| import { Activity, AlertTriangle, CheckCircle, Loader2, RefreshCw, TrendingUp } from 'lucide-react'; |
| import { useCallback, useEffect, useState } from 'react'; |
|
|
| interface ForecastInterval { |
| horizon: string; |
| point: number; |
| lower80: number; |
| upper80: number; |
| lower95?: number; |
| upper95?: number; |
| confidence: number; |
| unit?: string; |
| } |
|
|
| interface ForecastHead { |
| headName: string; |
| label: string; |
| description?: string; |
| intervals: ForecastInterval[]; |
| featureAttribution?: Record<string, number>; |
| provenance: { |
| modelId: string; |
| modelName: string; |
| modelVersion: string; |
| adapterId: string; |
| algorithmFamily?: string; |
| calibrationMethod?: string; |
| generatedAt: string; |
| }; |
| alertThreshold?: number; |
| thresholdBreached?: boolean; |
| driftScore?: number; |
| driftStatus?: string; |
| tceDistribution?: { |
| mean: number; |
| p5: number; |
| p25: number; |
| p75: number; |
| p95: number; |
| stdDev: number; |
| breakEvenProbability: number; |
| }; |
| } |
|
|
| interface ForecastApiResponse { |
| heads: ForecastHead[]; |
| headsCount: number; |
| breachedCount: number; |
| generatedAt: string; |
| modelRegistry: string; |
| calibrationMethod: string; |
| note: string; |
| } |
|
|
| function IntervalBar({ iv, threshold, unit }: { iv: ForecastInterval; threshold?: number; unit?: string }) { |
| const isUsd = unit === 'USD/day'; |
| const scale = isUsd ? 1 / 50000 : 1; |
| const pointPct = Math.min(Math.max(iv.point * scale * 100, 0), 100); |
| const lowerPct = Math.min(Math.max((iv.lower80 ?? iv.point * 0.7) * scale * 100, 0), 100); |
| const upperPct = Math.min(Math.max((iv.upper80 ?? iv.point * 1.3) * scale * 100, 0), 100); |
| const thresholdPct = threshold ? Math.min(Math.max(threshold * scale * 100, 0), 100) : undefined; |
| const breached = threshold !== undefined && (isUsd ? iv.point < threshold : iv.upper80 > threshold); |
|
|
| const fmt = (v: number) => |
| isUsd ? `$${v.toLocaleString()}/d` : `${(v * 100).toFixed(1)}%`; |
|
|
| return ( |
| <div className="space-y-1"> |
| <div className="flex items-center justify-between text-[10px] font-mono"> |
| <span className="text-muted-foreground">{iv.horizon}</span> |
| <span className={breached ? 'text-red-400' : 'text-foreground'}> |
| {fmt(iv.point)}{' '} |
| <span className="text-muted-foreground">[{fmt(iv.lower80 ?? iv.point * 0.7)}–{fmt(iv.upper80 ?? iv.point * 1.3)}]</span> |
| </span> |
| </div> |
| <div className="relative h-2 bg-muted rounded-full overflow-hidden"> |
| <div |
| className="absolute h-full bg-primary/20 rounded-full" |
| style={{ left: `${lowerPct}%`, width: `${Math.max(0, upperPct - lowerPct)}%` }} |
| /> |
| <div |
| className={`absolute h-full w-0.5 rounded-full ${breached ? 'bg-red-400' : 'bg-primary'}`} |
| style={{ left: `${pointPct}%` }} |
| /> |
| {thresholdPct !== undefined && ( |
| <div className="absolute h-full w-px bg-amber-400/60" style={{ left: `${thresholdPct}%` }} /> |
| )} |
| </div> |
| <div className="text-[9px] text-muted-foreground font-mono">conf: {(iv.confidence * 100).toFixed(0)}%</div> |
| </div> |
| ); |
| } |
|
|
| function FeatureAttribution({ attrs }: { attrs: Record<string, number> }) { |
| const sorted = Object.entries(attrs).sort((a, b) => b[1] - a[1]).slice(0, 5); |
| return ( |
| <div className="space-y-1 pt-1"> |
| <p className="text-[9px] font-mono text-muted-foreground uppercase tracking-wider">Feature Attribution</p> |
| {sorted.map(([key, val]) => ( |
| <div key={key} className="flex items-center gap-2"> |
| <div className="flex-1 h-1 bg-muted rounded-full overflow-hidden"> |
| <div className="h-full bg-sky-400/60 rounded-full" style={{ width: `${val * 100}%` }} /> |
| </div> |
| <span className="text-[9px] font-mono text-muted-foreground w-8 text-right">{(val * 100).toFixed(0)}%</span> |
| <span className="text-[9px] font-mono text-muted-foreground w-28 truncate">{key.replace(/_/g, ' ')}</span> |
| </div> |
| ))} |
| </div> |
| ); |
| } |
|
|
| function TceCard({ dist }: { dist: NonNullable<ForecastHead['tceDistribution']> }) { |
| return ( |
| <div className="pt-2 space-y-2 border-t border-border"> |
| <p className="text-[9px] font-mono text-muted-foreground uppercase tracking-wider">TCE Distribution (Monte Carlo · 5k iter)</p> |
| <div className="grid grid-cols-5 gap-1"> |
| {[ |
| { label: 'P5', val: dist.p5 }, |
| { label: 'P25', val: dist.p25 }, |
| { label: 'P50', val: dist.mean }, |
| { label: 'P75', val: dist.p75 }, |
| { label: 'P95', val: dist.p95 }, |
| ].map(({ label, val }) => ( |
| <div key={label} className="text-center"> |
| <p className="text-[8px] font-mono text-muted-foreground">{label}</p> |
| <p className="text-[10px] font-mono text-foreground">${(val / 1000).toFixed(1)}k</p> |
| </div> |
| ))} |
| </div> |
| <div className="flex items-center gap-2 text-[9px] font-mono text-muted-foreground"> |
| <span>Break-even prob:</span> |
| <span className={dist.breakEvenProbability >= 0.75 ? 'text-emerald-400' : 'text-amber-400'}> |
| {(dist.breakEvenProbability * 100).toFixed(0)}% |
| </span> |
| <span>· σ = ${(dist.stdDev / 1000).toFixed(1)}k/d</span> |
| </div> |
| </div> |
| ); |
| } |
|
|
| function HeadCard({ head }: { head: ForecastHead }) { |
| const [expanded, setExpanded] = useState(false); |
| const latest = head.intervals[head.intervals.length - 1]; |
| const isUsd = latest?.unit === 'USD/day'; |
|
|
| return ( |
| <div |
| className={`rounded-lg border p-4 space-y-3 cursor-pointer transition-all ${ |
| head.thresholdBreached |
| ? 'border-red-500/40 bg-red-500/5' |
| : 'border-border bg-card hover:bg-accent/30' |
| }`} |
| onClick={() => setExpanded((v) => !v)} |
| > |
| <div className="flex items-start justify-between gap-2"> |
| <div className="space-y-0.5"> |
| <p className="text-[10px] font-mono text-muted-foreground uppercase tracking-wider">{head.headName}</p> |
| <p className="text-sm font-medium text-foreground leading-tight">{head.label}</p> |
| {head.description && ( |
| <p className="text-[10px] text-muted-foreground leading-relaxed hidden sm:block">{head.description}</p> |
| )} |
| </div> |
| {head.thresholdBreached ? ( |
| <AlertTriangle className="w-4 h-4 text-red-400 shrink-0 mt-0.5" /> |
| ) : ( |
| <CheckCircle className="w-4 h-4 text-emerald-400 shrink-0 mt-0.5" /> |
| )} |
| </div> |
| |
| {latest && ( |
| <IntervalBar iv={latest} threshold={head.alertThreshold} unit={latest.unit} /> |
| )} |
| |
| {expanded && head.intervals.slice(0, -1).map((iv) => ( |
| <IntervalBar key={iv.horizon} iv={iv} threshold={head.alertThreshold} unit={iv.unit} /> |
| ))} |
| |
| {expanded && head.tceDistribution && <TceCard dist={head.tceDistribution} />} |
| |
| {expanded && head.featureAttribution && ( |
| <FeatureAttribution attrs={head.featureAttribution} /> |
| )} |
| |
| {expanded && ( |
| <div className="pt-1 border-t border-border space-y-1"> |
| <p className="text-[9px] font-mono text-muted-foreground">model: {head.provenance.modelName} v{head.provenance.modelVersion}</p> |
| <p className="text-[9px] font-mono text-muted-foreground">algorithm: {head.provenance.algorithmFamily ?? head.provenance.adapterId}</p> |
| <p className="text-[9px] font-mono text-muted-foreground">calibration: {head.provenance.calibrationMethod ?? 'monte_carlo'}</p> |
| {head.driftScore !== undefined && ( |
| <p className="text-[9px] font-mono text-muted-foreground"> |
| drift PSI: {head.driftScore.toFixed(4)} · status:{' '} |
| <span className={head.driftStatus === 'nominal' ? 'text-emerald-400' : 'text-amber-400'}> |
| {head.driftStatus ?? 'unknown'} |
| </span> |
| </p> |
| )} |
| <p className="text-[9px] font-mono text-muted-foreground">generated: {new Date(head.provenance.generatedAt).toLocaleString()}</p> |
| </div> |
| )} |
| </div> |
| ); |
| } |
|
|
| export function ForecastPanel() { |
| const [data, setData] = useState<ForecastApiResponse | null>(null); |
| const [loading, setLoading] = useState(false); |
| const [error, setError] = useState<string | null>(null); |
| const [lastFetchedAt, setLastFetchedAt] = useState<string | null>(null); |
|
|
| const fetchForecasts = useCallback(async () => { |
| setLoading(true); |
| setError(null); |
| try { |
| const res = await fetch('/api/vessels/forecasts/heads', { credentials: 'include' }); |
| if (!res.ok) throw new Error(`HTTP ${res.status}`); |
| const json = await res.json(); |
| setData(json.data ?? json); |
| setLastFetchedAt(new Date().toISOString()); |
| } catch (e) { |
| setError(e instanceof Error ? e.message : 'Failed to load forecasts'); |
| } finally { |
| setLoading(false); |
| } |
| }, []); |
|
|
| useEffect(() => { |
| fetchForecasts(); |
| }, [fetchForecasts]); |
|
|
| const heads = data?.heads ?? []; |
| const breachedCount = data?.breachedCount ?? heads.filter((h) => h.thresholdBreached).length; |
|
|
| return ( |
| <div className="space-y-4"> |
| <div className="flex items-center justify-between"> |
| <div className="flex items-center gap-2"> |
| <TrendingUp className="w-4 h-4 text-primary" /> |
| <h2 className="text-sm font-semibold text-foreground">Forecast Fabric — Vessels Heads</h2> |
| </div> |
| <div className="flex items-center gap-2"> |
| {breachedCount > 0 && ( |
| <span className="inline-flex items-center gap-1 text-[10px] font-mono bg-red-500/10 text-red-400 border border-red-500/30 rounded px-2 py-0.5"> |
| <Activity className="w-3 h-3" /> |
| {breachedCount} threshold{breachedCount > 1 ? 's' : ''} breached |
| </span> |
| )} |
| <button |
| onClick={fetchForecasts} |
| disabled={loading} |
| className="inline-flex items-center gap-1 text-[10px] font-mono text-muted-foreground hover:text-foreground transition-colors disabled:opacity-50" |
| > |
| {loading ? <Loader2 className="w-3 h-3 animate-spin" /> : <RefreshCw className="w-3 h-3" />} |
| {loading ? 'Refreshing…' : 'Refresh'} |
| </button> |
| {data && ( |
| <span className="text-[10px] font-mono text-muted-foreground"> |
| {heads.length} heads · {data.modelRegistry} |
| </span> |
| )} |
| </div> |
| </div> |
| |
| <p className="text-xs text-muted-foreground"> |
| Calibrated ML forecast intervals for all Vessels maritime-intelligence heads. Monte Carlo calibration (≥2k iterations). Click any card to expand horizons, feature attribution, and drift report. |
| </p> |
| |
| {error && ( |
| <div className="rounded-lg border border-red-500/30 bg-red-500/5 p-3 text-xs text-red-400"> |
| Failed to load live forecasts: {error} |
| </div> |
| )} |
| |
| {loading && heads.length === 0 && ( |
| <div className="flex items-center justify-center py-8 gap-2 text-muted-foreground text-sm"> |
| <Loader2 className="w-4 h-4 animate-spin" /> |
| Loading forecast heads… |
| </div> |
| )} |
| |
| {heads.length > 0 && ( |
| <div className="grid grid-cols-1 sm:grid-cols-2 gap-3"> |
| {heads.map((head) => <HeadCard key={head.headName} head={head} />)} |
| </div> |
| )} |
| |
| <div className="space-y-1"> |
| <p className="text-[10px] font-mono text-muted-foreground"> |
| source: vessels-ml-v2 · calibration: {data?.calibrationMethod ?? 'monte_carlo_isotonic'} · interval bars show [P10–P90] · amber line = alert threshold |
| </p> |
| {lastFetchedAt && ( |
| <p className="text-[9px] font-mono text-muted-foreground"> |
| last fetched: {new Date(lastFetchedAt).toLocaleTimeString()} · auto-refreshes every 5 min |
| </p> |
| )} |
| {data?.note && ( |
| <p className="text-[9px] font-mono text-muted-foreground">{data.note}</p> |
| )} |
| </div> |
| </div> |
| ); |
| } |
|
|