feat: guarantee 6 top stories with vector DB fallback

src/api/routes/top_stories.py

@@ -14,8 +14,9 @@ from fastapi import APIRouter, Query, Depends
 from pydantic import BaseModel
 from datetime import datetime
 
-from src.api.dependencies import get_cache_port, get_live_search_port
+from src.api.dependencies import get_cache_port, get_live_search_port, get_vector_store_port
 from src.core.ports.cache_port import CachePort
+from src.core.ports.vector_store_port import VectorStorePort
 from src.infrastructure.adapters.duckduckgo_adapter import DuckDuckGoAdapter
 
 try:
@@ -232,8 +233,8 @@ async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> L
         return []
     
     try:
-        # Focus on Ethiopia-related world-wide news
-        query = "Ethiopia related world-wide news today"
+        # Broaden search to ensure we get enough results
+        query = "Ethiopia latest news breaking world"
         results = await adapter.search(query)
         
         stories = []
@@ -261,7 +262,8 @@ async def fetch_live_stories(n: int = 6, adapter: DuckDuckGoAdapter = None) -> L
 async def get_top_stories(
     force_refresh: bool = Query(default=False, description="Force cache refresh"),
     cache: CachePort = Depends(get_cache_port),
-    adapter: DuckDuckGoAdapter = Depends(get_live_search_port)
+    adapter: DuckDuckGoAdapter = Depends(get_live_search_port),
+    vector_store: VectorStorePort = Depends(get_vector_store_port)
 ):
     """
     Get top 6 news stories for the landing page.
@@ -296,12 +298,39 @@ async def get_top_stories(
     all_stories: List[TopStory] = []
     seen_titles: set = set()
 
-    for story in live_stories + kafka_stories:  # Prioritize live for today's top stories
-        title_key = story.title.lower()[:60]
+    for story in live_stories + kafka_stories:  # Prioritize live
+        title_key = story.title.lower().strip()[:60]
         if title_key not in seen_titles:
             seen_titles.add(title_key)
             all_stories.append(story)
 
+    # ── FALLBACK: If still less than 6, pull from Vector DB (Qdrant) ──────────
+    if len(all_stories) < 6:
+        needed = 6 - len(all_stories)
+        logger.info(f"Top stories fallback: pulling {needed} from Vector DB")
+        try:
+            db_res = vector_store.browse(limit=needed * 2, days_back=7)
+            for p in db_res.get("articles", []):
+                payload = p.payload or {}
+                title = payload.get("title") or payload.get("content", "")[:100]
+                url = payload.get("url") or "#"
+                
+                title_key = title.lower().strip()[:60]
+                if title_key not in seen_titles and len(all_stories) < 6:
+                    seen_titles.add(title_key)
+                    all_stories.append(TopStory(
+                        title=title,
+                        url=url,
+                        source=payload.get("source", "ARKI Intelligence"),
+                        published_at=payload.get("published_at", datetime.utcnow().isoformat()),
+                        category="UPDATE",
+                        excerpt=payload.get("content", "")[:150],
+                        image_url=payload.get("image_url") or payload.get("thumbnail"),
+                        origin="db"
+                    ))
+        except Exception as e:
+            logger.error(f"Top stories DB fallback failed: {e}")
+
     now_iso = datetime.utcnow().isoformat()
     final_stories = all_stories[:6]