feat: article scanner with translation pair detection

1 files changed, 34 insertions, 0 deletions

diff --git a/core/article_scanner.py b/core/article_scanner.py
new file mode 100644
index 0000000..ea6579a
--- /dev/null
+++ b/core/article_scanner.py
@@ -0,0 +1,34 @@
+from __future__ import annotations
+from pathlib import Path
+from core.models import Article
+from core.frontmatter import parse_frontmatter
+
+def scan_articles(blog_root: Path) -> list[Article]:
+    articles: list[Article] = []
+    by_slug: dict[tuple[str, str], Path] = {}
+
+    for lang in ("it", "en"):
+        content_dir = blog_root / "content" / lang / "articles"
+        if not content_dir.exists():
+            continue
+        for index_md in sorted(content_dir.glob("*/index.md")):
+            slug = index_md.parent.name
+            by_slug[(lang, slug)] = index_md
+
+    for (lang, slug), path in by_slug.items():
+        other_lang = "en" if lang == "it" else "it"
+        translation_path = by_slug.get((other_lang, slug))
+        try:
+            fm, _ = parse_frontmatter(path)
+        except (ValueError, Exception):
+            fm = {}
+        articles.append(Article(
+            slug=slug,
+            lang=lang,
+            path=path,
+            frontmatter=fm,
+            has_translation=translation_path is not None,
+            translation_path=translation_path,
+        ))
+
+    return articles