Open source Loom alternative. Beautiful, shareable screen recordings.
Rust appcapcossloommacnextjsnextjs15open-sourceossreactrecordrustscreen-capturescreen-recorderscreenshotsolidjstauritauri-apptypescriptvite
import type {
CaptionSegment,
SegmentRecordings,
TimelineSegment,
} from "~/utils/tauri";
import {
getCaptionTextFromWords,
mapCaptionsToEditedTimeline,
} from "./captions";
import type { ClipTransition } from "./clip-transitions";
import type { TextSegment } from "./text";
export type CaptionExportFormat = "srt" | "vtt";
export interface CaptionExportCue {
startMs: number;
endMs: number;
text: string;
}
const DOUBLE_QUOTE = String.fromCharCode(34);
const INVALID_FILE_NAME_CHARS = new Set([
"<",
">",
":",
DOUBLE_QUOTE,
"/",
"\\",
"|",
"?",
"*",
]);
function millisecondsFromSeconds(seconds: number) {
return Math.max(0, Math.round(seconds * 1000));
}
function formatTimestamp(ms: number, separator: "," | ".") {
const hours = Math.floor(ms / 3_600_000);
const minutes = Math.floor((ms % 3_600_000) / 60_000);
const seconds = Math.floor((ms % 60_000) / 1000);
const milliseconds = ms % 1000;
return `${hours.toString().padStart(2, "0")}:${minutes.toString().padStart(2, "0")}:${seconds.toString().padStart(2, "0")}${separator}${milliseconds.toString().padStart(3, "0")}`;
}
function textFromCaptionSegment(segment: CaptionSegment) {
const words = segment.words ?? [];
if (words.length > 0) return getCaptionTextFromWords(words);
return segment.text;
}
function cueRangeFromCaptionSegment(segment: CaptionSegment) {
const words = segment.words ?? [];
const firstWord = words[0];
const lastWord = words[words.length - 1];
return {
start: firstWord?.start ?? segment.start,
end: lastWord?.end ?? segment.end,
};
}
function normalizeCueText(text: string) {
return text
.replace(/\r\n?/g, "\n")
.split("")
.filter((char) => {
const code = char.charCodeAt(0);
return code === 10 || (code >= 32 && code !== 127);
})
.join("")
.split("\n")
.map((line) => line.trim().replace(/\s+/g, " "))
.filter((line) => line.length > 0)
.join("\n")
.trim();
}
function normalizeVttCueText(text: string) {
return text
.replace(/&/g, "&")
.replace(/</g, "<")
.replace(/>/g, ">");
}
function cueFromCaptionSegment(
segment: CaptionSegment,
): CaptionExportCue | null {
const text = normalizeCueText(textFromCaptionSegment(segment));
const { start, end } = cueRangeFromCaptionSegment(segment);
if (
text.length === 0 ||
!Number.isFinite(start) ||
!Number.isFinite(end) ||
end <= start
) {
return null;
}
const startMs = millisecondsFromSeconds(start);
const roundedEndMs = millisecondsFromSeconds(end);
const endMs = roundedEndMs <= startMs ? startMs + 1 : roundedEndMs;
return { startMs, endMs, text };
}
export function createCaptionExportCues(
segments: CaptionSegment[],
timelineSegments: TimelineSegment[],
recordingSegments: SegmentRecordings[],
transitions: ClipTransition[] = [],
textSegments?: readonly TextSegment[],
): CaptionExportCue[] {
return mapCaptionsToEditedTimeline(
segments,
timelineSegments,
recordingSegments,
transitions,
textSegments,
)
.map(cueFromCaptionSegment)
.filter((cue): cue is CaptionExportCue => cue !== null)
.sort((a, b) => a.startMs - b.startMs || a.endMs - b.endMs);
}
export function formatCaptionCuesAsSrt(cues: CaptionExportCue[]) {
if (cues.length === 0) return "";
return `${cues
.map(
(cue, index) =>
`${index + 1}\n${formatTimestamp(cue.startMs, ",")} --> ${formatTimestamp(cue.endMs, ",")}\n${cue.text}`,
)
.join("\n\n")}\n`;
}
export function formatCaptionCuesAsVtt(cues: CaptionExportCue[]) {
if (cues.length === 0) return "WEBVTT\n";
return `WEBVTT\n\n${cues
.map(
(cue, index) =>
`${index + 1}\n${formatTimestamp(cue.startMs, ".")} --> ${formatTimestamp(cue.endMs, ".")}\n${normalizeVttCueText(cue.text)}`,
)
.join("\n\n")}\n`;
}
export function formatCaptionCues(
cues: CaptionExportCue[],
format: CaptionExportFormat,
) {
return format === "srt"
? formatCaptionCuesAsSrt(cues)
: formatCaptionCuesAsVtt(cues);
}
export function captionExportDefaultPath(
name: string,
format: CaptionExportFormat,
) {
const cleanedName = name
.trim()
.split("")
.map((char) => {
const code = char.charCodeAt(0);
return INVALID_FILE_NAME_CHARS.has(char) || code < 32 ? "-" : char;
})
.join("")
.replace(/\s+/g, " ")
.replace(/\.+$/g, "")
.slice(0, 120)
.trim();
return `${cleanedName || "captions"}.${format}`;
}
if (import.meta.vitest) {
const { describe, expect, it } = import.meta.vitest;
describe("caption exports", () => {
const recordings = [{ display: { duration: 8 } } as SegmentRecordings];
it("formats projected edited timeline cues as SRT", () => {
const cues = createCaptionExportCues(
[
{
id: "caption",
start: 1,
end: 5,
text: "hello world",
words: [
{ text: "hello", start: 1.1114, end: 1.4446 },
{ text: "world", start: 4.2, end: 4.6 },
],
},
],
[
{ start: 1, end: 2, timescale: 1, recordingSegment: 0 },
{ start: 4, end: 6, timescale: 1, recordingSegment: 0 },
],
recordings,
);
expect(formatCaptionCuesAsSrt(cues)).toBe(
"1\n00:00:00,111 --> 00:00:00,445\nhello\n\n2\n00:00:01,200 --> 00:00:01,600\nworld\n",
);
});
it("formats VTT with escaped cue text", () => {
expect(
formatCaptionCuesAsVtt([
{
startMs: 0,
endMs: 1250,
text: "Cap <Rend> & fast --> captions",
},
]),
).toBe(
"WEBVTT\n\n1\n00:00:00.000 --> 00:00:01.250\nCap <Rend> & fast --> captions\n",
);
});
it("cleans invalid filenames and falls back when empty", () => {
expect(
captionExportDefaultPath(["bad/name:", DOUBLE_QUOTE].join(""), "srt"),
).toBe("bad-name--.srt");
expect(captionExportDefaultPath("...", "vtt")).toBe("captions.vtt");
});
it("skips empty and invalid cues", () => {
const cues = createCaptionExportCues(
[
{
id: "empty",
start: 0,
end: 1,
text: " ",
words: [],
},
{
id: "backwards",
start: 2,
end: 1,
text: "backwards",
words: [],
},
{
id: "valid",
start: 1,
end: 1.0001,
text: "valid",
words: [],
},
],
[{ start: 0, end: 8, timescale: 1, recordingSegment: 0 }],
recordings,
);
expect(cues).toEqual([{ startMs: 1000, endMs: 1001, text: "valid" }]);
});
});
}