Sentiment Analyzer Block

'use client';

/**
 * @file sentiment-analyzer.tsx
 * @description Sentiment Analyzer block — DistilBERT SST-2 sentiment for one or many texts (one per line, or a .txt/.csv first column) with streaming results, determinate throughput, aggregate stats, and a 100-row windowed list; model download gated behind Run.
 */
import { useEffect, useRef, useState } from 'react';
import { FileUp, Play, Sparkles, Square, Trash2 } from 'lucide-react';
import {
  useModelLoad,
  useSequentialBatch,
  toAppError,
  type UseModelLoadReturn,
} from '@localmode/react';
import { classify, type ClassificationModel } from '@localmode/core';
import { transformers, isModelCached } from '@localmode/transformers';

import { ConfidenceScoreBadge } from '@/components/confidence-score-badge';
import { EntityStatsBar } from '@/components/entity-stats-bar';
import { ModelLoadingPanel } from '@/components/model-loading-panel';
import { CacheBadge } from '@/components/cache-badge';
import { ErrorAlert } from '@/components/error-alert';
import { ModeErrorBoundary } from '@/components/mode-error-boundary';
import { cn } from '@/lib/utils';

const SENTIMENT_MODEL_ID = 'Xenova/distilbert-base-uncased-finetuned-sst-2-english';
const SENTIMENT_MODEL_NAME = 'DistilBERT Sentiment';
const SENTIMENT_MODEL_SIZE = '67 MB';

const SAMPLE_REVIEWS = [
  'This product is amazing! Best purchase I ever made.',
  'Terrible experience. The item broke after one day.',
  'Pretty average product, nothing special about it.',
  'I love how easy this is to use. Highly recommend!',
  'Waste of money. Customer support was unhelpful too.',
  'Great quality and fast shipping. Will buy again.',
];

const WINDOW_CAP = 100;

function formatScore(score: number) {
  return `${(score * 100).toFixed(1)}%`;
}

function formatDuration(ms: number) {
  if (ms < 1000) return `${Math.round(ms)}ms`;
  return `${(ms / 1000).toFixed(1)}s`;
}

function parseItems(raw: string, options?: { csv?: boolean }): string[] {
  return raw
    .split('\n')
    .map((line) => {
      const trimmed = line.trim();
      if (!trimmed) return '';
      return options?.csv && trimmed.includes(',') ? trimmed.split(',')[0].trim() : trimmed;
    })
    .filter((line) => line.length > 0);
}

interface SentimentItem {
  text: string;
  label: string;
  score: number;
  status: 'ok' | 'error';
}

function SentimentLabel({ label }: { label: string }) {
  const positive = label.toUpperCase() === 'POSITIVE';
  return (
    <span
      className={cn(
        'inline-flex items-center rounded-full px-2 py-0.5 text-[10px] font-semibold uppercase tracking-wide',
        positive
          ? 'bg-emerald-500/15 text-emerald-700 dark:text-emerald-400'
          : 'bg-rose-500/15 text-rose-700 dark:text-rose-400',
      )}
    >
      {label}
    </span>
  );
}

export function SentimentAnalyzerBlock() {
  return (
    <div className="mx-auto flex max-w-4xl flex-col gap-4 p-4">
      <p className="text-xs text-muted-foreground">
        Sentiment Analyzer: DistilBERT SST-2. Model loads only behind an explicit action.
      </p>
      <ModeErrorBoundary>
        <AnalyzeInner />
      </ModeErrorBoundary>
    </div>
  );
}

function AnalyzeInner() {
  const load = useModelLoad<ClassificationModel>({
    key: `text-insights-sentiment:${SENTIMENT_MODEL_ID}`,
    create: (onProgress) =>
      transformers.classifier(SENTIMENT_MODEL_ID, {
        onProgress: (p) => onProgress(p as Parameters<typeof onProgress>[0]),
      }),
    warmup: (model) => classify({ model, text: 'ready' }),
    isCached: () => isModelCached(SENTIMENT_MODEL_ID),
  });

  const model = load.model;
  if (!model) return <p className="text-sm text-muted-foreground">Preparing…</p>;
  return <AnalyzeSurface load={load} model={model} />;
}

function AnalyzeSurface({
  load,
  model,
}: {
  load: UseModelLoadReturn<ClassificationModel>;
  model: ClassificationModel;
}) {
  const [input, setInput] = useState('');
  const [fileName, setFileName] = useState<string | null>(null);
  const [isCsv, setIsCsv] = useState(false);
  const [runItems, setRunItems] = useState<string[]>([]);
  const [elapsedMs, setElapsedMs] = useState(0);
  const startRef = useRef(0);
  const fileInputRef = useRef<HTMLInputElement>(null);

  const batch = useSequentialBatch<string, { label: string; score: number }>({
    fn: async (text, signal) => {
      const r = await classify({ model, text, abortSignal: signal });
      return { label: r.label, score: r.score };
    },
  });

  useEffect(() => {
    if (!batch.isRunning) return;
    const id = window.setInterval(() => setElapsedMs(Date.now() - startRef.current), 200);
    return () => window.clearInterval(id);
  }, [batch.isRunning]);

  const items = parseItems(input, { csv: isCsv });
  const lineCount = items.length;
  const charCount = input.length;
  const appErr = toAppError(batch.error) ?? (load.error ? toAppError(load.error) : null);

  const results: SentimentItem[] = [];
  for (let i = 0; i < batch.results.length; i++) {
    const r = batch.results[i];
    const text = runItems[i] ?? '';
    if (r !== null) {
      results.push({ text, label: r.label, score: r.score, status: 'ok' });
    } else if (batch.itemErrors[i]) {
      results.push({ text, label: 'ERROR', score: 0, status: 'error' });
    }
  }

  const ok = results.filter((r) => r.status === 'ok');
  const positive = ok.filter((r) => r.label.toUpperCase() === 'POSITIVE').length;
  const negative = ok.filter((r) => r.label.toUpperCase() === 'NEGATIVE').length;
  const total = ok.length;
  const avgScore = total > 0 ? ok.reduce((s, r) => s + r.score, 0) / total : 0;

  const done = batch.progress.current;
  const totalItems = batch.progress.total;
  const elapsedSec = elapsedMs / 1000;
  const rate = elapsedSec > 0 ? done / elapsedSec : 0;
  const etaSec = rate > 0 ? (totalItems - done) / rate : 0;

  const run = async () => {
    if (items.length === 0 || batch.isRunning) return;
    setRunItems(items);
    setElapsedMs(0);
    startRef.current = Date.now();
    try {
      await load.load();
    } catch {
      return;
    }
    await batch.execute(items);
    setElapsedMs(Date.now() - startRef.current);
  };

  const loadSample = () => {
    setInput(SAMPLE_REVIEWS.join('\n'));
    setFileName(null);
    setIsCsv(false);
  };

  const clear = () => {
    setInput('');
    setFileName(null);
    setIsCsv(false);
    setRunItems([]);
    batch.reset();
  };

  const onFile = async (file: File | undefined) => {
    if (!file) return;
    const text = await file.text();
    setInput(text);
    setFileName(file.name);
    setIsCsv(/\.csv$/i.test(file.name));
  };

  const onKeyDown = (e: React.KeyboardEvent) => {
    if ((e.metaKey || e.ctrlKey) && e.key === 'Enter') {
      e.preventDefault();
      void run();
    }
  };

  const visible = results.length > WINDOW_CAP ? results.slice(0, WINDOW_CAP) : results;
  const windowed = results.length > WINDOW_CAP;

  return (
    <div className="flex flex-col gap-4">
      <div className="flex flex-wrap items-center gap-2 text-xs text-muted-foreground">
        <span>
          {SENTIMENT_MODEL_NAME} · {SENTIMENT_MODEL_SIZE}
        </span>
        {load.cached === true && (
          <span>
            <CacheBadge cached latencyMs={undefined} label="model cached" />
          </span>
        )}
      </div>

      {}
      <div>
        <textarea
          aria-label="Texts to analyze, one per line"
          value={input}
          onChange={(e) => {
            setInput(e.target.value);
            setIsCsv(false);
          }}
          onKeyDown={onKeyDown}
          placeholder="Enter one text per line: reviews, comments, messages…"
          rows={6}
          className="w-full resize-y rounded-md border border-border bg-background p-2.5 text-sm outline-none focus-visible:ring-2 focus-visible:ring-ring"
        />
        <div className="mt-1 flex flex-wrap items-center justify-between gap-2">
          <span className="text-xs text-muted-foreground">
            {lineCount} {lineCount === 1 ? 'line' : 'lines'} · {charCount} chars
            {fileName && <span className="ml-1">· {fileName}</span>}
          </span>
          <div className="flex items-center gap-1.5">
            <button
              type="button"
              onClick={loadSample}
              className="inline-flex h-7 items-center gap-1 rounded-md border border-border px-2 text-xs font-medium hover:bg-muted focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
            >
              <Sparkles className="h-3 w-3" aria-hidden /> Load samples
            </button>
            <button
              type="button"
              onClick={() => fileInputRef.current?.click()}
              className="inline-flex h-7 items-center gap-1 rounded-md border border-border px-2 text-xs font-medium hover:bg-muted focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
            >
              <FileUp className="h-3 w-3" aria-hidden /> Upload
            </button>
            <input
              ref={fileInputRef}
              type="file"
              aria-label="Upload a .txt or .csv file"
              accept=".txt,.csv,text/plain,text/csv"
              className="hidden"
              onChange={(e) => void onFile(e.target.files?.[0])}
            />
          </div>
        </div>
      </div>

      {}
      <div className="flex items-center gap-2">
        <button
          type="button"
          onClick={() => void run()}
          disabled={items.length === 0 || batch.isRunning}
          className="inline-flex h-8 items-center gap-1.5 rounded-md bg-primary px-3 text-sm font-medium text-primary-foreground transition-colors hover:bg-primary/90 focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring disabled:opacity-50"
        >
          <Play className="h-3.5 w-3.5" aria-hidden /> Analyze
        </button>
        {batch.isRunning && (
          <button
            type="button"
            onClick={batch.cancel}
            className="inline-flex h-8 items-center gap-1.5 rounded-md border border-border px-3 text-sm font-medium hover:bg-muted focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
          >
            <Square className="h-3.5 w-3.5" aria-hidden /> Stop
          </button>
        )}
        {results.length > 0 && !batch.isRunning && (
          <button
            type="button"
            onClick={clear}
            className="inline-flex h-8 items-center gap-1.5 rounded-md border border-border px-3 text-sm font-medium hover:bg-muted focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
          >
            <Trash2 className="h-3.5 w-3.5" aria-hidden /> Clear
          </button>
        )}
      </div>

      {}
      {load.status === 'loading' && (
        <div>
          <ModelLoadingPanel
            name={SENTIMENT_MODEL_NAME}
            size={SENTIMENT_MODEL_SIZE}
            progress={load.progressValue}
            cached={load.cached === true}
          />
        </div>
      )}

      {appErr && (
        <span>
          <ErrorAlert message={appErr.message} onRetry={() => void run()} onDismiss={batch.reset} />
        </span>
      )}

      {}
      {(batch.isRunning || totalItems > 0) && (
        <div className="flex flex-col gap-2 rounded-md border border-border p-3">
          <div className="flex items-center justify-between text-xs">
            <span
              data-current={done}
              data-total={totalItems}
              data-running={batch.isRunning}
            >
              {done} / {totalItems} analyzed
            </span>
            <span
              role="status"
              aria-label="Throughput"
              data-rate={rate.toFixed(2)}
              data-elapsed={elapsedMs}
              data-eta={etaSec.toFixed(1)}
              className="font-mono text-muted-foreground"
            >
              {rate.toFixed(1)}/s · {formatDuration(elapsedMs)}
              {batch.isRunning && ` · ETA ${etaSec.toFixed(0)}s`}
            </span>
          </div>
          {batch.isRunning && (
            <div className="h-1.5 w-full overflow-hidden rounded-full bg-muted">
              <div
                className="h-full rounded-full bg-primary transition-all"
                style={{ width: `${totalItems > 0 ? (done / totalItems) * 100 : 0}%` }}
              />
            </div>
          )}
        </div>
      )}

      {}
      {total > 0 && (
        <div
          data-positive={positive}
          data-negative={negative}
          data-total={total}
          data-avg={avgScore.toFixed(4)}
          className="flex flex-col gap-2 rounded-md border border-border p-3"
        >
          <EntityStatsBar
            counts={{ POSITIVE: positive, NEGATIVE: negative }}
            itemNoun="result"
            registry={{
              POSITIVE: { label: 'Positive', color: 'var(--color-emerald-500, #10b981)' },
              NEGATIVE: { label: 'Negative', color: 'var(--color-rose-500, #f43f5e)' },
            }}
          />
          <div className="flex flex-wrap gap-x-4 gap-y-1 text-xs text-muted-foreground">
            <span>Positive: {total > 0 ? formatScore(positive / total) : '0%'}</span>
            <span>Negative: {total > 0 ? formatScore(negative / total) : '0%'}</span>
            <span>Total analyzed: {total}</span>
            <span>Avg confidence: {formatScore(avgScore)}</span>
          </div>
        </div>
      )}

      {}
      {results.length > 0 && (
        <div className="flex flex-col gap-1.5">
          {windowed && (
            <p className="text-xs text-muted-foreground">
              Showing first {WINDOW_CAP} of {results.length} results (stats cover all).
            </p>
          )}
          <div role="list" aria-label="Sentiment results" className="flex flex-col gap-1.5">
          {visible.map((r, i) => (
            <div
              key={i}
              role="listitem"
              data-label={r.label}
              data-score={r.score.toFixed(4)}
              data-status={r.status}
              className={cn(
                'flex items-center gap-3 rounded-md border p-2 text-sm',
                r.status === 'error' ? 'border-destructive/40 bg-destructive/5' : 'border-border',
              )}
            >
              <span className="min-w-0 flex-1 truncate" title={r.text}>
                {r.text}
              </span>
              <SentimentLabel label={r.label} />
              {r.status === 'ok' && <ConfidenceScoreBadge score={r.score} />}
            </div>
          ))}
          </div>
        </div>
      )}
    </div>
  );
}