x15904151985/iptvnator
1
0 Fork 0
mirror of https://github.com/4gray/iptvnator.git synced 2026-10-14 13:16:16 -08:00
Code Issues Packages Projects Releases Wiki Activity
2,826 Commits 419 Branches 40 Tags
7d1503fd316a30bd11b8297a7f7b4c7ddd5a16e3
Commit Graph
6 Commits
Author SHA1 Message Date
4gray 82a2948ffd fix(matching): keep abbreviation titles out of bare-year keys (#1426) 
A leading 2-5 character uppercase token before a dash, pipe or colon was
always read as a provider tag, so a film whose NAME is such a token lost it and
normalized down to its release year alone. "AKA - 2023" became the key "2023",
where it collided with BDE, BRO, OUT, WIL and IF — and they were offered to
each other as alternative sources in VOD multi-source.

"IT - 65 (2023)" (the Italian copy of the film "65") is structurally identical,
so only the token's meaning can separate them. The leading token is now tested
against a vocabulary — TRAILING_TAG_VOCABULARY plus a prefix-only list derived
from the real catalog — but only when the strip would leave no real word
behind. A compound is read by its head, so the open-ended "4K-<lang>" family
keeps working while "INU-OH" and "PC-4L" are recognized as film names. An
unknown token keeps its title: a refused strip costs one unmatched copy, a
wrong one corrupts that film's identity in VOD multi-source, the TMDB Similar
rail, DB_MATCH_TITLES and pin keys.

"No real word" is decided by running the rest of the pipeline on the stripped
form and looking at what comes out, never by re-implementing what later stages
remove. Quality tags, trailing and underscore tags, double-dash suffixes and
season markers each otherwise smuggle the strip through, and a stage added
later is covered for free.

Validated over the live catalog, movies and series: 83 keys fixed, 0 corrupted
across 1,616,111 titles. Deriving the vocabulary from movies alone missed AMZ,
D+ and P+ and broke the Paramount+/Disney+ copies of the numeric series 1923,
1883, 24 and 9-1-1.
 2026-08-13 17:01:33 +02:00
4grayandClaude Fable 5 36c2867d36 feat(xtream): recognize more language tags in VOD multi-source (#1417) 
* feat(xtream): recognize more language tags in VOD multi-source

The sources popover's language filter and copy chips now read prefixes
with Unicode pipe lookalikes, brackets and spaced dashes, Cyrillic tags
and MULTI. When a stream title carries no tag, the language falls back
to what the stream's visible categories unambiguously state ("EN |
Netflix") — discovery aggregates category names per (playlist, stream)
in SQL, and category prefixes must pass a known-language gate because
everyday category words like new/top/hot are real ISO 639-3 codes.

Both signals stay parsed guesses: browse filter and chips only, never
ranking, failover or dub-warning inputs.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(xtream): address Codex review on multi-source language detection

Gate the new bracket and dash title forms through isKnownLanguageTag:
those positions carry quality/rip tags ([HD], [CAM], NEW -) whose
fabricated "language" would outrank and mask a real category-derived
one. The legacy pipe form stays permissive.

Overlay a late-arriving route category onto the existing route row in
the same-key refresh path — cold/direct routes load categories after
discovery, and the category is outside the movie key on purpose. The
mid-flight case is redelivered by the bind() effect re-running on the
controller's sources signal; that tracked read is now documented as
load-bearing and pinned by a session spec.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(xtream): pair brackets and strip the new tag forms when matching

Greptile: the bracket prefix chose its opening and closing delimiter
independently, so a malformed "[EN)" was read as a language tag.

Codex: recognizing a prefix is only half the job — normalizeTitleKeys
has to strip the same tag, or the tagged copy never matches the bare
one and multi-source cannot offer the film at all. Its leading-tag rule
now shares the pipe-lookalike set and, on the pipe branch only, takes
the same Latin+Cyrillic any-case alphabet with no required trailing
space. Dash and colon keep their uppercase-Latin spaced form: those are
ordinary punctuation, and loosening them would amputate "ОНО: Часть 2"
the way a case-insensitive rule amputates "It: Chapter Two".

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(xtream): keep normalization uppercase-only, measured on real catalogs

The previous commit widened the pipe branch of normalizeTitleKeys to any
case and to Cyrillic, on the theory that nothing but a tag precedes a
pipe. Checked against 1.27M real catalog titles that theory is wrong in
two ways at once: "Akira | 1988" and "Coco | 2017" put the film's name
before the pipe and the year after it, and Russian catalogs write
"Момо | Momo" — localized title, then original. The widening corrupted
349 keys and rescued none, so it is reverted.

What survives is what the data supports: the pipe-lookalike set (0
changed keys, and correct for panels that use them) and dropping the
required space after a pipe (35 changed keys, genuine welded tags like
"EN|Dark Shadows" and "|FR|VO|Le dernier empereur").

A leading-tag guard that refused to strip when no letter remained is
also dropped: it fixes "AKA | 2023" but breaks "IT - 65", so telling
those apart needs a tag vocabulary and belongs in its own change.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs(xtream): cite the measured evidence for the category language gate

The gate's rationale named hypothetical category shapes. On a real
catalog the four it actually turns away are VOD (5,245 movies), KIDS
(1,010), SHOW and WWE — without it the language select offers "VOD" and
"KIDS" as languages. Comments, doc and one spec case only.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs(xtream): restore the docblock currentSourceRow lost to an insertion

routeCategoryLanguage was added between currentSourceRow's docblock and
its signature, so the paragraph describing "the row standing for the
source the route is already playing" ended up introducing a function
that returns a language string. Moved below; no behavior change.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(xtream): stop grouping the scan tier, it can drop a matching source

content is unique per (category, type, stream), so one stream sitting in
several categories is several rows and nothing forces their titles to
agree. The GROUP BY added for group_concat let SQLite keep an arbitrary
row's title, and the normalized confirmation then rejected the whole
stream on a title a sibling row would have matched — the source vanished.

The FTS tier can afford that grouping because its window makes it
necessary; the scan tier takes no window at all, so it now returns a row
per category and their names are merged per stream in TypeScript, which
also keeps the rejected sibling's category in the language derivation.

Found by Codex. Latent rather than active on the catalog I measured (0
streams currently carry differing titles across categories), but the
schema permits it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs(xtream): record why category names stay scoped to matched rows

Codex flagged that the FTS predicate runs before the aggregate, so a
sibling row under a localized title contributes no category. True, and
deliberate: the field is a guess feeding a chip and a browse filter, and
completing it costs measured latency — 0.74s to 2.0s for a correlated
subquery on a 3.9GB catalog, 19.7s for a second bounded lookup — to
correct a cosmetic guess in a shape that occurs 0 times in 2.7M rows.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-08-12 07:59:07 +02:00
4gray 78df3e7dbb fix(portals): match Greek titles whichever sigma the provider typed (#1310) 
Greek Σ has two lowercase forms — medial σ and word-final ς — and neither the
candidate query nor the confirmation treated them as one letter.

The GLOB scan built each character's class from a one-way reach that only
arrived at ς when it started from ς, so a request for "ΑΣ" never admitted a
stored "Ας". Classes are now built from a fold group — every character sharing
an uppercase form — derived by scanning the cased ranges at module load the way
ACCENTED_BY_BASE already is. It generalises past sigma on its own: dotless ı
folds with i, long ſ with s, historic Cyrillic letterforms with В Д О С Т Ъ Ѣ.
Only the 24 groups of 767 that a per-character fold would miss are kept.

Admitting the row was only half of it. normalizeTitleKeys then compared "ασ"
against "ας" and discarded it, because toLowerCase picks the sigma form by
position. Both SQL tiers already folded them together — SQLite's trigram
tokenizer does full Unicode folding natively, unlike LOWER() — so the JS
confirmation was the only tier that did not, making this a pre-existing gap on
the FTS path as well. Normalization now rewrites ς to σ after lowercasing,
which is what Unicode case folding does.

Guards unchanged: a case mapping that changes length (ß → SS, İ) or a GLOB
metacharacter still returns null rather than a partial pattern.
 2026-07-29 23:14:59 +02:00
4grayandClaude Fable 5 0273ded8e2 fix(tmdb): resolve season number from title markers for per-season series slices (#1229) 
* fix(tmdb): resolve season number from title markers for per-season series slices

Providers often slice a show into per-season catalog items ("The
Mandalorian (2 season)", "Пацаны 2 сезон", "The Boys S05") and renumber
the single contained season to 1, so season enrichment fetched the wrong
TMDB season (season 1 metadata for a season 2 item).

- new season-marker.util.ts in shared/interfaces: extractSeasonFromTitle
  (word-first, number-first and S-form markers, bracketed or trailing)
  and resolveEnrichmentSeasonNumber (title marker wins only for
  single-season items whose provider number disagrees)
- wired into Xtream enrichSerialSeasonWithTmdb and the Stalker
  series-view season service (cache/overlay still keyed by provider
  season key)
- SEASON_SUFFIX_PATTERN now also strips number-first season suffixes
  ("2 season", "2 сезон", "2-й сезон" incl. NFD-decomposed ordinals) so
  such titles match the show at all

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): wait for the season map before TMDB season fetch

The TMDB match can arrive before the async season resource; fetching
then passed seasonCount 0, suppressed the title-marker override and
cached the wrong season forever (fetchSeason is idempotent). The effect
now reads the season map tracked and skips while it is empty —
overlay-driven re-runs are safe because fetchSeason early-returns per
(tmdbId, seasonKey).

Addresses Greptile review on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): reset season selection on detail-to-detail navigation

The router reuses the series view for detail-to-detail navigation, and a
retained season selection let the NEW item's tmdb_id pair with the
PREVIOUS series' season context in the TMDB fetch effect, poisoning its
idempotent per-season cache before the new season resource loaded. The
selection is now a linkedSignal keyed on the displayed item's identity —
compared inside the computation, since displayItem produces a fresh
object on every recomputation.

Addresses Greptile review round 2 on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): include the resolved season in the TMDB season cache identity

Per-season slices of one show share (tmdbId, provider season key "1")
but resolve to different TMDB seasons — plain key idempotency served the
first slice's episodes to every later slice. Each cache entry now
records the resolved season it was fetched for: a call resolving a
different season refetches and overwrites (also self-healing a fetch
made with stale navigation context), an in-flight marker dedups
concurrent runs, and a superseded fetch may not store its result.
Failed fetches stay uncached so later triggers retry.

Addresses Codex review (P1) and Greptile review round 3 on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): drop a mismatched season cache entry before its replacement fetch

If a replacement fetch (same key, different resolved season) failed, the
previous slice's entry stayed visible indefinitely through overlay() and
descriptions(). The mismatched entry is now removed up front, so a
failed replacement falls back to provider data until a retry succeeds.

Addresses Codex review (P2) on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): title-based season reset identity and o_name marker support

- The season-selection reset identity now combines provider id and title:
  distinct items can share or lack provider ids, and an id-only identity
  retained the previous item's selection across such navigation
- The season marker is read from whichever title field carries it via
  pickSeasonMarkedTitle: providers put the descriptive title in o_name
  while name stays generic, and the show-level match already used o_name

Addresses Greptile review round 4 (P1) and Codex review (P2) on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(stalker): gate TMDB season fetch on resource coherence, not selection resets

Resetting the parent season selection on item identity change (previous
round) silently disabled enrichment after detail-to-detail navigation
between items sharing one season-key set: the season container keeps its
own selection and deduplicates seasonSelected emissions, so the parent
key stayed null forever. The reset is gone; instead the fetch effect
gates on coherence — it waits while the season resource reloads (the
window in which a reused component pairs the new item's tmdb_id with the
previous item's map) and requires the selected key to exist in the map
with episodes. Stale-snapshot fetches remain self-healing through the
resolution-aware cache.

Addresses Codex review (P2) and Greptile review round 5 on #1229.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-07-24 11:46:48 +02:00
4grayandClaude Fable 5 402382421c feat(matching): strip appended language/quality tags in title normalization (#1211) 
* feat(matching): strip appended language/quality tags in title normalization

Real-world portal catalogs duplicate one show under dozens of tagged
variants ("|ALB| Fallout", "4K-DE - The Pitt (2025) (US)",
"Breaking Bad-eng", "Fallout_esp", "The Last of Us (2023) AF"). A third
of them normalized to polluted keys, silently skipping TMDB enrichment
and staying invisible to cross-portal title matching.

normalizeTitleKeys() now handles, conservatively:
- wrapped pipe tags:      "|ALB| X", "|MULTI| X"
- longer/compound leads:  "EXYU| X", "4K-DE - X", "AR-SUBS - X",
                          "4K-OSN+ - X" (dash/pipe only; colon stays
                          2-3 chars so "NCIS: LA" is untouched)
- underscore suffixes:    "X_eng", "(US)_msub" (single-underscore only,
                          "The_Last_of_Us" stays intact)
- double-dash suffixes:   "X--esp"
- joined dash tags:       "X-DE", "X-eng" (vocabulary-gated and
                          case-uniform only; "Spider-Man", "Kick-It",
                          "Peut-être" are untouched)
- bare trailing tags:     "X (2025) DE", "Breaking Bad ES" (UPPERCASE
                          vocabulary only, skipped for ALL-CAPS titles;
                          "Rocky II", "Made in USA", "Making It" are
                          untouched)

Every leading-tag segment must contain a letter, so numeric titles
("1917 - ...") are never treated as tags. The display-side
stripCountryPrefix() learns the same compound/plus-sign prefixes and the
numeric guard.

buildSearchLookupKey() gets a |v2 suffix so cached negative TMDB match
resolutions keyed on old polluted titles are invalidated.

Measured on 248 real catalog names from four shows (The Pitt, Fallout,
The Last of Us, Breaking Bad): clean matching keys 65% -> 99%, display
strip 91% -> 100%. The corpora are committed as spec fixtures.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(matching): use ES2015-safe trailing trim in title normalization

String.prototype.trimEnd is ES2019; the shared-interfaces lib compiles
against an older lib target (TS2550 in typecheck:web). Replace with a
regex-based trimRight helper. Jest uses its own tsconfig, so this only
surfaced in the CI typecheck, not local unit runs.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(matching): guard tag stripping against real-title false positives

Address code-review findings on the tag-stripping rules:

- underscore suffix is now vocabulary-gated, so "Mr_Robot",
  "Cowboy_Bebop", "Mrs_Davis" keep their second word
- leading single-segment tags before a spaced dash stay 2-3 chars
  (only hyphen-compounds like "4K-DE" and pipe-tags like "EXYU|" may be
  wider), so "DUNE - Part Two" and "ALIEN - Covenant" are left intact
- "IN" is excluded from the weak joined-dash/underscore paths so
  "drive-in" and "Plug-in" are not truncated (India still strips via
  the strong "IN| " / "IN - " forms)

The display-side stripCountryPrefix() mirrors the narrowed dash rule.
Corpus coverage is unchanged at 99% (245/248); new counter-example
tests lock in the guards.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-07-19 19:42:20 +02:00
4grayandClaude Fable 5 dc05a2566e feat(tmdb): opt-in TMDB metadata enrichment for Xtream and Stalker portals (#1123) 
* feat(tmdb): opt-in TMDB metadata enrichment for Xtream and Stalker portals

Adds an opt-in TMDB integration (Settings > Metadata) that enriches
detail views with a field-level merge — the provider stays authoritative
for stream data, TMDB fills editorial fields when the match is confident.

Enrichment:
- Movie/series details: plot, cast (avatar chips), director, genres,
  rating, poster/backdrop, official YouTube trailers
- Confidence-gated matching: provider tmdb_id trusted; otherwise
  normalized-title search with year gate (±1; series accept earlier
  premieres), season-suffix stripping, Cyrillic search-language override,
  and language-prefix fallback variants
- Lazy season/episode enrichment: real episode names, overviews, stills
- "Similar" rail (Xtream): TMDB recommendations matched to the catalog
- Actor pages per portal with full filmography, availability filter and
  an Electron-only "All portals" scope backed by a batched DB_MATCH_TITLES
  worker op over the trigram FTS index

Infrastructure:
- SQLite cache table tmdb_metadata (details, search verdicts, seasons,
  persons; per-language, TTL-guarded), in-memory fallback for the PWA
- Settings: enable toggle, own-API-key override with a live "check key"
  button; TMDB attribution in Settings and About
- Embedded key stays an empty placeholder; CI injects TMDB_API_KEY via
  tools/tmdb/inject-tmdb-key.mjs when the secret is configured
- normalizeTitle shared between renderer and DB worker
- CSP: allow YouTube embeds (frame-src was 'none'; trailers never worked)

Fixes and refactors along the way:
- fix(stalker): Advanced Search sent bare get_ordered_list requests and
  skipped the auth handshake when isFullStalkerPortal was missing on the
  active-playlist meta — full portals answered "Authorization failed."
  and search looked empty; now mirrors the catalog request shape and
  routes through makeAuthenticatedRequest with URL-based detection
- fix(stalker): TMDB fields survive info re-normalization; detail views
  prefer the store copy patched by async enrichment over stale snapshots
- refactor(xtream): split oversized vod/serial detail components into
  component-scoped playback services; detail routes re-initialize on
  route param changes (router reuses them for detail-to-detail nav)
- i18n: all new keys translated across the 18 locales

Docs: docs/architecture/tmdb-metadata-enrichment.md + CLAUDE.md updates.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(tmdb): provide route params observable to inline collection details, linearize regexes

The global-collection inline detail host builds a fake ActivatedRoute for
VodDetailsRouteComponent/SerialDetailsComponent with only snapshot.params.
Since the detail components now read route.params via toSignal() (detail->
detail re-init), the missing observable crashed component construction and
the content hero never rendered — broke dashboard-activation, favorites and
recent Electron E2E on all platforms. Provide the params observable
alongside the snapshot and assert it in the component spec.

Also resolves both CodeQL js/polynomial-redos alerts: bracket-stripping in
normalizeTitle now excludes opening delimiters inside the classes, and
youtubeEmbedUrl extracts watch?v= ids with a linear two-pass match instead
of "watch\?.*v=". Combining-diacritics range rewritten as explicit \u
escapes (greptile note).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(tmdb): surface TMDB-only VOD score in the rating badge, drop youtube.com from CSP

Review follow-ups on PR #1123: the Xtream VOD detail badge renders
rating_imdb, but the merge wrote the TMDB score only into `rating`, so a
TMDB-only score was never displayed (Codex P2) — fill rating_imdb when the
provider left it empty, mirroring the Stalker merge. All trailer iframes
are normalized to youtube-nocookie.com, so the extra youtube.com frame-src
allowance was dead surface (greptile) — removed.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(tmdb): resolve confirmed review findings — matching correctness, race guards, cache schema

Fixes the confirmed findings from the PR #1123 code review:

- Stalker search: setSelectedContentType now runs BEFORE setSelectedItem,
  so the TMDB enrichment gate in the selection hook no longer sees the
  content type of the previously open tab (wrong/no enrichment after
  ITV -> search -> movie).
- Title normalization is now two-tier (normalizeTitleKeys): the exact
  normalized form keeps a trailing year, the base form strips it and
  remembers the tag. Year stripping is anchored to the end of the title
  ("2001: A Space Odyssey" keeps its year) and language-prefix stripping
  is UPPERCASE-only ("It: Chapter Two" is no longer amputated).
- All catalog matching (similar rail, actor pages, DB worker
  DB_MATCH_TITLES) compares exact forms first and only accepts
  year-stripped matches when the stripped tag is year-compatible (+-1)
  with the TMDB year — "Blade Runner" (1982) can no longer claim a
  catalog "Blade Runner 2049". CatalogTitleMatch carries the stripped
  trailingYear so the renderer can apply the guard to worker matches.
- mergedBackdrops tolerates a plain-string backdrop_path; enrichment
  merge+patch blocks are wrapped in try/catch so a malformed provider
  payload can no longer become an unhandled rejection.
- loadGlobalMatches (both actor routes) guards against actor->actor
  navigation races — a slow match for the previous person no longer
  overwrites the current one's results.
- tmdb_metadata media_type CHECK widened to ('movie','tv','person') and
  person rows now use the honest 'person' type (TmdbCacheMediaType).
  Pre-release dev DBs with the narrow CHECK are rebuilt in place — the
  table is a pure cache, so the migration is a self-healing
  drop-and-recreate keyed off sqlite_master.

Docs updated (tmdb-metadata-enrichment.md, CLAUDE.md). New regression
coverage: title-normalization.util.spec.ts, two-tier cases in
tmdb-similar.util.spec.ts and title-match.operations.spec.ts.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-07-04 17:07:46 +02:00