@torrent-tv/proxy 2.86.0 → 2.87.1

package/services/{tracks → media/tracks}/AudioTrack.js

@@ -1,131 +1,131 @@
-/**
- * @file Audio track — extends ContainerTrack with audio-specific flags.
- *
- * Matroska RFC 9559 §5.1.4.1:
- *  - FlagOriginal 0x55AE (default 0) — track is original language.
- *  - FlagCommentary 0x55AF — commentary track.
- *  - FlagVisualImpaired 0x55AC — audio description for visually impaired.
- *  - FlagTextDescriptions 0x55AD, FlagHearingImpaired etc. may also appear
- *    but the three above drive the audio menu (research/container-spec-conformance).
- * MP4 ISO/IEC 14496-12: tkhd alternate_group groups alternate audio tracks.
- */
-
-import { ContainerTrack } from "./ContainerTrack.js";
-
-/**
- * Codec identifiers as containers write them, against the name ffmpeg prints.
- *
- * Needed because the browser decides whether it can play a soundtrack from that
- * name, and for a sidecar file there is no ffmpeg banner to read it from — the
- * track came from the container's own table, where Matroska writes `A_AC3` and
- * MP4 writes `ac-3` for the thing ffmpeg calls `ac3`. Only what a soundtrack can
- * actually be is listed; an identifier not here is reported as it was written,
- * which the browser treats as one it does not know and therefore transcodes.
- */
-const CODEC_NAMES = new Map([
-  ["A_AAC", "aac"],
-  ["A_AC3", "ac3"],
-  ["A_EAC3", "eac3"],
-  ["A_DTS", "dts"],
-  ["A_FLAC", "flac"],
-  ["A_OPUS", "opus"],
-  ["A_VORBIS", "vorbis"],
-  ["A_TRUEHD", "truehd"],
-  ["A_MPEG/L3", "mp3"],
-  ["A_MPEG/L2", "mp2"],
-  ["A_ALAC", "alac"],
-  ["mp4a", "aac"],
-  ["ac-3", "ac3"],
-  ["ec-3", "eac3"],
-  ["alac", "alac"],
-  ["opus", "opus"],
-  ["Opus", "opus"],
-  ["fLaC", "flac"],
-  ["flac", "flac"]
-]);
-
-/**
- * Extensions of raw elementary streams, against the codec they carry.
- *
- * A bare `.ac3` has no track table to read, and its extension is the only thing
- * that states its codec — which for an elementary stream is exactly what the
- * extension means.
- */
-const CODEC_BY_EXTENSION = new Map([
-  [".aac", "aac"],
-  [".ac3", "ac3"],
-  [".eac3", "eac3"],
-  [".dts", "dts"],
-  [".dtshd", "dts"],
-  [".flac", "flac"],
-  [".mp3", "mp3"],
-  [".mp2", "mp2"],
-  [".opus", "opus"],
-  [".ogg", "vorbis"],
-  [".oga", "vorbis"],
-  [".wav", "pcm"],
-  [".thd", "truehd"],
-  [".mlp", "truehd"],
-  [".m4a", "aac"]
-]);
-
-export class AudioTrack extends ContainerTrack {
-  /**
-   * @param {object} params - See ContainerTrack plus:
-   * @param {boolean} params.isOriginal - FlagOriginal
-   * @param {boolean} params.isCommentary - FlagCommentary
-   * @param {boolean} params.isVisualImpaired - FlagVisualImpaired / descriptive audio
-   * @param {number | null} params.channels
-   * @param {number | null} params.samplingFrequency
-   */
-  constructor(params) {
-    super({ ...params, type: "audio" });
-    this.isOriginal = params.isOriginal === true;
-    this.isCommentary = params.isCommentary === true;
-    this.isVisualImpaired = params.isVisualImpaired === true;
-    this.channels = Number.isFinite(params.channels) ? params.channels : null;
-    this.samplingFrequency = Number.isFinite(params.samplingFrequency) ? params.samplingFrequency : null;
-  }
-
-  /**
-   * The ffmpeg-side codec name for one track, from whatever the reading gave.
-   *
-   * @param {{ codec?: string, codecId?: string }} track
-   * @param {string} [extension] - The sidecar file's extension, when the track
-   *   came from a file with no readable table.
-   * @returns {string}
-   */
-  static codecNameOf(track, extension = "") {
-    const fromBanner = typeof track?.codec === "string" ? track.codec.trim() : "";
-    if (fromBanner.length > 0) {
-      return fromBanner.toLowerCase();
-    }
-    const codecId = typeof track?.codecId === "string" ? track.codecId.trim() : "";
-    if (codecId.length > 0) {
-      const known = CODEC_NAMES.get(codecId);
-      if (known) {
-        return known;
-      }
-      // Matroska allows a suffix — `A_AAC/MPEG4/LC`, `A_PCM/INT/LIT` — so the
-      // family is what the first two segments say.
-      const family = codecId.split("/").slice(0, 2).join("/");
-      const byFamily = CODEC_NAMES.get(family) ?? CODEC_NAMES.get(codecId.split("/")[0]);
-      if (byFamily) {
-        return byFamily;
-      }
-      if (codecId.startsWith("A_PCM")) {
-        return "pcm";
-      }
-      return codecId.toLowerCase();
-    }
-    return CODEC_BY_EXTENSION.get(extension) ?? "";
-  }
-
-  /** Human label helper — commentary and descriptive audio must not look like main track. */
-  audioRoleLabel() {
-    if (this.isCommentary) return "commentary";
-    if (this.isVisualImpaired) return "descriptive";
-    if (this.isOriginal) return "original";
-    return "";
-  }
-}
+/**
+ * @file Audio track — extends ContainerTrack with audio-specific flags.
+ *
+ * Matroska RFC 9559 §5.1.4.1:
+ *  - FlagOriginal 0x55AE (default 0) — track is original language.
+ *  - FlagCommentary 0x55AF — commentary track.
+ *  - FlagVisualImpaired 0x55AC — audio description for visually impaired.
+ *  - FlagTextDescriptions 0x55AD, FlagHearingImpaired etc. may also appear
+ *    but the three above drive the audio menu (research/container-spec-conformance).
+ * MP4 ISO/IEC 14496-12: tkhd alternate_group groups alternate audio tracks.
+ */
+
+import { ContainerTrack } from "./ContainerTrack.js";
+
+/**
+ * Codec identifiers as containers write them, against the name ffmpeg prints.
+ *
+ * Needed because the browser decides whether it can play a soundtrack from that
+ * name, and for a sidecar file there is no ffmpeg banner to read it from — the
+ * track came from the container's own table, where Matroska writes `A_AC3` and
+ * MP4 writes `ac-3` for the thing ffmpeg calls `ac3`. Only what a soundtrack can
+ * actually be is listed; an identifier not here is reported as it was written,
+ * which the browser treats as one it does not know and therefore transcodes.
+ */
+const CODEC_NAMES = new Map([
+  ["A_AAC", "aac"],
+  ["A_AC3", "ac3"],
+  ["A_EAC3", "eac3"],
+  ["A_DTS", "dts"],
+  ["A_FLAC", "flac"],
+  ["A_OPUS", "opus"],
+  ["A_VORBIS", "vorbis"],
+  ["A_TRUEHD", "truehd"],
+  ["A_MPEG/L3", "mp3"],
+  ["A_MPEG/L2", "mp2"],
+  ["A_ALAC", "alac"],
+  ["mp4a", "aac"],
+  ["ac-3", "ac3"],
+  ["ec-3", "eac3"],
+  ["alac", "alac"],
+  ["opus", "opus"],
+  ["Opus", "opus"],
+  ["fLaC", "flac"],
+  ["flac", "flac"]
+]);
+
+/**
+ * Extensions of raw elementary streams, against the codec they carry.
+ *
+ * A bare `.ac3` has no track table to read, and its extension is the only thing
+ * that states its codec — which for an elementary stream is exactly what the
+ * extension means.
+ */
+const CODEC_BY_EXTENSION = new Map([
+  [".aac", "aac"],
+  [".ac3", "ac3"],
+  [".eac3", "eac3"],
+  [".dts", "dts"],
+  [".dtshd", "dts"],
+  [".flac", "flac"],
+  [".mp3", "mp3"],
+  [".mp2", "mp2"],
+  [".opus", "opus"],
+  [".ogg", "vorbis"],
+  [".oga", "vorbis"],
+  [".wav", "pcm"],
+  [".thd", "truehd"],
+  [".mlp", "truehd"],
+  [".m4a", "aac"]
+]);
+
+export class AudioTrack extends ContainerTrack {
+  /**
+   * @param {object} params - See ContainerTrack plus:
+   * @param {boolean} params.isOriginal - FlagOriginal
+   * @param {boolean} params.isCommentary - FlagCommentary
+   * @param {boolean} params.isVisualImpaired - FlagVisualImpaired / descriptive audio
+   * @param {number | null} params.channels
+   * @param {number | null} params.samplingFrequency
+   */
+  constructor(params) {
+    super({ ...params, type: "audio" });
+    this.isOriginal = params.isOriginal === true;
+    this.isCommentary = params.isCommentary === true;
+    this.isVisualImpaired = params.isVisualImpaired === true;
+    this.channels = Number.isFinite(params.channels) ? params.channels : null;
+    this.samplingFrequency = Number.isFinite(params.samplingFrequency) ? params.samplingFrequency : null;
+  }
+
+  /**
+   * The ffmpeg-side codec name for one track, from whatever the reading gave.
+   *
+   * @param {{ codec?: string, codecId?: string }} track
+   * @param {string} [extension] - The sidecar file's extension, when the track
+   *   came from a file with no readable table.
+   * @returns {string}
+   */
+  static codecNameOf(track, extension = "") {
+    const fromBanner = typeof track?.codec === "string" ? track.codec.trim() : "";
+    if (fromBanner.length > 0) {
+      return fromBanner.toLowerCase();
+    }
+    const codecId = typeof track?.codecId === "string" ? track.codecId.trim() : "";
+    if (codecId.length > 0) {
+      const known = CODEC_NAMES.get(codecId);
+      if (known) {
+        return known;
+      }
+      // Matroska allows a suffix — `A_AAC/MPEG4/LC`, `A_PCM/INT/LIT` — so the
+      // family is what the first two segments say.
+      const family = codecId.split("/").slice(0, 2).join("/");
+      const byFamily = CODEC_NAMES.get(family) ?? CODEC_NAMES.get(codecId.split("/")[0]);
+      if (byFamily) {
+        return byFamily;
+      }
+      if (codecId.startsWith("A_PCM")) {
+        return "pcm";
+      }
+      return codecId.toLowerCase();
+    }
+    return CODEC_BY_EXTENSION.get(extension) ?? "";
+  }
+
+  /** Human label helper — commentary and descriptive audio must not look like main track. */
+  audioRoleLabel() {
+    if (this.isCommentary) return "commentary";
+    if (this.isVisualImpaired) return "descriptive";
+    if (this.isOriginal) return "original";
+    return "";
+  }
+}

package/services/{tracks → media/tracks}/language-detect.js

@@ -1,174 +1,174 @@
-/**
- * @file Content-based subtitle language detection (proxy side).
- *
- * Uses `franc` (n-gram / trigram frequency against per-language reference
- * profiles — MIT). Runs on the proxy where the full subtitle text and
- * node_modules live, so no detection model ships to the browser. Detection is
- * restricted to a curated set of plausible subtitle languages via franc's
- * `only` option: this both maps ISO 639-3 → ISO 639-1 + English name and
- * avoids exotic false positives on short text (e.g. English mis-detected as
- * Scots). Returns null when franc is not confident (too little text, or
- * undetermined).
- */
-
-import { franc } from "franc";
-
-/** ISO 639-3 (franc output) → { code: ISO 639-1 / BCP-47, name }. Curated allowlist. */
-const LANG_3_TO_1 = {
-  eng: { code: "en", name: "English" },
-  rus: { code: "ru", name: "Russian" },
-  ukr: { code: "uk", name: "Ukrainian" },
-  bel: { code: "be", name: "Belarusian" },
-  jpn: { code: "ja", name: "Japanese" },
-  kor: { code: "ko", name: "Korean" },
-  cmn: { code: "zh", name: "Chinese" },
-  spa: { code: "es", name: "Spanish" },
-  fra: { code: "fr", name: "French" },
-  deu: { code: "de", name: "German" },
-  ita: { code: "it", name: "Italian" },
-  por: { code: "pt", name: "Portuguese" },
-  pol: { code: "pl", name: "Polish" },
-  nld: { code: "nl", name: "Dutch" },
-  arb: { code: "ar", name: "Arabic" },
-  tur: { code: "tr", name: "Turkish" },
-  vie: { code: "vi", name: "Vietnamese" },
-  tha: { code: "th", name: "Thai" },
-  hin: { code: "hi", name: "Hindi" },
-  ind: { code: "id", name: "Indonesian" },
-  zlm: { code: "ms", name: "Malay" },
-  ces: { code: "cs", name: "Czech" },
-  slk: { code: "sk", name: "Slovak" },
-  ron: { code: "ro", name: "Romanian" },
-  hun: { code: "hu", name: "Hungarian" },
-  srp: { code: "sr", name: "Serbian" },
-  hrv: { code: "hr", name: "Croatian" },
-  bul: { code: "bg", name: "Bulgarian" },
-  ell: { code: "el", name: "Greek" },
-  heb: { code: "he", name: "Hebrew" },
-  dan: { code: "da", name: "Danish" },
-  fin: { code: "fi", name: "Finnish" },
-  nob: { code: "no", name: "Norwegian" },
-  swe: { code: "sv", name: "Swedish" },
-  fas: { code: "fa", name: "Persian" }
-};
-
-const ONLY = Object.keys(LANG_3_TO_1);
-
-/**
- * The least text, in characters, that supports an answer of each language.
- *
- * MEASURED, not chosen — `research/franc-boundary-2026-09-02.md`. Method:
- * Wikipedia extracts per language (deliberately NOT the UDHR, which is what
- * franc's own profiles are built from and would read optimistically), 120
- * windows cut at random from them at each length of a ladder from 40 to 1300
- * characters, and the figure recorded is the shortest length from which franc
- * answered correctly in at least 95 % of trials AND kept doing so at every
- * longer length measured.
- *
- * Why it is per language rather than one number: the answer is not equally hard
- * to reach, and the spread is fivefold. Greek and Korean settle at 40
- * characters because their script settles it; English needs 130; Russian and
- * Czech need 650, because each competes with neighbours in this very list for
- * the same trigrams — Russian with Bulgarian, Serbian and Ukrainian, Czech with
- * Slovak.
- *
- * Two languages are deliberately ABSENT. Swedish and Chinese did not settle
- * anywhere in the ladder on the corpora collected, so no figure for them is
- * measured and none is invented; they take the fallback below.
- *
- * A language with no entry gets the WORST measured figure. That is the
- * conservative reading and it is still a measurement rather than a guess: it
- * says "no better than the hardest language we have measured".
- *
- * @type {Record<string, number>}
- */
-const LEAST_TEXT = {
-  bel: 100,
-  bul: 80,
-  ces: 650,
-  deu: 200,
-  ell: 40,
-  eng: 130,
-  fra: 80,
-  heb: 60,
-  ita: 160,
-  kor: 40,
-  nld: 130,
-  pol: 200,
-  por: 200,
-  rus: 650,
-  spa: 100,
-  srp: 100,
-  tur: 160,
-  ukr: 250
-};
-
-/** The worst measured figure, used for any language not in the table. */
-const LEAST_TEXT_WORST = Math.max(...Object.values(LEAST_TEXT), 0);
-
-/** franc's own floor: below this it is not asked at all. */
-const FRANC_FLOOR = 15;
-
-/** One space between words, nothing else — the form every figure above is in. */
-function normalise(text) {
-  return typeof text === "string" ? text.replace(/\s+/g, " ").trim() : "";
-}
-
-/**
- * franc's answer for this text, or null when it has none.
- *
- * @param {string} text
- * @returns {string | null} ISO 639-3.
- */
-function ask(text) {
-  if (text.length < FRANC_FLOOR) {
-    return null;
-  }
-  const iso3 = franc(text, { only: ONLY, minLength: FRANC_FLOOR });
-  return iso3 === "und" ? null : iso3;
-}
-
-/**
- * Best-effort detect the language of subtitle text.
- *
- * **Give this the words a viewer reads and nothing else.** franc scores letter
- * trigrams over the whole string it is handed, so anything around the words
- * competes with them. An ASS file is markup by half: measured 2026-09-01 on
- * `[HorribleSubs] Drifters - 03 [1080p].ass`, 5040 Latin characters of Aegisub
- * headers, style and font names, `Format:`/`Dialogue:` field prefixes and
- * `{\…}` override groups against 5983 Cyrillic characters of dialogue —
- * `franc(the file) = eng`, `franc(the dialogue) = rus`. The proxy had already
- * built the markup-free text and detected on the file anyway, so a Russian
- * track was offered to the viewer as English.
- *
- * `detectLanguageFromVtt` below is the safe entry point for a whole document;
- * this one is for text that is already only text.
- *
- * @param {string} text - Subtitle text with no markup left in it.
- * @returns {{ code: string, name: string } | null} Detected language, or null when uncertain.
- */
-export function detectLanguage(text) {
-  const words = normalise(text);
-  if (words.length < FRANC_FLOOR) {
-    return null;
-  }
-  const candidate = ask(words);
-  if (candidate === null) {
-    return null;
-  }
-  // Enough text to support THIS answer. The figure is the language's own,
-  // because the languages are not alike: Russian shares its trigrams with
-  // Bulgarian, Serbian and Ukrainian and needs several times what English does.
-  if (words.length < (LEAST_TEXT[candidate] ?? LEAST_TEXT_WORST)) {
-    return null;
-  }
-  // And an answer that does not survive losing half the text was an accident of
-  // where the text happened to stop, not a reading of it. Free — franc costs
-  // about 2 ms whatever the size, measured — and it needs no figure of its own,
-  // because the test is taken on the text in hand.
-  const middle = Math.floor(words.length / 2);
-  if (ask(words.slice(0, middle)) !== candidate || ask(words.slice(middle)) !== candidate) {
-    return null;
-  }
-  return LANG_3_TO_1[candidate] ?? null;
-}
+/**
+ * @file Content-based subtitle language detection (proxy side).
+ *
+ * Uses `franc` (n-gram / trigram frequency against per-language reference
+ * profiles — MIT). Runs on the proxy where the full subtitle text and
+ * node_modules live, so no detection model ships to the browser. Detection is
+ * restricted to a curated set of plausible subtitle languages via franc's
+ * `only` option: this both maps ISO 639-3 → ISO 639-1 + English name and
+ * avoids exotic false positives on short text (e.g. English mis-detected as
+ * Scots). Returns null when franc is not confident (too little text, or
+ * undetermined).
+ */
+
+import { franc } from "franc";
+
+/** ISO 639-3 (franc output) → { code: ISO 639-1 / BCP-47, name }. Curated allowlist. */
+const LANG_3_TO_1 = {
+  eng: { code: "en", name: "English" },
+  rus: { code: "ru", name: "Russian" },
+  ukr: { code: "uk", name: "Ukrainian" },
+  bel: { code: "be", name: "Belarusian" },
+  jpn: { code: "ja", name: "Japanese" },
+  kor: { code: "ko", name: "Korean" },
+  cmn: { code: "zh", name: "Chinese" },
+  spa: { code: "es", name: "Spanish" },
+  fra: { code: "fr", name: "French" },
+  deu: { code: "de", name: "German" },
+  ita: { code: "it", name: "Italian" },
+  por: { code: "pt", name: "Portuguese" },
+  pol: { code: "pl", name: "Polish" },
+  nld: { code: "nl", name: "Dutch" },
+  arb: { code: "ar", name: "Arabic" },
+  tur: { code: "tr", name: "Turkish" },
+  vie: { code: "vi", name: "Vietnamese" },
+  tha: { code: "th", name: "Thai" },
+  hin: { code: "hi", name: "Hindi" },
+  ind: { code: "id", name: "Indonesian" },
+  zlm: { code: "ms", name: "Malay" },
+  ces: { code: "cs", name: "Czech" },
+  slk: { code: "sk", name: "Slovak" },
+  ron: { code: "ro", name: "Romanian" },
+  hun: { code: "hu", name: "Hungarian" },
+  srp: { code: "sr", name: "Serbian" },
+  hrv: { code: "hr", name: "Croatian" },
+  bul: { code: "bg", name: "Bulgarian" },
+  ell: { code: "el", name: "Greek" },
+  heb: { code: "he", name: "Hebrew" },
+  dan: { code: "da", name: "Danish" },
+  fin: { code: "fi", name: "Finnish" },
+  nob: { code: "no", name: "Norwegian" },
+  swe: { code: "sv", name: "Swedish" },
+  fas: { code: "fa", name: "Persian" }
+};
+
+const ONLY = Object.keys(LANG_3_TO_1);
+
+/**
+ * The least text, in characters, that supports an answer of each language.
+ *
+ * MEASURED, not chosen — `research/franc-boundary-2026-09-02.md`. Method:
+ * Wikipedia extracts per language (deliberately NOT the UDHR, which is what
+ * franc's own profiles are built from and would read optimistically), 120
+ * windows cut at random from them at each length of a ladder from 40 to 1300
+ * characters, and the figure recorded is the shortest length from which franc
+ * answered correctly in at least 95 % of trials AND kept doing so at every
+ * longer length measured.
+ *
+ * Why it is per language rather than one number: the answer is not equally hard
+ * to reach, and the spread is fivefold. Greek and Korean settle at 40
+ * characters because their script settles it; English needs 130; Russian and
+ * Czech need 650, because each competes with neighbours in this very list for
+ * the same trigrams — Russian with Bulgarian, Serbian and Ukrainian, Czech with
+ * Slovak.
+ *
+ * Two languages are deliberately ABSENT. Swedish and Chinese did not settle
+ * anywhere in the ladder on the corpora collected, so no figure for them is
+ * measured and none is invented; they take the fallback below.
+ *
+ * A language with no entry gets the WORST measured figure. That is the
+ * conservative reading and it is still a measurement rather than a guess: it
+ * says "no better than the hardest language we have measured".
+ *
+ * @type {Record<string, number>}
+ */
+const LEAST_TEXT = {
+  bel: 100,
+  bul: 80,
+  ces: 650,
+  deu: 200,
+  ell: 40,
+  eng: 130,
+  fra: 80,
+  heb: 60,
+  ita: 160,
+  kor: 40,
+  nld: 130,
+  pol: 200,
+  por: 200,
+  rus: 650,
+  spa: 100,
+  srp: 100,
+  tur: 160,
+  ukr: 250
+};
+
+/** The worst measured figure, used for any language not in the table. */
+const LEAST_TEXT_WORST = Math.max(...Object.values(LEAST_TEXT), 0);
+
+/** franc's own floor: below this it is not asked at all. */
+const FRANC_FLOOR = 15;
+
+/** One space between words, nothing else — the form every figure above is in. */
+function normalise(text) {
+  return typeof text === "string" ? text.replace(/\s+/g, " ").trim() : "";
+}
+
+/**
+ * franc's answer for this text, or null when it has none.
+ *
+ * @param {string} text
+ * @returns {string | null} ISO 639-3.
+ */
+function ask(text) {
+  if (text.length < FRANC_FLOOR) {
+    return null;
+  }
+  const iso3 = franc(text, { only: ONLY, minLength: FRANC_FLOOR });
+  return iso3 === "und" ? null : iso3;
+}
+
+/**
+ * Best-effort detect the language of subtitle text.
+ *
+ * **Give this the words a viewer reads and nothing else.** franc scores letter
+ * trigrams over the whole string it is handed, so anything around the words
+ * competes with them. An ASS file is markup by half: measured 2026-09-01 on
+ * `[HorribleSubs] Drifters - 03 [1080p].ass`, 5040 Latin characters of Aegisub
+ * headers, style and font names, `Format:`/`Dialogue:` field prefixes and
+ * `{\…}` override groups against 5983 Cyrillic characters of dialogue —
+ * `franc(the file) = eng`, `franc(the dialogue) = rus`. The proxy had already
+ * built the markup-free text and detected on the file anyway, so a Russian
+ * track was offered to the viewer as English.
+ *
+ * `detectLanguageFromVtt` below is the safe entry point for a whole document;
+ * this one is for text that is already only text.
+ *
+ * @param {string} text - Subtitle text with no markup left in it.
+ * @returns {{ code: string, name: string } | null} Detected language, or null when uncertain.
+ */
+export function detectLanguage(text) {
+  const words = normalise(text);
+  if (words.length < FRANC_FLOOR) {
+    return null;
+  }
+  const candidate = ask(words);
+  if (candidate === null) {
+    return null;
+  }
+  // Enough text to support THIS answer. The figure is the language's own,
+  // because the languages are not alike: Russian shares its trigrams with
+  // Bulgarian, Serbian and Ukrainian and needs several times what English does.
+  if (words.length < (LEAST_TEXT[candidate] ?? LEAST_TEXT_WORST)) {
+    return null;
+  }
+  // And an answer that does not survive losing half the text was an accident of
+  // where the text happened to stop, not a reading of it. Free — franc costs
+  // about 2 ms whatever the size, measured — and it needs no figure of its own,
+  // because the test is taken on the text in hand.
+  const middle = Math.floor(words.length / 2);
+  if (ask(words.slice(0, middle)) !== candidate || ask(words.slice(middle)) !== candidate) {
+    return null;
+  }
+  return LANG_3_TO_1[candidate] ?? null;
+}