@link-assistant/web-capture 1.10.7 → 1.10.9

package/CHANGELOG.md

@@ -1,5 +1,20 @@
 # @link-assistant/web-capture
 
+## 1.10.9
+
+### Patch Changes
+
+- Auto-release unreleased changes
+
+## 1.10.8
+
+### Patch Changes
+
+- 8cee5e4: Keep JS/Rust kreuzberg parity for the html-to-markdown 3.6 image dimensions
+  change (issue #137). Export `normalizeStructuredKeys` and add regression tests
+  asserting inline image `dimensions` are exposed as `{ width, height }`, mirroring
+  the Rust `inline_image_to_json` fix.
+
 ## 1.10.7
 
 ### Patch Changes

package/jest.config.mjs

@@ -9,6 +9,7 @@ export default {
   testEnvironment: 'node',
   verbose: true,
   testMatch: [
+    '**/tests/ci/**/*.test.js',
     '**/tests/unit/**/*.test.js',
     '**/tests/mock/**/*.test.js',
     '**/tests/e2e/**/*.test.js',

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@link-assistant/web-capture",
-  "version": "1.10.7",
+  "version": "1.10.9",
   "description": "CLI and microservice to render web pages as HTML, Markdown, or PNG",
   "main": "src/index.js",
   "bin": {

package/src/kreuzberg.js

@@ -39,7 +39,19 @@ function toSnakeCase(key) {
   return key.replace(/[A-Z]/g, (match) => `_${match.toLowerCase()}`);
 }
 
-function normalizeStructuredKeys(value) {
+/**
+ * Recursively normalize object keys to snake_case.
+ *
+ * Used to give the structured `metadata`, `tables`, and `images` payloads a
+ * stable shape regardless of the casing the native binding emits. In
+ * particular, inline image `dimensions` are exposed as `{ width, height }` —
+ * mirroring the Rust `inline_image_to_json` mapping after the
+ * html-to-markdown 3.6 `ImageDimensions` change (see issue #137).
+ *
+ * @param {*} value - Value to normalize
+ * @returns {*} The value with all object keys converted to snake_case
+ */
+export function normalizeStructuredKeys(value) {
   if (Array.isArray(value)) {
     return value.map(normalizeStructuredKeys);
   }

package/tests/ci/detect-code-changes.test.js

@@ -0,0 +1,174 @@
+import { execFileSync } from 'child_process';
+import {
+  mkdirSync,
+  mkdtempSync,
+  readFileSync,
+  rmSync,
+  writeFileSync,
+} from 'fs';
+import { tmpdir } from 'os';
+import { dirname, join, resolve } from 'path';
+import { fileURLToPath } from 'url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const repoRoot = resolve(__dirname, '../../..');
+const detectorPath = resolve(repoRoot, 'scripts/detect-code-changes.mjs');
+
+function run(command, args, cwd, options = {}) {
+  return execFileSync(command, args, {
+    cwd,
+    encoding: 'utf8',
+    stdio: ['ignore', 'pipe', 'pipe'],
+    ...options,
+  });
+}
+
+function git(repoPath, args) {
+  return run('git', args, repoPath);
+}
+
+function writeTrackedFile(repoPath, relativePath, contents) {
+  const filePath = join(repoPath, relativePath);
+  mkdirSync(dirname(filePath), { recursive: true });
+  writeFileSync(filePath, contents);
+}
+
+function commitAll(repoPath, message) {
+  git(repoPath, ['add', '.']);
+  git(repoPath, ['commit', '-m', message]);
+}
+
+function createMergedFeatureRepo() {
+  const repoPath = mkdtempSync(join(tmpdir(), 'web-capture-detect-'));
+
+  initializeRepo(repoPath);
+
+  writeTrackedFile(repoPath, 'README.md', '# fixture\n');
+  commitAll(repoPath, 'initial commit');
+
+  git(repoPath, ['checkout', '-b', 'feature']);
+  writeTrackedFile(repoPath, 'rust/src/lib.rs', 'pub fn capture() {}\n');
+  commitAll(repoPath, 'add rust code');
+
+  writeTrackedFile(repoPath, 'docs/notes.md', '# docs only tail commit\n');
+  commitAll(repoPath, 'document rust change');
+
+  git(repoPath, ['checkout', 'main']);
+  git(repoPath, ['merge', '--no-ff', 'feature', '-m', 'merge feature']);
+
+  return repoPath;
+}
+
+function createSingleCommitRepo() {
+  return createSingleCommitRepoWithFiles({
+    'js/src/index.js': 'export const ok = true;\n',
+  });
+}
+
+function createSingleCommitRepoWithFiles(files) {
+  const repoPath = mkdtempSync(join(tmpdir(), 'web-capture-detect-'));
+
+  initializeRepo(repoPath);
+  for (const [relativePath, contents] of Object.entries(files)) {
+    writeTrackedFile(repoPath, relativePath, contents);
+  }
+  commitAll(repoPath, 'initial commit');
+
+  return repoPath;
+}
+
+function initializeRepo(repoPath) {
+  git(repoPath, ['init']);
+  git(repoPath, ['config', 'user.email', 'ci@example.test']);
+  git(repoPath, ['config', 'user.name', 'CI Test']);
+  git(repoPath, ['checkout', '-b', 'main']);
+}
+
+function parseOutputs(outputFile) {
+  return Object.fromEntries(
+    readFileSync(outputFile, 'utf8')
+      .trim()
+      .split('\n')
+      .filter(Boolean)
+      .map((line) => {
+        const separatorIndex = line.indexOf('=');
+        return [line.slice(0, separatorIndex), line.slice(separatorIndex + 1)];
+      })
+  );
+}
+
+function runDetector(repoPath, eventName) {
+  const outputFile = join(repoPath, `github-output-${eventName}.txt`);
+  const env = {
+    ...process.env,
+    GITHUB_EVENT_NAME: eventName,
+    GITHUB_OUTPUT: outputFile,
+  };
+  delete env.GITHUB_BASE_SHA;
+  delete env.GITHUB_HEAD_SHA;
+
+  run(process.execPath, [detectorPath], repoPath, { env });
+
+  return parseOutputs(outputFile);
+}
+
+describe('detect-code-changes', () => {
+  let repoPath;
+
+  afterEach(() => {
+    if (repoPath) {
+      rmSync(repoPath, { recursive: true, force: true });
+      repoPath = undefined;
+    }
+  });
+
+  test('detects code introduced by a real merge commit pushed to main', () => {
+    repoPath = createMergedFeatureRepo();
+
+    const outputs = runDetector(repoPath, 'push');
+
+    expect(outputs['rust-code-changed']).toBe('true');
+    expect(outputs['rust-changed']).toBe('true');
+    expect(outputs['any-rust-code-changed']).toBe('true');
+    expect(outputs['any-code-changed']).toBe('true');
+    expect(outputs['docs-changed']).toBe('true');
+  });
+
+  test('keeps pull request merge commits scoped to the PR head commit', () => {
+    repoPath = createMergedFeatureRepo();
+
+    const outputs = runDetector(repoPath, 'pull_request');
+
+    expect(outputs['rust-code-changed']).toBe('false');
+    expect(outputs['rust-changed']).toBe('false');
+    expect(outputs['any-rust-code-changed']).toBe('false');
+    expect(outputs['any-code-changed']).toBe('false');
+    expect(outputs['docs-changed']).toBe('true');
+  });
+
+  test('detects files when a push only has one commit', () => {
+    repoPath = createSingleCommitRepo();
+
+    const outputs = runDetector(repoPath, 'push');
+
+    expect(outputs['js-code-changed']).toBe('true');
+    expect(outputs['js-changed']).toBe('true');
+    expect(outputs['any-js-code-changed']).toBe('true');
+    expect(outputs['any-code-changed']).toBe('true');
+  });
+
+  test('runs JS checks for CI tests without requiring a changeset', () => {
+    repoPath = createSingleCommitRepoWithFiles({
+      'js/jest.config.mjs': 'export default {};\n',
+      'js/tests/ci/workflow-policy.test.js': 'test("ok", () => {});\n',
+    });
+
+    const outputs = runDetector(repoPath, 'push');
+
+    expect(outputs['js-code-changed']).toBe('false');
+    expect(outputs['js-changed']).toBe('true');
+    expect(outputs['any-js-code-changed']).toBe('true');
+    expect(outputs['any-code-changed']).toBe('true');
+  });
+});

package/tests/ci/rust-dockerfile.test.js

@@ -0,0 +1,54 @@
+import { readFileSync } from 'fs';
+import { dirname, resolve } from 'path';
+import { fileURLToPath } from 'url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const repoRoot = resolve(__dirname, '../../..');
+
+function parseVersion(version) {
+  return version.split('.').map((part) => Number(part));
+}
+
+function compareVersions(left, right) {
+  const leftParts = parseVersion(left);
+  const rightParts = parseVersion(right);
+  const maxLength = Math.max(leftParts.length, rightParts.length);
+
+  for (let index = 0; index < maxLength; index += 1) {
+    const leftPart = leftParts[index] || 0;
+    const rightPart = rightParts[index] || 0;
+
+    if (leftPart !== rightPart) {
+      return leftPart - rightPart;
+    }
+  }
+
+  return 0;
+}
+
+describe('Rust Dockerfile', () => {
+  test('uses a Rust builder image that satisfies Cargo.toml rust-version', () => {
+    const cargoToml = readFileSync(
+      resolve(repoRoot, 'rust/Cargo.toml'),
+      'utf8'
+    );
+    const dockerfile = readFileSync(
+      resolve(repoRoot, 'rust/Dockerfile'),
+      'utf8'
+    );
+
+    const requiredRustVersion = cargoToml.match(
+      /^rust-version = "([^"]+)"/m
+    )?.[1];
+    const dockerRustVersion = dockerfile.match(
+      /^FROM rust:(\d+(?:\.\d+){1,2})[-\s]/m
+    )?.[1];
+
+    expect(requiredRustVersion).toBeDefined();
+    expect(dockerRustVersion).toBeDefined();
+    expect(
+      compareVersions(dockerRustVersion, requiredRustVersion)
+    ).toBeGreaterThanOrEqual(0);
+  });
+});

package/tests/ci/workflow-policy.test.js

@@ -0,0 +1,63 @@
+import { readFileSync } from 'fs';
+import { dirname, resolve } from 'path';
+import { fileURLToPath } from 'url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const repoRoot = resolve(__dirname, '../../..');
+
+const workflowFiles = [
+  '.github/workflows/js.yml',
+  '.github/workflows/rust.yml',
+  '.github/workflows/parity.yml',
+];
+
+function readRootFile(relativePath) {
+  return readFileSync(resolve(repoRoot, relativePath), 'utf8');
+}
+
+function readWorkflow(relativePath) {
+  return readRootFile(relativePath);
+}
+
+describe('workflow policy', () => {
+  test.each(workflowFiles)(
+    '%s uses release-safe concurrency cancellation',
+    (workflowFile) => {
+      const workflow = readWorkflow(workflowFile);
+
+      expect(workflow).toContain(
+        "cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}"
+      );
+      expect(workflow).not.toContain(
+        "cancel-in-progress: ${{ github.ref == 'refs/heads/main' }}"
+      );
+    }
+  );
+
+  test.each(workflowFiles)(
+    '%s uses the current checkout action',
+    (workflowFile) => {
+      const workflow = readWorkflow(workflowFile);
+
+      expect(workflow).toContain('uses: actions/checkout@v6');
+      expect(workflow).not.toContain('uses: actions/checkout@v4');
+    }
+  );
+
+  test('Rust workflow uses the current cache action', () => {
+    const workflow = readWorkflow('.github/workflows/rust.yml');
+
+    expect(workflow).toContain('uses: actions/cache@v5');
+    expect(workflow).not.toContain('uses: actions/cache@v4');
+  });
+
+  test('parity check ignores repository-level CI tests', () => {
+    const parityScript = readRootFile('scripts/check-js-rust-parity.mjs');
+
+    expect(parityScript).toContain('"js/tests/ci/"');
+    expect(parityScript).toContain(
+      'matchingFiles(files, JS_PARITY_PATHS, JS_PARITY_IGNORE_PATHS)'
+    );
+  });
+});

package/tests/unit/kreuzberg.test.js

@@ -1,6 +1,7 @@
 import {
   convertWithKreuzberg,
   isKreuzbergAvailable,
+  normalizeStructuredKeys,
 } from '../../src/kreuzberg.js';
 
 let available;
@@ -207,3 +208,42 @@ describe('kreuzberg html-to-markdown integration', () => {
     expect(result.content).toContain('Item 3');
   });
 });
+
+describe('kreuzberg structured key normalization (issue #137 parity)', () => {
+  // Mirrors the Rust `inline_image_to_json` regression tests: after the
+  // html-to-markdown 3.6 change, inline image dimensions are a structured
+  // `{ width, height }` value rather than a tuple. The normalization pipeline
+  // that backs `result.images` must preserve those `width`/`height` keys.
+  it('preserves width/height keys on inline image dimensions', () => {
+    const images = normalizeStructuredKeys([
+      {
+        format: 'png',
+        filename: 'pixel.png',
+        description: 'one pixel',
+        dimensions: { width: 800, height: 600 },
+        source: 'ImgDataUri',
+      },
+    ]);
+
+    expect(images).toHaveLength(1);
+    expect(images[0].dimensions).toEqual({ width: 800, height: 600 });
+    expect(images[0].format).toBe('png');
+    expect(images[0].filename).toBe('pixel.png');
+  });
+
+  it('leaves images without dimensions untouched', () => {
+    const images = normalizeStructuredKeys([
+      { format: 'svg', source: 'SvgElement', dimensions: null },
+    ]);
+
+    expect(images[0].dimensions).toBeNull();
+  });
+
+  it('snake_cases nested structured keys', () => {
+    const normalized = normalizeStructuredKeys({
+      openGraph: { siteName: 'Example' },
+    });
+
+    expect(normalized).toEqual({ open_graph: { site_name: 'Example' } });
+  });
+});