feat: add WhisperX diarization and speaker transcript UI

- add WhisperX diarization support to the Whisper VM server
- normalize speaker timestamp segments from Whisper responses
- document Hugging Face/pyannote VM setup and health checks
- show diarized speaker transcript blocks in record and transcript views
- group consecutive segments from the same speaker
- remove duplicate paragraph transcript display when diarized segments exist
- let diarized transcript content expand without an inner scrollbar
This commit is contained in:
KevinB-T committed 2026-05-20 16:34:50 +05:30
1 parent 3abb5e9281
commit 9517bad3dc
10 files changed
+674 -73

No files matched your search

+79 -7
View File
@@ -4,6 +4,7 @@ import { useEffect, useMemo, useRef, useState } from "react";
import { motion } from "framer-motion";
import {
Check,
Clock,
Loader2,
Mic,
Pause,
@@ -14,11 +15,12 @@ import {
Sparkles,
Square,
Trash2,
UserRound,
} from "lucide-react";
import { toast } from "sonner";
import { transcribeAudio } from "@/services/audio";
import { transcriptService } from "@/services/transcripts";
import type { Transcript } from "@/services/types";
import type { Transcript, TranscriptSegment } from "@/services/types";
import { AudioPlayer } from "@/components/audio-player";
import { SendTranscriptDialog } from "@/components/send-transcript-dialog";
@@ -483,12 +485,42 @@ function RecordPage() {
</Link>
</div>
</div>
<textarea
value={transcript}
onChange={(event) => setTranscript(event.target.value)}
rows={16}
className="mt-5 w-full resize-none rounded-xl border border-border bg-background/35 p-4 text-sm leading-relaxed outline-none transition focus:border-primary/60"
/>
{currentTranscript.timestamps.length > 0 ? (
<div className="mt-5 space-y-3 rounded-xl border border-border bg-background/35 p-4">
{groupTranscriptSegments(currentTranscript.timestamps).map((segment, index) => {
const speaker = segment.speaker ?? "Speaker";
return (
<div
key={`${segment.start}-${segment.end}-${index}`}
className="rounded-lg border border-border bg-secondary/20 p-3 text-sm"
>
<div className="mb-2 flex flex-wrap items-center gap-2">
<span
className={`inline-flex items-center gap-1.5 rounded-full border px-2.5 py-1 text-xs font-semibold ${speakerTone(
speaker,
)}`}
>
<UserRound className="h-3.5 w-3.5" />
{speaker}
</span>
<span className="inline-flex items-center gap-1.5 font-mono text-xs text-primary">
<Clock className="h-3.5 w-3.5" />
{formatStamp(segment.start)} - {formatStamp(segment.end)}
</span>
</div>
<p className="leading-relaxed text-muted-foreground">{segment.text}</p>
</div>
);
})}
</div>
) : (
<textarea
value={transcript}
onChange={(event) => setTranscript(event.target.value)}
rows={16}
className="mt-5 w-full resize-none rounded-xl border border-border bg-background/35 p-4 text-sm leading-relaxed outline-none transition focus:border-primary/60"
/>
)}
</section>
)}
@@ -506,3 +538,43 @@ function RecordPage() {
function formatTime(seconds: number) {
return `${String(Math.floor(seconds / 60)).padStart(2, "0")}:${String(seconds % 60).padStart(2, "0")}`;
}
function speakerTone(speaker: string) {
const tones = [
"border-primary/25 bg-primary/15 text-primary",
"border-emerald-400/25 bg-emerald-400/10 text-emerald-300",
"border-sky-400/25 bg-sky-400/10 text-sky-300",
"border-amber-400/25 bg-amber-400/10 text-amber-300",
"border-rose-400/25 bg-rose-400/10 text-rose-300",
];
const number = Number(speaker.match(/\d+/)?.[0] ?? 1);
return tones[(Math.max(number, 1) - 1) % tones.length];
}
function groupTranscriptSegments(segments: TranscriptSegment[]) {
return segments.reduce<TranscriptSegment[]>((groups, segment) => {
const previous = groups.at(-1);
const speaker = segment.speaker ?? "";
if (previous && (previous.speaker ?? "") === speaker) {
previous.end = segment.end;
previous.text = `${previous.text} ${segment.text}`.trim();
return groups;
}
groups.push({ ...segment });
return groups;
}, []);
}
function formatStamp(seconds: number) {
const safeSeconds = Number.isFinite(seconds) ? Math.max(seconds, 0) : 0;
const totalSeconds = Math.floor(safeSeconds);
const hours = Math.floor(totalSeconds / 3600);
const minutes = Math.floor((totalSeconds % 3600) / 60);
const secs = totalSeconds % 60;
if (hours > 0) {
return `${String(hours).padStart(2, "0")}:${String(minutes).padStart(2, "0")}:${String(
secs,
).padStart(2, "0")}`;
}
return `${String(minutes).padStart(2, "0")}:${String(secs).padStart(2, "0")}`;
}
@@ -1,11 +1,22 @@
import { createFileRoute, Link, useNavigate } from "@tanstack/react-router";
import { useQuery, useQueryClient } from "@tanstack/react-query";
import { useEffect, useRef, useState } from "react";
import { ArrowLeft, Copy, Download, Loader2, Save, Send, Sparkles, Trash2 } from "lucide-react";
import {
ArrowLeft,
Clock,
Copy,
Download,
Loader2,
Save,
Send,
Sparkles,
Trash2,
UserRound,
} from "lucide-react";
import { toast } from "sonner";
import { transcriptService } from "@/services/transcripts";
import { useAuth } from "@/context/auth";
import type { Transcript } from "@/services/types";
import type { Transcript, TranscriptSegment } from "@/services/types";
import { AudioPlayer } from "@/components/audio-player";
import { ConfirmDialog } from "@/components/confirm-dialog";
import { SendTranscriptDialog } from "@/components/send-transcript-dialog";
@@ -104,7 +115,7 @@ function TranscriptDetailPage() {
}
async function copy() {
await navigator.clipboard.writeText(text);
await navigator.clipboard.writeText(formatTranscriptForClipboard(transcript));
toast.success("Transcript copied");
}
@@ -277,34 +288,19 @@ function TranscriptDetailPage() {
</button>
)}
</div>
<textarea
value={text}
onChange={(event) => setText(event.target.value)}
rows={18}
readOnly={isReceived}
className="w-full resize-none bg-transparent text-sm leading-relaxed outline-none read-only:text-muted-foreground"
/>
{transcript.timestamps.length > 0 ? (
<SpeakerTranscript transcript={transcript} />
) : (
<textarea
value={text}
onChange={(event) => setText(event.target.value)}
rows={18}
readOnly={isReceived}
className="w-full resize-none bg-transparent text-sm leading-relaxed outline-none read-only:text-muted-foreground"
/>
)}
</div>
{transcript.timestamps.length > 0 && (
<div className="glass rounded-2xl p-6">
<h2 className="font-display text-lg font-semibold">Timestamps</h2>
<div className="mt-4 max-h-72 space-y-2 overflow-y-auto">
{transcript.timestamps.map((segment, index) => (
<div
key={`${segment.start}-${index}`}
className="grid gap-3 rounded-xl border border-border bg-secondary/20 p-3 text-sm md:grid-cols-[120px_1fr]"
>
<span className="font-mono text-xs text-primary">
{formatStamp(segment.start)} - {formatStamp(segment.end)}
</span>
<span className="text-muted-foreground">{segment.text}</span>
</div>
))}
</div>
</div>
)}
{showSend && (
<SendTranscriptDialog sending={sending} onClose={() => setShowSend(false)} onSend={send} />
)}
@@ -322,8 +318,105 @@ function TranscriptDetailPage() {
);
}
function formatStamp(seconds: number) {
const minutes = Math.floor(seconds / 60);
const secs = Math.floor(seconds % 60);
return `${minutes}:${String(secs).padStart(2, "0")}`;
function SpeakerTranscript({ transcript }: { transcript: Transcript }) {
const hasSpeakers = transcript.timestamps.some((segment) => segment.speaker);
const groupedSegments = groupTranscriptSegments(transcript.timestamps);
return (
<div>
<div className="flex flex-wrap items-center justify-between gap-3">
<h2 className="font-display text-lg font-semibold">
{hasSpeakers ? "Speaker transcript" : "Timestamps"}
</h2>
{hasSpeakers && (
<span className="rounded-full border border-primary/20 bg-primary/10 px-3 py-1 text-xs font-medium text-primary">
{speakerCount(transcript)} {speakerCount(transcript) === 1 ? "speaker" : "speakers"}
</span>
)}
</div>
<div className="mt-4 space-y-3">
{groupedSegments.map((segment, index) => {
const speaker = segment.speaker ?? "Speaker";
return (
<div
key={`${segment.start}-${segment.end}-${index}`}
className="rounded-xl border border-border bg-secondary/20 p-4 text-sm"
>
<div className="mb-2 flex flex-wrap items-center gap-2">
{segment.speaker && (
<span
className={`inline-flex items-center gap-1.5 rounded-full border px-2.5 py-1 text-xs font-semibold ${speakerTone(
speaker,
)}`}
>
<UserRound className="h-3.5 w-3.5" />
{speaker}
</span>
)}
<span className="inline-flex items-center gap-1.5 font-mono text-xs text-primary">
<Clock className="h-3.5 w-3.5" />
{formatStamp(segment.start)} - {formatStamp(segment.end)}
</span>
</div>
<p className="leading-relaxed text-muted-foreground">{segment.text}</p>
</div>
);
})}
</div>
</div>
);
}
function formatTranscriptForClipboard(transcript: Transcript | null) {
if (!transcript?.timestamps.length) return transcript?.transcriptText ?? "";
return groupTranscriptSegments(transcript.timestamps)
.map((segment) => {
const speaker = segment.speaker ? `${segment.speaker}\n` : "";
return `${speaker}${formatStamp(segment.start)} - ${formatStamp(segment.end)}\n${segment.text}`;
})
.join("\n\n");
}
function speakerCount(transcript: Transcript) {
return new Set(transcript.timestamps.map((segment) => segment.speaker).filter(Boolean)).size;
}
function groupTranscriptSegments(segments: TranscriptSegment[]) {
return segments.reduce<TranscriptSegment[]>((groups, segment) => {
const previous = groups.at(-1);
const speaker = segment.speaker ?? "";
if (previous && (previous.speaker ?? "") === speaker) {
previous.end = segment.end;
previous.text = `${previous.text} ${segment.text}`.trim();
return groups;
}
groups.push({ ...segment });
return groups;
}, []);
}
function speakerTone(speaker: string) {
const tones = [
"border-primary/25 bg-primary/15 text-primary",
"border-emerald-400/25 bg-emerald-400/10 text-emerald-300",
"border-sky-400/25 bg-sky-400/10 text-sky-300",
"border-amber-400/25 bg-amber-400/10 text-amber-300",
"border-rose-400/25 bg-rose-400/10 text-rose-300",
];
const number = Number(speaker.match(/\d+/)?.[0] ?? 1);
return tones[(Math.max(number, 1) - 1) % tones.length];
}
function formatStamp(seconds: number) {
const safeSeconds = Number.isFinite(seconds) ? Math.max(seconds, 0) : 0;
const totalSeconds = Math.floor(safeSeconds);
const hours = Math.floor(totalSeconds / 3600);
const minutes = Math.floor((totalSeconds % 3600) / 60);
const secs = totalSeconds % 60;
if (hours > 0) {
return `${String(hours).padStart(2, "0")}:${String(minutes).padStart(2, "0")}:${String(
secs,
).padStart(2, "0")}`;
}
return `${String(minutes).padStart(2, "0")}:${String(secs).padStart(2, "0")}`;
}
+1
View File
@@ -10,6 +10,7 @@ export type User = {
export type UserSummary = Pick<User, "id" | "fullName" | "username" | "email">;
export type TranscriptSegment = {
speaker?: string;
start: number;
end: number;
text: string;
+45 -1
View File
@@ -1,3 +1,7 @@
import { execFileSync } from "node:child_process";
import { existsSync, mkdirSync, readFileSync } from "node:fs";
import { tmpdir } from "node:os";
import path from "node:path";
import { fileURLToPath, URL } from "node:url";
import tailwindcss from "@tailwindcss/vite";
import { tanstackRouter } from "@tanstack/router-plugin/vite";
@@ -5,11 +9,49 @@ import viteReact from "@vitejs/plugin-react";
import { defineConfig, loadEnv } from "vite";
import tsConfigPaths from "vite-tsconfig-paths";
export default defineConfig(({ mode }) => {
function getDevHttpsOptions() {
const certDir = path.join(tmpdir(), "orphion-vite-https");
const keyPath = path.join(certDir, "localhost-key.pem");
const certPath = path.join(certDir, "localhost-cert.pem");
if (!existsSync(keyPath) || !existsSync(certPath)) {
mkdirSync(certDir, { recursive: true });
execFileSync(
"openssl",
[
"req",
"-x509",
"-newkey",
"rsa:2048",
"-nodes",
"-sha256",
"-days",
"365",
"-subj",
"/CN=localhost",
"-addext",
"subjectAltName=DNS:localhost,IP:127.0.0.1,IP:0.0.0.0",
"-keyout",
keyPath,
"-out",
certPath,
],
{ stdio: "ignore" },
);
}
return {
key: readFileSync(keyPath),
cert: readFileSync(certPath),
};
}
export default defineConfig(({ command, mode }) => {
const viteEnv = loadEnv(mode, process.cwd(), "");
const serviceHost = viteEnv.VITE_ORPHION_SERVICE_HOST || "127.0.0.1";
const apiPort = viteEnv.VITE_API_PORT || "4000";
const apiProxyTarget = viteEnv.VITE_API_PROXY_TARGET || `http://${serviceHost}:${apiPort}`;
const https = command === "serve" ? getDevHttpsOptions() : undefined;
return {
plugins: [
@@ -30,6 +72,7 @@ export default defineConfig(({ mode }) => {
dedupe: ["react", "react-dom", "@tanstack/react-query", "@tanstack/query-core"],
},
server: {
...(https ? { https } : {}),
proxy: {
"/api": {
target: apiProxyTarget,
@@ -37,6 +80,7 @@ export default defineConfig(({ mode }) => {
},
},
},
...(https ? { preview: { https } } : {}),
build: {
rollupOptions: {
output: {