Skip to content

Speakr is a personal, self-hosted web application designed for transcribing audio recordings

Python

Fix queue #1 — recording Stats tab — new Stats tab in the right rail (shown only when transcript segments carry start/end times) with: (a) headline metric cards for total length / number of speakers / total turns / total words; (b) per-speaker table sorted by speaking time with speaking duration (mm:ss), %% of total audio, turn count, word count and words-per-minute, each speaker row painting a slim proportion bar in the accent colour so the distribution reads at a glance; (c) a trailing Silence row aggregating leading/trailing gaps and between-segment gaps, with its own proportion bar in the muted colour; computation is a pure Vue computed (recordingStats) derived from processedTranscription.simpleSegments + recording.duration — no backend change; hasRecordingStats computed gates the tab button so the Stats tab only appears when the data is meaningful; falls back to an .empty primitive when shown without data; queue tracking file at internal-docs/ui-fix-queue.md

Murtaza Nasir8d338543d103
diff --git a/static/js/app.modular.js b/static/js/app.modular.js
index 775de555..d55509c4 100644
--- a/static/js/app.modular.js
+++ b/static/js/app.modular.js
@@ -1118,6 +1118,119 @@ document.addEventListener('DOMContentLoaded', async () => {
return n ? _formatCount(n) : '';
});
+ // ---------------------------------------------------------------
+ // Recording stats — per-speaker speaking time, % of audio, turn
+ // counts, word counts and WPM, plus a silence row aggregating the
+ // gaps between consecutive segments. Surfaces a Stats tab in the
+ // right rail only when diarized transcript segments with
+ // start/end times are available; otherwise empty/null.
+ //
+ // Pure derivation from processedTranscription.simpleSegments +
+ // recording.duration. No backend change.
+ // ---------------------------------------------------------------
+ const _fmtMmSs = (totalSec) => {
+ if (totalSec == null || !isFinite(totalSec)) return '—';
+ const s = Math.max(0, Math.round(totalSec));
+ const m = Math.floor(s / 60);
+ const r = s % 60;
+ return `${m}:${String(r).padStart(2, '0')}`;
+ };
+ const recordingStats = computed(() => {
+ const rec = selectedRecording.value;
+ const pt = processedTranscription.value;
+ if (!rec || !pt) return null;
+ const segs = pt.simpleSegments || [];
+ if (segs.length === 0) return null;
+ // Require time info on at least the first segment.
+ const hasTimes = segs.every(s =>
+ typeof s.startTime === 'number' &&
+ typeof s.endTime === 'number' &&
+ isFinite(s.startTime) && isFinite(s.endTime)
+ );
+ if (!hasTimes) return null;
+
+ // Per-speaker aggregation.
+ const bySpeaker = new Map(); // speakerId → { name, color, seconds, turns, words, firstStart }
+ let totalSpeakingSeconds = 0;
+ segs.forEach((seg, idx) => {
+ const dur = Math.max(0, (seg.endTime || 0) - (seg.startTime || 0));
+ const wordCount = (seg.sentence || '').trim().split(/\s+/).filter(Boolean).length;
+ const id = seg.speakerId || 'unknown';
+ let agg = bySpeaker.get(id);
+ if (!agg) {
+ agg = {
+ speakerId: id,
+ name: seg.speaker || id,
+ color: seg.color || 'speaker-color-1',
+ seconds: 0,
+ words: 0,
+ turns: 0,
+ firstStart: seg.startTime,
+ };
+ bySpeaker.set(id, agg);
+ }
+ agg.seconds += dur;
+ agg.words += wordCount;
+ // A new "turn" = previous segment had a different speaker.
+ const prev = segs[idx - 1];
+ if (!prev || prev.speakerId !== seg.speakerId) {
+ agg.turns += 1;
+ }
+ totalSpeakingSeconds += dur;
+ });
+
+ // Silence: gaps between consecutive segments + leading/trailing gap.
+ const totalAudioSeconds = Number(rec.duration) || segs[segs.length - 1].endTime || 0;
+ let silenceSeconds = 0;
+ for (let i = 0; i < segs.length - 1; i++) {
+ const gap = (segs[i + 1].startTime || 0) - (segs[i].endTime || 0);
+ if (gap > 0) silenceSeconds += gap;
+ }
+ // Leading silence before first segment.
+ if (segs[0].startTime > 0) silenceSeconds += segs[0].startTime;
+ // Trailing silence after last segment (only if duration known).
+ if (totalAudioSeconds > 0) {
+ const trail = totalAudioSeconds - segs[segs.length - 1].endTime;
+ if (trail > 0) silenceSeconds += trail;
+ }
+
+ // Build sorted speaker rows.
+ const denominator = totalAudioSeconds > 0 ? totalAudioSeconds : (totalSpeakingSeconds + silenceSeconds);
+ const speakerRows = Array.from(bySpeaker.values())
+ .sort((a, b) => b.seconds - a.seconds)
+ .map(agg => ({
+ speakerId: agg.speakerId,
+ name: agg.name,
+ color: agg.color,
+ seconds: agg.seconds,
+ durationLabel: _fmtMmSs(agg.seconds),
+ pct: denominator > 0 ? (agg.seconds / denominator) * 100 : 0,
+ turns: agg.turns,
+ words: agg.words,
+ wpm: agg.seconds > 0 ? Math.round((agg.words / agg.seconds) * 60) : 0,
+ }));
+
+ return {
+ speakerRows,
+ silence: {
+ seconds: silenceSeconds,
+ durationLabel: _fmtMmSs(silenceSeconds),
+ pct: denominator > 0 ? (silenceSeconds / denominator) * 100 : 0,
+ },
+ total: {
+ seconds: denominator,
+ durationLabel: _fmtMmSs(denominator),
+ speakingSeconds: totalSpeakingSeconds,
+ speakers: speakerRows.length,
+ turns: segs.length,
+ words: speakerRows.reduce((s, r) => s + r.words, 0),
+ },
+ };
+ });
+ const hasRecordingStats = computed(() =>
+ !!recordingStats.value && recordingStats.value.speakerRows.length > 0
+ );
+
// Aggregate `{{name}}` variables across the currently selected
// tags / folder / user / admin prompt chain. The pure helpers
// live in modules/utils/prompt-variables.js so they can be unit
@@ -1245,7 +1358,7 @@ document.addEventListener('DOMContentLoaded', async () => {
currentColorScheme, showColorSchemeModal, windowWidth, mobileTab, mobileMoreOpen, isMetadataExpanded, expandedSection,
showSortOptions, currentLanguage, currentLanguageName, availableLanguages, showLanguageMenu,
colorSchemes, isMobileScreen, isMobileDevice,
- summaryWordCount, notesWordCount,
+ summaryWordCount, notesWordCount, recordingStats, hasRecordingStats,
// Upload
uploadQueue, allJobs, currentlyProcessingFile, processingProgress, processingMessage,
diff --git a/templates/components/detail/desktop-right-panel.html b/templates/components/detail/desktop-right-panel.html
index f6057cf3..d4ed3e8d 100644
--- a/templates/components/detail/desktop-right-panel.html
+++ b/templates/components/detail/desktop-right-panel.html
@@ -22,6 +22,13 @@
<span>${ t('notes.title') }</span>
<span v-if="notesWordCount" class="tab__count">${ notesWordCount }</span>
</button>
+ <button v-if="hasRecordingStats"
+ @click="selectedTab = 'stats'"
+ class="tab"
+ :class="{'is-active': selectedTab === 'stats'}">
+ <span>Stats</span>
+ <span class="tab__count">${ recordingStats.total.speakers }</span>
+ </button>
<button v-if="selectedRecording.events && selectedRecording.events.length > 0"
@click="selectedTab = 'events'"
class="tab"
@@ -67,6 +74,7 @@
<div class="flex-1 overflow-hidden">
{% include 'components/detail/desktop-summary-tab.html' %}
{% include 'components/detail/desktop-notes-tab.html' %}
+ {% include 'components/detail/desktop-stats-tab.html' %}
{% include 'components/detail/desktop-events-tab.html' %}
</div>
</div>
diff --git a/templates/components/detail/desktop-stats-tab.html b/templates/components/detail/desktop-stats-tab.html
new file mode 100644
index 00000000..33140a51
--- /dev/null
+++ b/templates/components/detail/desktop-stats-tab.html
@@ -0,0 +1,88 @@
+<!-- Recording stats tab — per-speaker speaking time, %, turns, words and
+ WPM, plus a silence row. Only shown when the transcript has speaker
+ segments with start/end times (see hasRecordingStats computed). -->
+<div v-show="selectedTab === 'stats'" class="h-full overflow-y-auto px-5 py-4">
+ <div v-if="hasRecordingStats" class="space-y-5">
+ <!-- Headline metrics -->
+ <div class="grid grid-cols-2 md:grid-cols-4 gap-3">
+ <div class="bg-[var(--bg-secondary)] border border-[var(--border-primary)] rounded-md p-3">
+ <div class="t-label">Total length</div>
+ <div class="t-heading mt-1">${ recordingStats.total.durationLabel }</div>
+ </div>
+ <div class="bg-[var(--bg-secondary)] border border-[var(--border-primary)] rounded-md p-3">
+ <div class="t-label">Speakers</div>
+ <div class="t-heading mt-1">${ recordingStats.total.speakers }</div>
+ </div>
+ <div class="bg-[var(--bg-secondary)] border border-[var(--border-primary)] rounded-md p-3">
+ <div class="t-label">Turns</div>
+ <div class="t-heading mt-1">${ recordingStats.total.turns }</div>
+ </div>
+ <div class="bg-[var(--bg-secondary)] border border-[var(--border-primary)] rounded-md p-3">
+ <div class="t-label">Words</div>
+ <div class="t-heading mt-1">${ recordingStats.total.words.toLocaleString() }</div>
+ </div>
+ </div>
+
+ <!-- Per-speaker breakdown table -->
+ <div class="bg-[var(--bg-secondary)] border border-[var(--border-primary)] rounded-md overflow-hidden">
+ <div class="px-4 py-3 border-b border-[var(--border-primary)] flex items-center justify-between">
+ <h3 class="t-subhead">Per speaker</h3>
+ <span class="t-meta">sorted by speaking time</span>
+ </div>
+ <table class="w-full text-sm">
+ <thead>
+ <tr class="text-left t-label">
+ <th class="py-2 px-4">Speaker</th>
+ <th class="py-2 px-3 text-right">Time</th>
+ <th class="py-2 px-3 text-right">% of audio</th>
+ <th class="py-2 px-3 text-right">Turns</th>
+ <th class="py-2 px-3 text-right">Words</th>
+ <th class="py-2 px-3 text-right">WPM</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr v-for="row in recordingStats.speakerRows" :key="row.speakerId"
+ class="border-t border-[var(--border-primary)] hover:bg-[var(--bg-tertiary)] transition-colors">
+ <td class="py-2 px-4">
+ <div class="flex items-center gap-2">
+ <span :class="['speaker-tablet', row.color]" style="font-size: 11px; padding: 2px 8px;">
+ ${ row.name }
+ </span>
+ </div>
+ <!-- Slim proportion bar -->
+ <div class="mt-1.5 h-1.5 w-full rounded-full bg-[var(--bg-tertiary)] overflow-hidden">
+ <div class="h-full" :style="{ width: row.pct.toFixed(1) + '%', background: 'var(--text-accent)' }"></div>
+ </div>
+ </td>
+ <td class="py-2 px-3 font-mono text-right">${ row.durationLabel }</td>
+ <td class="py-2 px-3 text-right">${ row.pct.toFixed(1) }%</td>
+ <td class="py-2 px-3 text-right">${ row.turns }</td>
+ <td class="py-2 px-3 text-right">${ row.words.toLocaleString() }</td>
+ <td class="py-2 px-3 text-right">${ row.wpm || '—' }</td>
+ </tr>
+ <!-- Silence row -->
+ <tr v-if="recordingStats.silence.seconds > 0"
+ class="border-t border-[var(--border-primary)] bg-[var(--bg-tertiary)]/50">
+ <td class="py-2 px-4">
+ <span class="t-meta italic">Silence</span>
+ <div class="mt-1.5 h-1.5 w-full rounded-full bg-[var(--bg-tertiary)] overflow-hidden">
+ <div class="h-full" :style="{ width: recordingStats.silence.pct.toFixed(1) + '%', background: 'var(--text-muted)' }"></div>
+ </div>
+ </td>
+ <td class="py-2 px-3 font-mono text-right">${ recordingStats.silence.durationLabel }</td>
+ <td class="py-2 px-3 text-right">${ recordingStats.silence.pct.toFixed(1) }%</td>
+ <td class="py-2 px-3 text-right text-[var(--text-muted)]">—</td>
+ <td class="py-2 px-3 text-right text-[var(--text-muted)]">—</td>
+ <td class="py-2 px-3 text-right text-[var(--text-muted)]">—</td>
+ </tr>
+ </tbody>
+ </table>
+ </div>
+ </div>
+
+ <div v-else class="empty">
+ <div class="empty__icon"><i class="fas fa-chart-bar"></i></div>
+ <div class="empty__title">No stats yet</div>
+ <div class="empty__body">Stats appear once this recording has been transcribed with speaker diarization and segment timestamps.</div>
+ </div>
+</div>