jumpy-lion 0.1.5 → 0.1.6-beta.0

package/package.json

@@ -2,7 +2,7 @@
     "name": "jumpy-lion",
     "description": "",
     "type": "module",
-    "version": "0.1.5",
+    "version": "0.1.6-beta.0",
     "engines": {
         "node": ">=20.0.0"
     },
@@ -21,7 +21,8 @@
         "./package.json": "./package.json"
     },
     "files": [
-        "dist"
+        "dist",
+        "scripts/postinstall.cjs"
     ],
     "license": "Apache-2.0",
     "scripts": {
@@ -30,6 +31,9 @@
         "compile": "tsc -p tsconfig.build.json",
         "copy-extension": "cp -r src/fingerprinting/all-fingerprint-defender dist/fingerprinting/ && cp -r src/fingerprinting/anti-webgpu dist/fingerprinting/",
         "build": "npm run clean && npm run compile && npm run copy-extension",
+        "package:browser": "bash browser/scripts/package-binary.sh",
+        "release:browser": "node scripts/release-browser.cjs",
+        "postinstall": "node scripts/postinstall.cjs",
         "ci:build": "turbo run build --cache-dir=\".turbo\"",
         "test:unit": "vitest run unit",
         "test": "vitest run --silent",

package/scripts/postinstall.cjs

@@ -0,0 +1,242 @@
+#!/usr/bin/env node
+/**
+ * postinstall.cjs — Download the pre-built anti-detect browser binary
+ *
+ * Runs automatically on `npm install`. Downloads the correct platform archive
+ * from the public Apify key-value store `jumpy-lion` (no auth required) and
+ * extracts it into `.browser/` inside the package root.
+ *
+ * Environment variables:
+ *   SKIP_BROWSER_DOWNLOAD=1   Skip entirely (e.g. CI that ships its own Chrome)
+ *   BROWSER_DOWNLOAD_URL      Override the download URL verbatim (mirrors,
+ *                             staging, offline envs). When set, no other env
+ *                             vars are consulted.
+ *
+ * If the download fails for any reason (network error, bad override URL,
+ * asset missing) postinstall prints a warning and exits 0 — the crawler will
+ * fall back to the system Chrome at runtime. `npm install` never fails hard.
+ *
+ * Supported platforms:
+ *   linux-x64      → chrome-v<VERSION>-linux-x64.tar.gz
+ *   darwin-arm64   → chrome-v<VERSION>-darwin-arm64.tar.gz
+ *   darwin-x64     → chrome-v<VERSION>-darwin-x64.tar.gz
+ *   win32-x64      → chrome-v<VERSION>-win32-x64.zip
+ */
+
+'use strict';
+
+const https = require('https');
+const http = require('http');
+const fs = require('fs');
+const path = require('path');
+const { spawnSync } = require('child_process');
+
+// ─── Config ──────────────────────────────────────────────────────────────────
+
+// Public Apify key-value store. `generalAccess: ANYONE_WITH_URL` means record
+// reads do not require an Apify token. If we ever rotate the store, update
+// both this ID and the matching upload target in scripts/release-browser.cjs.
+const KV_STORE_ID = '6R1GXIyIQVzdYnGWd';
+const KV_STORE_URL = `https://api.apify.com/v2/key-value-stores/${KV_STORE_ID}/records`;
+
+const PKG_VERSION = require('../package.json').version;
+const INSTALL_DIR = path.join(__dirname, '..', '.browser');
+
+// Archive name → platform key. Matches package-binary.sh output.
+function archiveName(platformKey, version) {
+    const ext = platformKey === 'win32-x64' ? 'zip' : 'tar.gz';
+    return `chrome-v${version}-${platformKey}.${ext}`;
+}
+
+// Fall-back check for binaries bundled inside the npm tarball itself (beta
+// releases that ship the full binary, see browser/BUILD.md § Beta).
+const BUNDLED_BINARY_PATHS = {
+    'linux-x64': path.join(__dirname, '..', 'browser', 'dist', 'linux-x64', 'chrome'),
+    'darwin-arm64': path.join(__dirname, '..', 'browser', 'dist', 'darwin-arm64', 'Chromium.app', 'Contents', 'MacOS', 'Chromium'),
+    'darwin-x64': path.join(__dirname, '..', 'browser', 'dist', 'darwin-x64', 'Chromium.app', 'Contents', 'MacOS', 'Chromium'),
+    'win32-x64': path.join(__dirname, '..', 'browser', 'dist', 'win32-x64', 'chrome.exe'),
+};
+
+// ─── Main ─────────────────────────────────────────────────────────────────────
+
+async function main() {
+    if (process.env.SKIP_BROWSER_DOWNLOAD === '1') {
+        console.log('[cdp-crawler] SKIP_BROWSER_DOWNLOAD=1 — skipping browser download');
+        return;
+    }
+
+    const platformKey = `${process.platform}-${process.arch}`;
+
+    // Bundled (beta) binary takes precedence.
+    const bundledBinary = BUNDLED_BINARY_PATHS[platformKey];
+    if (bundledBinary && fs.existsSync(bundledBinary)) {
+        console.log('[cdp-crawler] Browser binary found in browser/dist/ — skipping download.');
+        return;
+    }
+
+    if (!BUNDLED_BINARY_PATHS[platformKey]) {
+        console.warn(`[cdp-crawler] No pre-built browser for platform: ${platformKey}`);
+        console.warn('[cdp-crawler] Will fall back to system Chrome on first use.');
+        return;
+    }
+
+    // Already installed at the correct version → skip.
+    const versionFile = path.join(INSTALL_DIR, '.version');
+    const installedVersion = fs.existsSync(versionFile)
+        ? fs.readFileSync(versionFile, 'utf8').trim()
+        : null;
+    if (installedVersion === PKG_VERSION) {
+        console.log(`[cdp-crawler] Anti-detect browser v${PKG_VERSION} already installed.`);
+        return;
+    }
+
+    const name = archiveName(platformKey, PKG_VERSION);
+    const downloadUrl = process.env.BROWSER_DOWNLOAD_URL
+        || `${KV_STORE_URL}/${name}`;
+
+    console.log(`[cdp-crawler] Downloading anti-detect browser for ${platformKey}...`);
+    console.log(`[cdp-crawler] URL: ${downloadUrl}`);
+
+    fs.mkdirSync(INSTALL_DIR, { recursive: true });
+    const assetPath = path.join(INSTALL_DIR, name);
+
+    try {
+        await download(downloadUrl, assetPath);
+        console.log(`[cdp-crawler] Downloaded: ${assetPath}`);
+
+        await extract(assetPath, INSTALL_DIR, platformKey);
+        console.log(`[cdp-crawler] Extracted to: ${INSTALL_DIR}`);
+
+        if (process.platform !== 'win32') {
+            setExecutablePermissions(INSTALL_DIR, platformKey);
+        }
+
+        fs.writeFileSync(versionFile, PKG_VERSION);
+        console.log(`[cdp-crawler] Anti-detect browser v${PKG_VERSION} installed successfully.`);
+
+        fs.unlinkSync(assetPath);
+    } catch (err) {
+        console.warn(`[cdp-crawler] Browser download failed: ${err.message}`);
+        console.warn('[cdp-crawler] Will fall back to system Chrome on first use.');
+        if (fs.existsSync(assetPath)) {
+            try { fs.unlinkSync(assetPath); } catch { /* ignore */ }
+        }
+    }
+}
+
+// ─── Helpers ──────────────────────────────────────────────────────────────────
+
+function download(url, destPath) {
+    return new Promise((resolve, reject) => {
+        const file = fs.createWriteStream(destPath);
+
+        // The KV store record endpoint returns a 302 to a signed S3 URL. No
+        // auth header is needed on either hop for a public store, so the
+        // redirect follower is just a simple recursion.
+        const request = (requestUrl) => {
+            const client = requestUrl.startsWith('https:') ? https : http;
+            const headers = { 'User-Agent': 'cdp-crawler-postinstall' };
+
+            client.get(requestUrl, { headers }, (res) => {
+                if (res.statusCode === 301 || res.statusCode === 302 || res.statusCode === 307) {
+                    const redirectUrl = res.headers.location;
+                    if (!redirectUrl) {
+                        reject(new Error(`Redirect with no location header (status ${res.statusCode})`));
+                        return;
+                    }
+                    res.resume();
+                    request(redirectUrl);
+                    return;
+                }
+
+                if (res.statusCode === 404) {
+                    reject(new Error(
+                        `Asset not found (404). The release archive for v${PKG_VERSION} ` +
+                        `has not been published to the public Apify store yet.\n` +
+                        `Expected: ${url}\n` +
+                        `To build locally: npm run build:browser`,
+                    ));
+                    return;
+                }
+
+                if (res.statusCode !== 200) {
+                    reject(new Error(`HTTP ${res.statusCode} from ${requestUrl}`));
+                    return;
+                }
+
+                const totalBytes = parseInt(res.headers['content-length'] || '0', 10);
+                let receivedBytes = 0;
+                let lastLogPct = 0;
+
+                res.on('data', (chunk) => {
+                    receivedBytes += chunk.length;
+                    if (totalBytes > 0) {
+                        const pct = Math.floor(receivedBytes / totalBytes * 100);
+                        if (pct >= lastLogPct + 10) {
+                            process.stdout.write(`\r[cdp-crawler] Downloading... ${pct}%`);
+                            lastLogPct = pct;
+                        }
+                    }
+                });
+
+                res.pipe(file);
+                file.on('finish', () => {
+                    file.close();
+                    process.stdout.write('\n');
+                    resolve();
+                });
+                file.on('error', reject);
+            }).on('error', reject);
+        };
+
+        request(url);
+    });
+}
+
+async function extract(archivePath, destDir, platformKey) {
+    if (platformKey === 'win32-x64') {
+        const result = spawnSync('powershell', [
+            '-NoProfile', '-NonInteractive', '-Command',
+            `Expand-Archive -Force -Path "${archivePath}" -DestinationPath "${destDir}"`,
+        ], { stdio: 'inherit' });
+        if (result.status !== 0) {
+            throw new Error('Failed to extract zip with PowerShell');
+        }
+    } else {
+        const result = spawnSync('tar', ['-xzf', archivePath, '-C', destDir], { stdio: 'inherit' });
+        if (result.status !== 0) {
+            throw new Error('Failed to extract tar.gz');
+        }
+    }
+}
+
+function setExecutablePermissions(installDir, platformKey) {
+    const executablePaths = {
+        'linux-x64': [
+            path.join(installDir, 'linux-x64', 'chrome'),
+            path.join(installDir, 'linux-x64', 'chrome_sandbox'),
+        ],
+        'darwin-arm64': [
+            path.join(installDir, 'darwin-arm64', 'Chromium.app', 'Contents', 'MacOS', 'Chromium'),
+        ],
+        'darwin-x64': [
+            path.join(installDir, 'darwin-x64', 'Chromium.app', 'Contents', 'MacOS', 'Chromium'),
+        ],
+    };
+
+    const paths = executablePaths[platformKey] || [];
+    for (const exePath of paths) {
+        if (fs.existsSync(exePath)) {
+            fs.chmodSync(exePath, 0o755);
+            console.log(`[cdp-crawler] Set executable: ${exePath}`);
+        }
+    }
+}
+
+// ─── Entry ────────────────────────────────────────────────────────────────────
+
+main().catch((err) => {
+    console.error('[cdp-crawler] postinstall error:', err);
+    // Exit 0 so `npm install` never fails on browser download issues.
+    process.exit(0);
+});