youtube-transcript-plus 1.1.2 → 2.0.0

package/dist/youtube-transcript-plus.js

@@ -1,5 +1,5 @@
-import fs from 'fs/promises';
-import path from 'path';
+import fs from 'node:fs/promises';
+import path from 'node:path';
 
 /******************************************************************************
 Copyright (c) Microsoft Corporation.
@@ -48,24 +48,30 @@ class YoutubeTranscriptVideoUnavailableError extends Error {
     constructor(videoId) {
         super(`The video with ID "${videoId}" is no longer available or has been removed. Please check the video URL or ID and try again.`);
         this.name = 'YoutubeTranscriptVideoUnavailableError';
+        this.videoId = videoId;
     }
 }
 class YoutubeTranscriptDisabledError extends Error {
     constructor(videoId) {
         super(`Transcripts are disabled for the video with ID "${videoId}". This may be due to the video owner disabling captions or the video not supporting transcripts.`);
         this.name = 'YoutubeTranscriptDisabledError';
+        this.videoId = videoId;
     }
 }
 class YoutubeTranscriptNotAvailableError extends Error {
     constructor(videoId) {
         super(`No transcripts are available for the video with ID "${videoId}". This may be because the video does not have captions or the captions are not accessible.`);
         this.name = 'YoutubeTranscriptNotAvailableError';
+        this.videoId = videoId;
     }
 }
 class YoutubeTranscriptNotAvailableLanguageError extends Error {
     constructor(lang, availableLangs, videoId) {
         super(`No transcripts are available in "${lang}" for the video with ID "${videoId}". Available languages: ${availableLangs.join(', ')}. Please try a different language.`);
         this.name = 'YoutubeTranscriptNotAvailableLanguageError';
+        this.videoId = videoId;
+        this.lang = lang;
+        this.availableLangs = availableLangs;
     }
 }
 class YoutubeTranscriptInvalidVideoIdError extends Error {
@@ -75,8 +81,21 @@ class YoutubeTranscriptInvalidVideoIdError extends Error {
     }
 }
 
+const RE_VIDEO_ID = /^[a-zA-Z0-9_-]{11}$/;
+const XML_ENTITIES = {
+    '&': '&',
+    '&lt;': '<',
+    '&gt;': '>',
+    '&quot;': '"',
+    '&#39;': "'",
+    '&apos;': "'",
+};
+const RE_XML_ENTITY = /&(?:amp|lt|gt|quot|apos|#39);/g;
+function decodeXmlEntities(text) {
+    return text.replace(RE_XML_ENTITY, (match) => { var _a; return (_a = XML_ENTITIES[match]) !== null && _a !== void 0 ? _a : match; });
+}
 function retrieveVideoId(videoId) {
-    if (videoId.length === 11) {
+    if (RE_VIDEO_ID.test(videoId)) {
         return videoId;
     }
     const matchId = videoId.match(RE_YOUTUBE);
@@ -100,15 +119,19 @@ function defaultFetch(params) {
     });
 }
 
+function sanitizeKey(key) {
+    return key.replace(/[^a-zA-Z0-9_-]/g, '_');
+}
 class FsCache {
     constructor(cacheDir = './cache', defaultTTL = DEFAULT_CACHE_TTL) {
         this.cacheDir = cacheDir;
         this.defaultTTL = defaultTTL;
-        fs.mkdir(cacheDir, { recursive: true }).catch(() => { });
+        this.ready = fs.mkdir(cacheDir, { recursive: true }).then(() => { });
     }
     get(key) {
         return __awaiter(this, void 0, void 0, function* () {
-            const filePath = path.join(this.cacheDir, key);
+            yield this.ready;
+            const filePath = path.join(this.cacheDir, sanitizeKey(key));
             try {
                 const data = yield fs.readFile(filePath, 'utf-8');
                 const { value, expires } = JSON.parse(data);
@@ -123,7 +146,8 @@ class FsCache {
     }
     set(key, value, ttl) {
         return __awaiter(this, void 0, void 0, function* () {
-            const filePath = path.join(this.cacheDir, key);
+            yield this.ready;
+            const filePath = path.join(this.cacheDir, sanitizeKey(key));
             const expires = Date.now() + (ttl !== null && ttl !== void 0 ? ttl : this.defaultTTL);
             yield fs.writeFile(filePath, JSON.stringify({ value, expires }), 'utf-8');
         });
@@ -165,7 +189,7 @@ class YoutubeTranscript {
     }
     fetchTranscript(videoId) {
         return __awaiter(this, void 0, void 0, function* () {
-            var _a, _b, _c, _d, _e, _f, _g, _h, _j, _k, _l, _m, _o;
+            var _a, _b, _c, _d, _e, _f, _g, _h, _j, _k, _l, _m, _o, _p;
             const identifier = retrieveVideoId(videoId);
             const lang = (_a = this.config) === null || _a === void 0 ? void 0 : _a.lang;
             const userAgent = (_c = (_b = this.config) === null || _b === void 0 ? void 0 : _b.userAgent) !== null && _c !== void 0 ? _c : DEFAULT_USER_AGENT;
@@ -179,7 +203,7 @@ class YoutubeTranscript {
                     try {
                         return JSON.parse(cached);
                     }
-                    catch (_p) {
+                    catch (_q) {
                         // ignore parse errors and continue
                     }
                 }
@@ -209,7 +233,7 @@ class YoutubeTranscript {
             }
             const apiKey = apiKeyMatch[1];
             // 3) Call Innertube player as ANDROID client to retrieve captionTracks
-            const playerEndpoint = `https://www.youtube.com/youtubei/v1/player?key=${apiKey}`;
+            const playerEndpoint = `${protocol}://www.youtube.com/youtubei/v1/player?key=${apiKey}`;
             const playerBody = {
                 context: {
                     client: {
@@ -234,17 +258,17 @@ class YoutubeTranscript {
             if (!playerRes.ok) {
                 throw new YoutubeTranscriptVideoUnavailableError(identifier);
             }
-            const playerJson = yield playerRes.json();
-            const tracklist = (_k = (_j = playerJson === null || playerJson === void 0 ? void 0 : playerJson.captions) === null || _j === void 0 ? void 0 : _j.playerCaptionsTracklistRenderer) !== null && _k !== void 0 ? _k : playerJson === null || playerJson === void 0 ? void 0 : playerJson.playerCaptionsTracklistRenderer;
+            const playerJson = (yield playerRes.json());
+            const tracklist = (_k = (_j = playerJson.captions) === null || _j === void 0 ? void 0 : _j.playerCaptionsTracklistRenderer) !== null && _k !== void 0 ? _k : playerJson.playerCaptionsTracklistRenderer;
             const tracks = tracklist === null || tracklist === void 0 ? void 0 : tracklist.captionTracks;
-            const isPlayableOk = ((_l = playerJson === null || playerJson === void 0 ? void 0 : playerJson.playabilityStatus) === null || _l === void 0 ? void 0 : _l.status) === 'OK';
+            const isPlayableOk = ((_l = playerJson.playabilityStatus) === null || _l === void 0 ? void 0 : _l.status) === 'OK';
             // If `captions` is entirely missing, treat as "not available"
-            if (!(playerJson === null || playerJson === void 0 ? void 0 : playerJson.captions) || !tracklist) {
-                // If video is playable but captions aren’t provided, treat as "disabled"
+            if (!playerJson.captions || !tracklist) {
+                // If video is playable but captions aren't provided, treat as "disabled"
                 if (isPlayableOk) {
                     throw new YoutubeTranscriptDisabledError(identifier);
                 }
-                // Otherwise we can’t assert they’re disabled; treat as "not available"
+                // Otherwise we can't assert they're disabled; treat as "not available"
                 throw new YoutubeTranscriptNotAvailableError(identifier);
             }
             // If `captions` exists but there are zero tracks, treat as "disabled"
@@ -252,22 +276,25 @@ class YoutubeTranscript {
                 throw new YoutubeTranscriptDisabledError(identifier);
             }
             // Respect requested language or fallback to first track
-            const selectedTrack = lang ? tracks.find((t) => t.languageCode === lang) : tracks[0];
+            const selectedTrack = lang
+                ? tracks.find((t) => t.languageCode === lang)
+                : tracks[0];
             if (!selectedTrack) {
                 const available = tracks.map((t) => t.languageCode).filter(Boolean);
                 throw new YoutubeTranscriptNotAvailableLanguageError(lang, available, identifier);
             }
             // 4) Build transcript URL; prefer XML by stripping fmt if present
-            let transcriptURL = selectedTrack.baseUrl || selectedTrack.url;
-            if (!transcriptURL) {
+            const transcriptBaseURL = (_m = selectedTrack.baseUrl) !== null && _m !== void 0 ? _m : selectedTrack.url;
+            if (!transcriptBaseURL) {
                 throw new YoutubeTranscriptNotAvailableError(identifier);
             }
-            transcriptURL = transcriptURL.replace(/&fmt=[^&]+$/, '');
-            if ((_m = this.config) === null || _m === void 0 ? void 0 : _m.disableHttps) {
+            let transcriptURL = transcriptBaseURL;
+            transcriptURL = transcriptURL.replace(/&fmt=[^&]+/, '');
+            if ((_o = this.config) === null || _o === void 0 ? void 0 : _o.disableHttps) {
                 transcriptURL = transcriptURL.replace(/^https:\/\//, 'http://');
             }
             // 5) Fetch transcript XML using the same hook surface as before
-            const transcriptResponse = ((_o = this.config) === null || _o === void 0 ? void 0 : _o.transcriptFetch)
+            const transcriptResponse = ((_p = this.config) === null || _p === void 0 ? void 0 : _p.transcriptFetch)
                 ? yield this.config.transcriptFetch({ url: transcriptURL, lang, userAgent })
                 : yield defaultFetch({ url: transcriptURL, lang, userAgent });
             if (!transcriptResponse.ok) {
@@ -281,7 +308,7 @@ class YoutubeTranscript {
             // 6) Parse XML into the existing TranscriptResponse shape
             const results = [...transcriptBody.matchAll(RE_XML_TRANSCRIPT)];
             const transcript = results.map((m) => ({
-                text: m[3],
+                text: decodeXmlEntities(m[3]),
                 duration: parseFloat(m[2]),
                 offset: parseFloat(m[1]),
                 lang: lang !== null && lang !== void 0 ? lang : selectedTrack.languageCode,
@@ -294,7 +321,7 @@ class YoutubeTranscript {
                 try {
                     yield cache.set(cacheKey, JSON.stringify(transcript), cacheTTL);
                 }
-                catch (_q) {
+                catch (_r) {
                     // non-fatal
                 }
             }