feat: multi-region parallel search for top stories

src/api/routes/top_stories.py

@@ -232,19 +232,23 @@ async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> L
         return []
     
 async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> List[TopStory]:
-    """Fetch N live stories from DuckDuckGo using multiple queries to ensure high yield"""
+    """Fetch N live stories from DuckDuckGo using multi-region queries for maximum yield"""
     if not adapter:
         return []
     
     try:
-        # Run multiple queries in parallel for better coverage
-        queries = [
-            "Ethiopia news breaking today",
-            "Addis Ababa latest updates",
-            "Ethiopia world news headlines"
+        # We run 4 parallel searches with different regional focuses
+        search_configs = [
+            {"q": "Ethiopia news breaking today", "reg": "et-en"},   # Local Focus
+            {"q": "Ethiopia latest breaking news", "reg": "wt-wt"},  # Global Focus (CNN, BBC, etc)
+            {"q": "Addis Ababa news updates", "reg": "et-en"},      # Capital Focus
+            {"q": "Ethiopia world news reporting", "reg": "us-en"}  # International Perspective
         ]
         
-        search_tasks = [adapter.search(q) for q in queries]
+        search_tasks = [
+            adapter.search(conf["q"], region=conf["reg"], max_results=10) 
+            for conf in search_configs
+        ]
         all_results_lists = await asyncio.gather(*search_tasks)
         
         # Flatten and deduplicate
@@ -258,7 +262,8 @@ async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> L
                 title = r.get("title", "Untitled")
                 title_key = title.lower().strip()[:60]
                 
-                if url in seen_urls or title_key in seen_titles:
+                # Check for duplicates or empty titles
+                if url in seen_urls or title_key in seen_titles or len(title) < 10:
                     continue
                 
                 seen_urls.add(url)
@@ -275,7 +280,10 @@ async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> L
                     origin="live",
                 ))
         
-        logger.info(f"Multi-query live search: collected {len(stories)} unique stories")
+        # Sorting: Prioritize those with images, then by freshness
+        stories.sort(key=lambda s: (1 if s.image_url else 0, s.published_at), reverse=True)
+        
+        logger.info(f"Multi-region search: collected {len(stories)} unique stories")
         return stories[:n]
     except Exception as e:
         logger.error(f"Live top stories error: {e}")