Skip to content

Write HTML. Render video. Built for agents.

fix(cli): retain WebVTT timestamps before cue settings (#5224)

Abdullah M. Naziff5aeecdbf4b4
diff --git a/packages/cli/src/whisper/normalize.test.ts b/packages/cli/src/whisper/normalize.test.ts
index 969bbff3b0..7aa187fc80 100644
--- a/packages/cli/src/whisper/normalize.test.ts
+++ b/packages/cli/src/whisper/normalize.test.ts
@@ -83,6 +83,24 @@ describe("detectFormat", () => {
});
describe("loadTranscript", () => {
+ it.each([
+ ["00:00:01.000", "00:00:03.500", "align:start"],
+ ["00:01.000", "00:03.500", "align:start"],
+ ["00:00:01.000", "00:00:03.500", "line:90% position:50%,center size:80% align:center"],
+ ["00:01.000", "00:03.500", "\tvertical:rl\tline:0"],
+ ["00:00:01.000", "00:00:03.500", "region:captions"],
+ ])("retains VTT end time before cue settings: %s --> %s %s", (start, end, settings) => {
+ const source = `WEBVTT\n\nfirst\n${start} --> ${end} ${settings}\nFirst phrase\n\nsecond\n00:04.000 --> 00:06.000\nSecond phrase\n`;
+ const { words } = loadTranscript(tmpFile("settings.vtt", source));
+ expect(words).toEqual([
+ { text: "First phrase", start: 1, end: 3.5, id: "w0" },
+ { text: "Second phrase", start: 4, end: 6, id: "w1" },
+ ]);
+ expect(formatSrt(words, { preGrouped: true })).toBe(
+ "1\n00:00:01,000 --> 00:00:03,500\nFirst phrase\n\n2\n00:00:04,000 --> 00:00:06,000\nSecond phrase\n",
+ );
+ });
+
it("reads an empty word list as a transcript with no words", () => {
expect(loadTranscript(tmpFile("transcript.json", "[]"))).toEqual({
words: [],
diff --git a/packages/cli/src/whisper/normalize.ts b/packages/cli/src/whisper/normalize.ts
index 6683d634af..9ddbc9ebbe 100644
--- a/packages/cli/src/whisper/normalize.ts
+++ b/packages/cli/src/whisper/normalize.ts
@@ -266,8 +266,9 @@ function parseSrtTimestamp(ts: string): number {
/** Parse VTT timestamp: 00:01:23.456 or 01:23.456 → seconds */
function parseVttTimestamp(ts: string): number {
- const parts = ts.split(":");
- if (parts.length === 3) return parseSrtTimestamp(ts);
+ const timestamp = ts.split(/[ \t]+/, 1)[0] ?? "";
+ const parts = timestamp.split(":");
+ if (parts.length === 3) return parseSrtTimestamp(timestamp);
// MM:SS.mmm
if (parts.length === 2) {
const [min, secMs] = parts;