scripts/validate-test-pmws.ts

import { readFileSync } from "fs";
import { resolve } from "path";

const envContent = readFileSync(resolve(import.meta.dirname, "../.env"), "utf-8");
for (const line of envContent.split("\n")) {
  const t = line.trim(); if (!t || t.startsWith("#")) continue;
  const e = t.indexOf("="); if (e < 0) continue;
  if (!process.env[t.slice(0, e)]) process.env[t.slice(0, e)] = t.slice(e + 1);
}

import { detectFormat } from "../src/lib/gts/formats";
import { richTextToPlain } from "../src/lib/gts/models/text";
import { runNetflixChecks } from "../src/lib/qc/netflix-checks";
import { runRuleChecks } from "../src/lib/qc/rule-checks";
import type { QCLine } from "../src/lib/qc/types";

const TEST_DIR = resolve(import.meta.dirname, "../test-pmws");
const manifest = JSON.parse(readFileSync(resolve(TEST_DIR, "manifest.json"), "utf-8")) as Array<{
  lang: string;
  file: string;
  sensitiveTermCount: number;
  glyphViolationCount: number;
  expectedIssues: number;
  [key: string]: any;
}>;

console.log(`=== QC VALIDATION: ${manifest.length} PMW files ===`);
console.log("");

let totalFiles = 0;
let passedFiles = 0;
let failedFiles = 0;
const failures: string[] = [];

for (const entry of manifest) {
  totalFiles++;
  const filePath = resolve(TEST_DIR, entry.file);

  try {
    // Parse PMW
    const buf = readFileSync(filePath);
    const handler = detectFormat(buf, entry.file);
    if (!handler) {
      console.log(`✗ ${entry.lang.padEnd(8)} | PARSE FAILED: format not detected`);
      failedFiles++;
      failures.push(`${entry.lang}: format not detected`);
      continue;
    }

    const doc = handler.read(buf, entry.file);
    const track = doc.tracks[0];

    const lines: QCLine[] = track.subtitles.map((sub, i) => ({
      id: i + 1,
      text: typeof sub.text === "string" ? sub.text : richTextToPlain(sub.text),
    }));

    // Run Netflix checks (glyph + sensitive terms)
    const netflixIssues = runNetflixChecks(lines, entry.lang);
    const glyphIssues = netflixIssues.filter((i) => i.category === "glyph");
    const sensitiveIssues = netflixIssues.filter((i) => i.category === "sensitive");

    // Also run rule checks
    const ruleIssues = runRuleChecks(lines, entry.lang);

    // Validate: did we detect at least as many as expected?
    // Note: some terms may match multiple times in different contexts, so actual >= expected
    const sensitiveOk = sensitiveIssues.length >= entry.sensitiveTermCount;
    const glyphOk = glyphIssues.length >= entry.glyphViolationCount;

    const status = sensitiveOk && glyphOk ? "✓" : "✗";
    if (sensitiveOk && glyphOk) {
      passedFiles++;
    } else {
      failedFiles++;
      if (!sensitiveOk) failures.push(`${entry.lang}: sensitive ${sensitiveIssues.length}/${entry.sensitiveTermCount}`);
      if (!glyphOk) failures.push(`${entry.lang}: glyph ${glyphIssues.length}/${entry.glyphViolationCount}`);
    }

    console.log(
      `${status} ${entry.lang.padEnd(8)} | ` +
      `sensitive: ${String(sensitiveIssues.length).padStart(4)}/${String(entry.sensitiveTermCount).padStart(4)} ` +
      `${sensitiveOk ? "✓" : "✗"} | ` +
      `glyph: ${String(glyphIssues.length).padStart(2)}/${String(entry.glyphViolationCount).padStart(2)} ` +
      `${glyphOk ? "✓" : "✗"} | ` +
      `rule: ${String(ruleIssues.length).padStart(3)} | ` +
      `total: ${netflixIssues.length + ruleIssues.length}`
    );
  } catch (err) {
    failedFiles++;
    const msg = (err as Error).message.substring(0, 60);
    console.log(`✗ ${entry.lang.padEnd(8)} | ERROR: ${msg}`);
    failures.push(`${entry.lang}: ${msg}`);
  }
}

console.log("");
console.log("=== SUMMARY ===");
console.log(`Files:  ${totalFiles}`);
console.log(`Passed: ${passedFiles}`);
console.log(`Failed: ${failedFiles}`);
console.log("");

if (failures.length > 0) {
  console.log("FAILURES:");
  for (const f of failures) console.log(`  ${f}`);
} else {
  console.log("ALL 32 LANGUAGES PASSED ✓");
}

// Aggregate stats
const totalSensitive = manifest.reduce((s, e) => s + e.sensitiveTermCount, 0);
const totalGlyphs = manifest.reduce((s, e) => s + e.glyphViolationCount, 0);
console.log("");
console.log(`Injected: ${totalSensitive} sensitive terms + ${totalGlyphs} illegal glyphs across ${totalFiles} files`);