api/rutube: support SRT transcript sources

api/src/processing/helpers/caption-formats.js

@@ -61,6 +61,30 @@ export const parseVtt = input => {
     return cues;
 };
 
+export const parseSrt = input => {
+    const normalized = input.replace(/^\uFEFF/, "").replace(/\r\n?/g, "\n");
+    const cues = [];
+
+    for (const block of normalized.split(/\n{2,}/)) {
+        const lines = block.split("\n").map(line => line.trimEnd());
+        const timingIndex = lines.findIndex(line => line.includes(" --> "));
+        if (timingIndex < 0) continue;
+
+        const timing = lines[timingIndex].match(
+            /^(\d{2}:)?\d{2}:\d{2},\d{3}\s+-->\s+(\d{2}:)?\d{2}:\d{2},\d{3}/
+        )?.[0];
+        if (!timing) continue;
+
+        const [ start, end ] = timing.split(/\s+-->\s+/).map(value =>
+            value.replace(",", ".")
+        );
+        const text = cleanCueText(lines.slice(timingIndex + 1).join("\n"));
+        if (text) cues.push({ start, end, text });
+    }
+
+    return cues;
+};
+
 const transcriptLines = cues => {
     const lines = [];
     for (const cue of cues) {
@@ -83,14 +107,21 @@ const srtTimestamp = timestamp => {
 
 const escapeMarkdown = value => value.replace(/[\\`*_{}\[\]<>#+.!|~-]/g, "\\$&");
 
-export const convertCaptions = (vtt, format, metadata = {}) => {
-    if (format === "vtt") {
-        return vtt.replace(/^\uFEFF/, "").replace(/\r\n?/g, "\n");
+export const convertCaptions = (input, format, metadata = {}, sourceFormat = "vtt") => {
+    if (sourceFormat === "vtt" && format === "vtt") {
+        return input.replace(/^\uFEFF/, "").replace(/\r\n?/g, "\n");
     }
 
-    const cues = parseVtt(vtt);
+    const cues = sourceFormat === "srt" ? parseSrt(input) : parseVtt(input);
     if (!cues.length) throw new Error("caption file has no cues");
 
+    if (format === "vtt") {
+        return "WEBVTT\n\n" + cues.map(cue => [
+            `${cue.start} --> ${cue.end}`,
+            cue.text,
+        ].join("\n")).join("\n\n") + "\n";
+    }
+
     if (format === "srt") {
         return cues.map((cue, index) => [
             index + 1,

api/src/processing/helpers/captions.js

@@ -12,6 +12,7 @@ export const matchesCaptionLanguage = (language, requested) => {
 export const createCaptionResponse = ({
     url,
     format,
+    sourceFormat = "vtt",
     language,
     service,
     id,
@@ -27,6 +28,7 @@ export const createCaptionResponse = ({
         urls: url,
         headers,
         captionFormat: format,
+        captionSourceFormat: sourceFormat,
         captionLanguage: language || "unknown",
         filenameAttributes: {
             service,

api/src/processing/helpers/tests/captions.test.js

@@ -2,7 +2,7 @@ import test from "node:test";
 import assert from "node:assert/strict";
 
 import { selectCaptionTrack } from "../captions.js";
-import { convertCaptions, parseVtt } from "../caption-formats.js";
+import { convertCaptions, parseSrt, parseVtt } from "../caption-formats.js";
 import { apiSchema } from "../../schema.js";
 
 const sample = `\uFEFFWEBVTT\r
@@ -21,6 +21,15 @@ Hello & welcome back\r
 <v Speaker>Final line</v>\r
 `;
 
+const sampleSrt = `1\r
+00:00:00,000 --> 00:00:02,500\r
+Hello &amp; welcome\r
+\r
+2\r
+00:00:02,500 --> 00:00:04,000\r
+Final line\r
+`;
+
 test("parseVtt extracts and cleans cues", () => {
     assert.deepEqual(parseVtt(sample), [
         { start: "00:00.000", end: "00:02.500", text: "Hello & welcome" },
@@ -29,6 +38,18 @@ test("parseVtt extracts and cleans cues", () => {
     ]);
 });
 
+test("SRT input converts to transcripts and VTT", () => {
+    assert.equal(parseSrt(sampleSrt).length, 2);
+    assert.equal(
+        convertCaptions(sampleSrt, "txt", {}, "srt"),
+        "Hello & welcome\n\nFinal line\n"
+    );
+    assert.match(
+        convertCaptions(sampleSrt, "vtt", {}, "srt"),
+        /^WEBVTT\n\n00:00:00\.000 --> 00:00:02\.500/
+    );
+});
+
 test("plain text collapses growing automatic-caption cues", () => {
     assert.equal(
         convertCaptions(sample, "txt"),

api/src/processing/services/rutube.js

@@ -48,7 +48,7 @@ export default async function(obj) {
 
     if (obj.isCaptionOnly) {
         const caption = play.captions?.find(caption =>
-            caption.format === "webvtt"
+            ["webvtt", "srt"].includes(caption.format)
             && matchesCaptionLanguage(caption.code, obj.captionLanguage)
         );
         if (!caption) return { error: "fetch.empty" };
@@ -56,6 +56,7 @@ export default async function(obj) {
         return createCaptionResponse({
             url: caption.file,
             format: obj.captionFormat,
+            sourceFormat: caption.format === "webvtt" ? "vtt" : caption.format,
             language: caption.code,
             service: "rutube",
             id: obj.id,

api/src/stream/captions.js

@@ -55,7 +55,8 @@ export default async function captions(streamInfo, res) {
         const body = convertCaptions(
             Buffer.concat(chunks).toString("utf8"),
             streamInfo.captionFormat,
-            streamInfo.captionMetadata
+            streamInfo.captionMetadata,
+            streamInfo.captionSourceFormat
         );
 
         res.setHeader("Cross-Origin-Resource-Policy", "cross-origin");

api/src/util/tests/rutube.json

@@ -10,11 +10,11 @@
     },
     {
         "name": "standalone transcript",
-        "url": "https://rutube.ru/video/b2f6c27649907c2fde0af411b03825eb/",
-        "canFail": true,
+        "url": "https://rutube.ru/video/76ee456fa0fe79a8e0013b6110cb0168/",
         "params": {
             "downloadMode": "captions",
-            "captionFormat": "txt"
+            "captionFormat": "txt",
+            "captionLanguage": "ru"
         },
         "expected": {
             "code": 200,