add RSSFeed and subheadline selector

flairNLP · Apr 19, 2024 · 98513f4 · 98513f4
1 parent ccefb98
commit 98513f4
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/src/fundus/publishers/de/__init__.py b/src/fundus/publishers/de/__init__.py
@@ -232,6 +232,10 @@ class DE(PublisherEnum):
     RheinischePost = PublisherSpec(
         name="Rheinische Post",
         domain="https://rp-online.de/",
-        sources=[NewsMap("https://rp-online.de/sitemap-news.xml"), Sitemap("https://rp-online.de/sitemap.xml")],
+        sources=[
+            RSSFeed("https://rp-online.de/feed.rss"),
+            NewsMap("https://rp-online.de/sitemap-news.xml"),
+            Sitemap("https://rp-online.de/sitemap.xml"),
+        ],
         parser=RheinischePostParser,
     )
diff --git a/src/fundus/publishers/de/rheinische_post.py b/src/fundus/publishers/de/rheinische_post.py
@@ -16,13 +16,15 @@ class RheinischePostParser(ParserProxy):
     class V1(BaseParser):
         _summary_selector = CSSSelector("strong[data-cy='intro']")
         _paragraph_selector = CSSSelector("div[data-cy='article-content'] p")
+        _subheadline_selector = CSSSelector("div[data-cy='article-content'] h2")
 
         @attribute
         def body(self) -> ArticleBody:
             return extract_article_body_with_selector(
                 self.precomputed.doc,
                 summary_selector=self._summary_selector,
                 paragraph_selector=self._paragraph_selector,
+                subheadline_selector=self._subheadline_selector,
             )
 
         @attribute