lint and benchmark

Author: willwade

scripts/README.md

@@ -71,6 +71,7 @@ Translation and localization tools.
 General-purpose utility scripts.
 
 - **image-map.js** - Image mapping and analysis tools
+- **bench.js** - Run simple performance benchmarks for load/extract/process paths
 
 ## 🚀 Usage
 
@@ -84,6 +85,15 @@ npm run build
 node scripts/analysis/analyze_pageset.js path/to/pageset.gridset
 node scripts/analysis/extract_vocabulary.js path/to/pageset.gridset
 npx ts-node scripts/conversion/txt-to-gridset.ts input.tsv output.gridset
+
+# Benchmark a processor by file type (loadIntoTree)
+node scripts/bench.js --file examples/example.gridset
+
+# Benchmark text extraction
+node scripts/bench.js --file examples/example.obz --mode extract --iterations 10
+
+# Benchmark translation processing
+node scripts/bench.js --file examples/example.sps --mode process --iterations 3
 ```
 
 ## 🔑 Environment Variables

scripts/bench.js

@@ -0,0 +1,138 @@
+#!/usr/bin/env node
+
+const fs = require('fs');
+const path = require('path');
+const os = require('os');
+const { performance } = require('perf_hooks');
+
+function requireLibrary() {
+  const distPath = path.resolve(__dirname, '../dist');
+  try {
+    // eslint-disable-next-line global-require, import/no-dynamic-require
+    return require(distPath);
+  } catch (error) {
+    console.error('Unable to load the built library from dist/. Have you run "npm run build"?');
+    throw error;
+  }
+}
+
+const ENGINE_TO_EXT = {
+  dot: '.dot',
+  opml: '.opml',
+  obf: '.obf',
+  obz: '.obz',
+  obfset: '.obfset',
+  gridset: '.gridset',
+  gridsetx: '.gridsetx',
+  snap: '.sps',
+  spb: '.spb',
+  touchchat: '.ce',
+  excel: '.xlsx',
+  applepanels: '.plist',
+  asterics: '.grd',
+};
+
+function parseArgs(argv) {
+  const args = {
+    file: '',
+    engine: '',
+    mode: 'load',
+    iterations: 5,
+  };
+  for (let i = 2; i < argv.length; i += 1) {
+    const value = argv[i + 1];
+    if (argv[i] === '--file' && value) {
+      args.file = value;
+      i += 1;
+    } else if (argv[i] === '--engine' && value) {
+      args.engine = value;
+      i += 1;
+    } else if (argv[i] === '--mode' && value) {
+      args.mode = value;
+      i += 1;
+    } else if (argv[i] === '--iterations' && value) {
+      args.iterations = Math.max(1, Number(value));
+      i += 1;
+    }
+  }
+  return args;
+}
+
+function detectExtension(inputPath, engineOverride) {
+  if (engineOverride) {
+    const key = engineOverride.toLowerCase();
+    return ENGINE_TO_EXT[key] || engineOverride;
+  }
+  const ext = path.extname(inputPath);
+  return ext || '';
+}
+
+function mean(values) {
+  return values.reduce((sum, v) => sum + v, 0) / values.length;
+}
+
+function percentile(values, p) {
+  const sorted = [...values].sort((a, b) => a - b);
+  const idx = Math.min(sorted.length - 1, Math.floor(sorted.length * p));
+  return sorted[idx];
+}
+
+async function main() {
+  const args = parseArgs(process.argv);
+  if (!args.file) {
+    console.error('Usage: node scripts/bench.js --file <path> [--engine <name>] [--mode load|extract|process] [--iterations N]');
+    process.exit(1);
+  }
+
+  const inputPath = path.resolve(process.cwd(), args.file);
+  if (!fs.existsSync(inputPath)) {
+    console.error(`File not found: ${inputPath}`);
+    process.exit(1);
+  }
+
+  const { getProcessor } = requireLibrary();
+  const extension = detectExtension(inputPath, args.engine);
+  const processor = getProcessor(extension);
+  const iterations = args.iterations;
+
+  let translations = null;
+  if (args.mode === 'process') {
+    const texts = await processor.extractTexts(inputPath);
+    translations = new Map();
+    texts.slice(0, 50).forEach((text) => {
+      if (typeof text === 'string' && text.trim().length > 0) {
+        translations.set(text, `${text}-bench`);
+      }
+    });
+  }
+
+  const timings = [];
+  for (let i = 0; i < iterations; i += 1) {
+    const start = performance.now();
+    if (args.mode === 'extract') {
+      await processor.extractTexts(inputPath);
+    } else if (args.mode === 'process') {
+      const outputPath = path.join(os.tmpdir(), `aac-bench-${Date.now()}-${i}${extension}`);
+      await processor.processTexts(inputPath, translations || new Map(), outputPath);
+    } else {
+      await processor.loadIntoTree(inputPath);
+    }
+    timings.push(performance.now() - start);
+  }
+
+  const avg = mean(timings);
+  const p95 = percentile(timings, 0.95);
+
+  console.log('Bench results');
+  console.log(`- file: ${inputPath}`);
+  console.log(`- engine: ${args.engine || extension}`);
+  console.log(`- mode: ${args.mode}`);
+  console.log(`- iterations: ${iterations}`);
+  console.log(`- avg ms: ${avg.toFixed(2)}`);
+  console.log(`- p95 ms: ${p95.toFixed(2)}`);
+}
+
+main().catch((error) => {
+  console.error(error);
+  process.exit(1);
+});

test/processors/gridset/symbols.test.ts

@@ -52,10 +52,7 @@ describe('gridset symbols utilities', () => {
     };
 
     const refs = extractSymbolReferences(tree);
-    expect(refs).toEqual([
-      '[tawasl]/animals/cat.png',
-      '[widgit]/food/apple.png',
-    ]);
+    expect(refs).toEqual(['[tawasl]/animals/cat.png', '[widgit]/food/apple.png']);
 
     const usage = analyzeSymbolUsage(tree);
     expect(usage.totalSymbols).toBe(2);
@@ -65,11 +62,7 @@ describe('gridset symbols utilities', () => {
   });
 
   it('creates embedded filenames for symbol references', () => {
-    expect(symbolReferenceToFilename('[widgit]/food/apple.png', 2, 3)).toBe(
-      '2-3-0-text-0.png'
-    );
-    expect(symbolReferenceToFilename('[widgit]/food/apple', 1, 1)).toBe(
-      '1-1-0-text-0.png'
-    );
+    expect(symbolReferenceToFilename('[widgit]/food/apple.png', 2, 3)).toBe('2-3-0-text-0.png');
+    expect(symbolReferenceToFilename('[widgit]/food/apple', 1, 1)).toBe('1-1-0-text-0.png');
   });
 });