fix: no more duplicates for already fully supported languages (#58)

Co-authored-by: hugo <hugoms@me.com>

.claude/settings.local.json

@@ -19,7 +19,13 @@
       "WebFetch(domain:ricmac.org)",
       "WebFetch(domain:docs.mistral.ai)",
       "Bash(npm uninstall dropbox date-fns @testing-library/user-event @types/dagre electron-store memfs)",
-      "WebSearch"
+      "WebSearch",
+      "WebFetch(domain:ai.google.dev)",
+      "Bash(gh api:*)",
+      "Bash(python3 -c \"import json,sys; [print\\(f[''''name'''']\\) for f in json.load\\(sys.stdin\\)]\")",
+      "Bash(grep -rn choices.*auto.*mistral /Users/gb/mlx-env/lib/python3.14/site-packages/vllm_mlx/ --include=*.py)",
+      "Bash(source ~/mlx-env/bin/activate)",
+      "Bash(python3 -c \":*)"
     ]
   }
 }

src/main/engine/BlogGenerationEngine.ts

@@ -628,6 +628,16 @@ export class BlogGenerationEngine {
       .map((lang) => lang.trim().toLowerCase())
       .filter((lang) => lang.length > 0 && lang !== mainLanguage);
 
+    // When a language has a dedicated subtree (e.g. /fr/), suppress its .lang
+    // translation variant pages (e.g. /2025/01/15/post.fr/) to avoid duplicate
+    // content.  The subtree already contains the translated version at a clean URL.
+    if (additionalLanguages.length > 0) {
+      const subtreeLanguages = new Set(additionalLanguages);
+      publishedRoutePosts = publishedRoutePosts.filter(
+        (p) => !(p as any).translationSourceSlug || !subtreeLanguages.has((p.language ?? '').trim().toLowerCase()),
+      );
+    }
+
     // Determine whether to use worker threads for page generation
     const useWorkers = !!options.dbPath;
 
@@ -1409,9 +1419,23 @@ export class BlogGenerationEngine {
       .map(([category]) => category);
 
     const { publishedPosts, publishedListPosts } = await loadPublishedGenerationSets(this.postEngine, listExcludedCategories);
-    const { routePosts: publishedRoutePosts } = await this.buildPublishedRoutePosts(publishedPosts);
+    let { routePosts: publishedRoutePosts } = await this.buildPublishedRoutePosts(publishedPosts);
     const generationPostIndex = buildGenerationPostIndex(publishedListPosts);
 
+    // --- Build per-language expected paths ---
+    const mainLanguage = (options.language ?? 'en').trim().toLowerCase();
+    const additionalLanguages = (options.blogLanguages ?? [])
+      .map((lang) => lang.trim().toLowerCase())
+      .filter((lang) => lang.length > 0 && lang !== mainLanguage);
+
+    // Suppress .lang variant pages for languages that have a dedicated subtree
+    if (additionalLanguages.length > 0) {
+      const subtreeLanguages = new Set(additionalLanguages);
+      publishedRoutePosts = publishedRoutePosts.filter(
+        (p) => !(p as any).translationSourceSlug || !subtreeLanguages.has((p.language ?? '').trim().toLowerCase()),
+      );
+    }
+
     const { sitemapXml } = buildSitemapAndFeeds({
       baseUrl: options.baseUrl,
       projectName: options.projectName,
@@ -1427,12 +1451,6 @@ export class BlogGenerationEngine {
     await fs.mkdir(htmlDir, { recursive: true });
     const sitemapPath = path.join(htmlDir, 'sitemap.xml');
 
-    // --- Build per-language expected paths ---
-    const mainLanguage = (options.language ?? 'en').trim().toLowerCase();
-    const additionalLanguages = (options.blogLanguages ?? [])
-      .map((lang) => lang.trim().toLowerCase())
-      .filter((lang) => lang.length > 0 && lang !== mainLanguage);
-
     let sitemapToWrite = sitemapXml;
     const additionalExpectedPaths: string[] = [];
     const additionalPostTimestampChecks: Array<{

tests/engine/BlogGenerationEngine.test.ts

@@ -2325,4 +2325,56 @@ describe('BlogGenerationEngine', () => {
     const page2 = await readFile(path.join(tempDir, 'html', 'tag', 'paginated', 'page', '2', 'index.html'), 'utf-8');
     expect(page2).toContain('/tag/paginated/');
   });
+
+  it('does not generate .lang variant pages when that language has a subtree', async () => {
+    const sourcePost = makePost({
+      id: '1',
+      slug: 'hello-world',
+      title: 'Hello World',
+      content: '# Hello World\n\nEnglish body',
+      language: 'en',
+      availableLanguages: ['en', 'fr'],
+      createdAt: new Date('2025-01-15T10:00:00Z'),
+      updatedAt: new Date('2025-01-15T10:00:00Z'),
+    });
+    const translationsByPostId = new Map<string, PostTranslationData[]>([
+      ['1', [{
+        id: 'translation-1-fr',
+        projectId: 'default',
+        translationFor: '1',
+        language: 'fr',
+        title: 'Bonjour le monde',
+        excerpt: 'Resume FR',
+        content: '# Bonjour le monde\n\nCorps FR',
+        status: 'published',
+        createdAt: new Date('2025-01-15T10:05:00Z'),
+        updatedAt: new Date('2025-01-15T10:05:00Z'),
+        publishedAt: new Date('2025-01-15T10:06:00Z'),
+        filePath: path.join(tempDir, 'posts', 'hello-world.fr.md'),
+      }]],
+    ]);
+
+    setupPosts([sourcePost]);
+    mockPostEngine.getPostTranslations.mockImplementation(async (postId: string) => translationsByPostId.get(postId) ?? []);
+
+    const { BlogGenerationEngine } = await import('../../src/main/engine/BlogGenerationEngine');
+    const engine = new BlogGenerationEngine(mockPostEngine, mockMediaEngine, mockPostMediaEngine);
+
+    await engine.generate({
+      projectId: 'test',
+      projectName: 'Test Blog',
+      dataDir: tempDir,
+      baseUrl: 'https://example.com',
+      language: 'en',
+      blogLanguages: ['en', 'fr'],
+    }, vi.fn());
+
+    // The fr subtree should exist
+    const frSubtreeExists = await fileExists(path.join(tempDir, 'html', 'fr', '2025', '01', '15', 'hello-world', 'index.html'));
+    expect(frSubtreeExists).toBe(true);
+
+    // The .fr variant page should NOT exist since fr has a language subtree
+    const variantExists = await fileExists(path.join(tempDir, 'html', '2025', '01', '15', 'hello-world.fr', 'index.html'));
+    expect(variantExists).toBe(false);
+  });
 });