', channelId: null, confidence: 0 },
language: {
code: languageCode,
short: languageCode === 'und' ? 'UND' : languageCode.toUpperCase(),
label: languageCode === 'und' ? 'Unknown' : new Intl.DisplayNames(['en'], { type: 'language' }).of(languageCode),
confidence: Number(row.language_confidence || 0)
},
name: row.display_name,
variants: [],
isFavorite: false,
isHidden: true
});
}
const group = groups.get(row.logical_channel_id);
const group = groups.get(languageKey);
group.variants.push(variant);
group.isFavorite ||= variant.isFavorite;
group.isHidden &&= variant.isHidden;
if (row.id === row.preferred_variant_id) group.preferredVariant = variant;
if (!group.preferredVariant || Number(variant.variantRank) < Number(group.preferredVariant.variantRank)) group.preferredVariant = variant;
}
return Array.from(groups.values()).map(group => {
group.preferredVariant ||= group.variants[0];
@ -401,6 +436,13 @@ function serializeChannels(rows, includeHidden = false) {
group.selectedVariantId = group.preferredVariant
? `${group.preferredVariant.sourceId}:${group.preferredVariant.id}`
: null;
const epgChannelId = group.preferredVariant?.tvgId || group.preferredVariant?.epg_channel_id || null;
group.epg = {
sourceId: null,
channelId: epgChannelId,
language: group.language.code,
matchConfidence: epgChannelId ? 100 : 0
};
return group;
});
}
@ -414,6 +456,7 @@ module.exports = {
identityCandidatePairs,
normalizeChannelName,
markProjectionBuilding,
matchesSearch,
projectCatalog,
projectionState,
qualityForName,