Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
/**
* Pipeline — Voice-Fingerprint Matrix (#2194, CWQE Phase 2 follow-up).
*
* A dedicated, deep-linkable surface for the deterministic voice-drift engine:
* a monospace issues×metrics matrix where every drafted issue's full fingerprint
* vector (sentence rhythm, fragment/long-sentence rates, paragraph shape,
* dialogue ratio, em-dash rate, abstract-noun/simile density, dominant opener,
* plus any configured vocabulary wells) is a row, and each statistical outlier
* cell (> the series drift threshold in σ) is highlighted. The finding cards on
* the Editorial Checks page only surface the FLAGGED outliers as prose; this view
* shows the whole matrix + the series mean/σ footer so the drift is legible at a
* glance. Read-only — no LLM cost.
*
* The series id is the route param (`/pipeline/series/:seriesId/voice-fingerprint`),
* so the view is shareable/bookmarkable per the "URL is the source of truth"
* convention. A stale/deleted id falls back to the pipeline index.
*/
import { useEffect, useState } from 'react';
import { Link, useParams, useNavigate } from 'react-router';
import { ArrowLeft, Fingerprint, BookOpen, Info } from 'lucide-react';
import toast from '../components/ui/Toast';
import PageSkeleton from '../components/ui/PageSkeleton';
import { getPipelineSeries, getVoiceFingerprint } from '../services/api';
// A cell is an outlier when the (issue, metricKey) pair is in the drift set.
// Keyed on `issue:metricKey` so a lookup is O(1) while rendering the grid.
const outlierKey = (issue, metricKey) => `${issue}:${metricKey}`;
// Round a metric value to 2dp for display (the matrix cells + the seriesFindings
// banner, which carries the raw mean/center).
const round2 = (n) => Math.round((Number(n) || 0) * 100) / 100;
export default function PipelineVoiceFingerprint() {
const { seriesId } = useParams();
const navigate = useNavigate();
const [series, setSeries] = useState(null);
const [data, setData] = useState(null);
const [loading, setLoading] = useState(true);
useEffect(() => {
let canceled = false;
setLoading(true);
Promise.all([
getPipelineSeries(seriesId, { silent: true }),
getVoiceFingerprint(seriesId, { silent: true }),
])
.then(([s, fp]) => {
if (canceled) return;
setSeries(s);
setData(fp);
})
.catch((err) => {
if (canceled) return;
toast.error(err.message || 'Failed to load voice fingerprint');
navigate('/pipeline');
})
.finally(() => { if (!canceled) setLoading(false); });
return () => { canceled = true; };
}, [seriesId, navigate]);
if (loading) {
return (
<PageSkeleton
label="Loading voice fingerprint"
fullHeight
padded
headerRowClass="flex flex-wrap items-center gap-2"
titleWidthClass="w-56"
showAction={false}
cards={3}
sidebar={false}
/>
);
}
const columns = data?.columns || [];
const issues = data?.matrix?.issues || [];
const seriesStats = data?.series || {};
const outlierSet = new Set((data?.outliers || []).map((o) => outlierKey(o.issue, o.metricKey)));
// #2248 — series-wide "the whole corpus is uniformly off the chosen voice"
// findings. These have no single cell to highlight (the column is uniform), so
// they surface as a banner above the matrix rather than an amber cell.
const seriesFindings = data?.seriesFindings || [];
const hasMatrix = issues.length > 0;
const baselineMode = data?.baselineMode || 'drafted';
// Only claim the chosen-voice baseline once it was actually applied: a gatedOff
// run (< minIssues) still reports the configured mode but computes no comparison
// and returns `series: {}`, so the footer row + intro copy must fall back to the
// plain description rather than promising a baseline that never ran.
const usesChosenVoice = data?.exemplarBaselineUsed === true && !data?.gatedOff;
// The baseline phrase for the intro copy — blended is the midpoint, NOT the
// exemplars alone, so it must not be described as "the chosen voice."
const baselineNoun = !usesChosenVoice
? 'series mean'
: (baselineMode === 'blended'
? "blend of the series mean and the style guide's chosen voice"
: "style guide's chosen voice (its voice exemplars)");
return (
<div className="h-full overflow-y-auto p-4 md:p-6">
<header className="flex flex-wrap items-center gap-2 mb-4">
<Link
to={`/pipeline/series/${seriesId}`}
className="inline-flex items-center gap-1 px-2 py-1 rounded text-xs text-gray-400 hover:text-white border border-port-border bg-port-card"
title="Back to series"
>
<ArrowLeft size={12} /> Series
</Link>
<h1 className="text-lg font-semibold text-white flex items-center gap-2">
<Fingerprint size={18} className="text-port-accent" /> Voice Fingerprint
</h1>
{series?.name ? <span className="text-sm text-gray-400 truncate">— {series.name}</span> : null}
<Link
to="/pipeline/editorial-checks"
className="ml-auto inline-flex items-center gap-1 px-2 py-1 rounded text-xs text-gray-400 hover:text-white border border-port-border bg-port-card"
title="Configure the style.voice-drift check (threshold, min issues, vocabulary wells)"
>
Editorial Checks
</Link>
</header>
<p className="text-xs text-gray-500 mb-4 max-w-3xl">
Every drafted issue's prose fingerprint. Each column is a metric; a cell
highlighted in <span className="text-port-warning">amber</span> is a
statistical outlier — more than {data?.threshold ?? 1.5}σ from the{' '}
{baselineNoun}{' '}
on that metric. The bottom rows are the series mean and standard
deviation (σ){usesChosenVoice ? `, plus the ${baselineMode === 'blended' ? 'blended' : 'chosen-voice'} baseline the outliers are measured against` : ''}.
This measures the same vectors the deterministic{' '}
<code className="text-gray-400">style.voice-drift</code> editorial check
flags; it just shows the whole matrix, not only the flagged outliers.
</p>
{!hasMatrix ? (
<EmptyState gatedOff={data?.gatedOff} issueCount={data?.issueCount ?? 0} minIssues={data?.config?.minIssues} />
) : (
<>
{data?.gatedOff ? (
<div className="mb-4 flex items-start gap-2 px-3 py-2 rounded-lg border border-port-border bg-port-card text-gray-400 text-sm">
<Info size={14} className="mt-0.5 shrink-0" />
<span>
Drift detection is off — only {data.issueCount} issue
{data.issueCount === 1 ? '' : 's'} drafted (minimum{' '}
{data?.config?.minIssues ?? 4}). The fingerprint matrix is shown for
reference, but no outliers are flagged until more issues exist.
</span>
</div>
) : null}
{seriesFindings.length ? (
<div className="mb-4 rounded-lg border border-port-warning/40 bg-port-warning/10 px-3 py-2 text-sm">
<div className="flex items-center gap-2 text-port-warning font-medium mb-1">
<Info size={14} className="shrink-0" />
{seriesFindings.length} series-wide voice mismatch
{seriesFindings.length === 1 ? '' : 'es'}
</div>
<ul className="space-y-1 text-gray-300">
{seriesFindings.map((f) => (
<li key={f.metricKey}>
The whole series sits at {round2(f.mean)}{f.unit || ''} on{' '}
<span className="text-white">{f.label}</span> vs the chosen voice's{' '}
{round2(f.center)}{f.unit || ''} — uniformly {f.direction === 'high' ? 'above' : 'below'} the
chosen voice (no single issue is an outlier).
</li>
))}
</ul>
</div>
) : null}
<FingerprintMatrix
columns={columns}
issues={issues}
seriesStats={seriesStats}
outlierSet={outlierSet}
baselineMode={baselineMode}
usesChosenVoice={usesChosenVoice}
/>
<p className="text-xs text-gray-500 mt-3">
{issues.length} issue{issues.length === 1 ? '' : 's'} · {columns.length} metrics
{data?.wells?.length ? ` · vocabulary wells: ${data.wells.join(', ')}` : ''}
{data?.outliers?.length
? ` · ${data.outliers.length} outlier${data.outliers.length === 1 ? '' : 's'} flagged`
: ''}
</p>
</>
)}
</div>
);
}
function EmptyState({ gatedOff, issueCount, minIssues }) {
if (gatedOff && issueCount > 0) {
return (
<div className="text-center text-gray-400 py-16">
<Fingerprint className="mx-auto mb-3 opacity-40" size={28} />
Only {issueCount} issue{issueCount === 1 ? '' : 's'} drafted — drift needs at
least {minIssues ?? 4} to compute a stable series voice. Draft more issues,
then return to see the fingerprint matrix.
</div>
);
}
return (
<div className="text-center text-gray-400 py-16">
<BookOpen className="mx-auto mb-3 opacity-40" size={28} />
Nothing is drafted yet. Write or import a manuscript, then return to see each
issue's voice fingerprint.
</div>
);
}
function FingerprintMatrix({ columns, issues, seriesStats, outlierSet, baselineMode, usesChosenVoice }) {
// The baseline an outlier tooltip / footer row names: the chosen voice (or a
// blend) when the exemplar baseline is active (#2179), else the series mean.
const centerNoun = usesChosenVoice
? (baselineMode === 'blended' ? 'blend of series + chosen voice' : 'chosen voice')
: 'series mean';
const centerOf = (col) => (usesChosenVoice
? (seriesStats?.[col.key]?.center ?? seriesStats?.[col.key]?.mean)
: seriesStats?.[col.key]?.mean);
// Below minIssues the server gates drift off and returns `series: {}`, so a
// mean/σ footer would render misleading 0s — only show it once stats exist.
const hasStats = seriesStats && Object.keys(seriesStats).length > 0;
return (
<div className="overflow-x-auto border border-port-border rounded-lg">
<table className="border-collapse font-mono text-xs w-full">
<thead>
<tr className="bg-port-card">
<th className="sticky left-0 z-10 bg-port-card text-left text-gray-400 font-medium px-3 py-2 min-w-[4rem]">
Issue
</th>
{columns.map((col) => (
<th
key={col.key}
className="px-2 py-2 text-right text-gray-400 font-medium whitespace-nowrap align-bottom"
title={`${col.label} — high = ${col.higher}; low = ${col.lower}`}
>
{col.label}
</th>
))}
</tr>
</thead>
<tbody>
{issues.map((it) => (
<tr key={it.issue} className="border-t border-port-border">
<td className="sticky left-0 z-10 bg-port-bg text-gray-300 font-medium px-3 py-1.5">
#{it.issue}
</td>
{columns.map((col) => {
const v = it.metrics?.[col.key] ?? 0;
const isOutlier = outlierSet.has(outlierKey(it.issue, col.key));
return (
<td
key={col.key}
className={`px-2 py-1.5 text-right tabular-nums ${
isOutlier
? 'bg-port-warning/20 text-port-warning font-semibold'
: 'text-gray-300'
}`}
title={isOutlier ? `Outlier — ${centerNoun} ${round2(centerOf(col))}${col.unit || ''}` : undefined}
>
{round2(v)}{col.unit || ''}
</td>
);
})}
</tr>
))}
</tbody>
{hasStats ? (
<tfoot>
<tr className="border-t-2 border-port-border bg-port-card">
<td className="sticky left-0 z-10 bg-port-card text-gray-500 px-3 py-1.5">mean</td>
{columns.map((col) => (
<td key={col.key} className="px-2 py-1.5 text-right text-gray-500 tabular-nums">
{round2(seriesStats?.[col.key]?.mean)}{col.unit || ''}
</td>
))}
</tr>
<tr className="bg-port-card">
<td className="sticky left-0 z-10 bg-port-card text-gray-500 px-3 py-1.5">σ</td>
{columns.map((col) => (
<td key={col.key} className="px-2 py-1.5 text-right text-gray-500 tabular-nums">
{round2(seriesStats?.[col.key]?.std)}
</td>
))}
</tr>
{usesChosenVoice ? (
<tr className="bg-port-card">
<td
className="sticky left-0 z-10 bg-port-card text-port-accent px-3 py-1.5"
title={`Outliers are measured against this ${centerNoun}, not the drafted mean.`}
>
{baselineMode === 'blended' ? 'blend' : 'voice'}
</td>
{columns.map((col) => (
<td key={col.key} className="px-2 py-1.5 text-right text-port-accent tabular-nums">
{round2(centerOf(col))}{col.unit || ''}
</td>
))}
</tr>
) : null}
</tfoot>
) : null}
</table>
</div>
);
}