Added memory benchmark (#979)

Balearica · web-flow · commit 18bd4188a2e1 · 2024-12-20T22:15:29.000-08:00
diff --git a/.eslintrc b/.eslintrc
@@ -18,6 +18,7 @@
     "no-restricted-syntax": ["error", "ForInStatement", "LabeledStatement", "WithStatement"],
     // The `prefer-destructuring` rule causes issues
     // https://github.com/naptha/tesseract.js/issues/847
-    "prefer-destructuring": 0
+    "prefer-destructuring": 0,
+    "no-plusplus": ["error", { "allowForLoopAfterthoughts": true }]
   }
 }
diff --git a/benchmarks/node/memory-benchmark.js b/benchmarks/node/memory-benchmark.js
@@ -0,0 +1,61 @@
+#!/usr/bin/env node
+const path = require('path');
+const { createWorker, createScheduler } = require('../../src');
+
+const formatBytes = (bytes) => `${(bytes / 1024 / 1024).toFixed(2)} MB`;
+const formatTime = (seconds) => `${seconds.toFixed(2)}s`;
+
+const getMemoryRow = (iteration, memUsage, time) => `| ${iteration} | ${formatTime(time)} | ${formatBytes(memUsage.heapUsed)} | ${formatBytes(memUsage.heapTotal)} | ${formatBytes(memUsage.rss)} | ${formatBytes(memUsage.external)} |`;
+
+const scheduler = createScheduler();
+
+if (!global.gc) {
+  console.log('Garbage collection unavailable.  Pass --expose-gc '
+    + 'when launching node to enable forced garbage collection.');
+}
+
+const workerGen = async () => {
+  const worker = await createWorker('eng', 1, { cachePath: '.' });
+  scheduler.addWorker(worker);
+};
+
+(async () => {
+  const fileArr = [
+    path.join(__dirname, '..', 'data', 'meditations.jpg'),
+    path.join(__dirname, '..', 'data', 'tyger.jpg'),
+    path.join(__dirname, '..', 'data', 'testocr.png'),
+  ];
+
+  const workerN = 4;
+  const resArr = Array(workerN);
+  for (let i = 0; i < workerN; i++) {
+    resArr[i] = workerGen();
+  }
+  await Promise.all(resArr);
+
+  // Print table header
+  console.log('| Iteration | Time | Heap Used | Heap Total | RSS | External |');
+  console.log('|-----------|------|------------|------------|-----|----------|');
+
+  for (let i = 0; i < 10; i++) {
+    let iterationTime = 0;
+    for (const file of fileArr) {
+      const time1 = Date.now();
+      const promises = [];
+      for (let j = 0; j < 10; j++) {
+        promises.push(scheduler.addJob('recognize', file));
+      }
+      // eslint-disable-next-line no-await-in-loop
+      await Promise.all(promises);
+
+      if (global.gc) global.gc();
+
+      const time2 = Date.now();
+      const timeDif = (time2 - time1) / 1e3;
+      iterationTime += timeDif;
+    }
+    // Print memory stats and time after each iteration
+    console.log(getMemoryRow(i + 1, process.memoryUsage(), iterationTime));
+  }
+  scheduler.terminate();
+})();

Original file line number	Diff line number	Diff line change
`@@ -18,6 +18,7 @@`
`18`	`18`	`"no-restricted-syntax": ["error", "ForInStatement", "LabeledStatement", "WithStatement"],`
`19`	`19`	// The `prefer-destructuring` rule causes issues
`20`	`20`	`// https://github.com/naptha/tesseract.js/issues/847`
`21`		`- "prefer-destructuring": 0`
	`21`	`+ "prefer-destructuring": 0,`
	`22`	`+ "no-plusplus": ["error", { "allowForLoopAfterthoughts": true }]`
`22`	`23`	`}`
`23`	`24`	`}`