Update rag/retriever.py

rag/retriever.py

@@ -1,24 +1,14 @@
-from app.deps import search, embed_texts
-import numpy as np
+from app.deps import search
 
 def retrieve(query: str, top_k=8):
-    hits = search(query, top_k=top_k)
-    return hits
+    return search(query, top_k=top_k)
 
 def format_citations(hits):
-    # 重複URLをまとめて [1], [2]...を付与
     url_to_id = {}
-    citations = {}
+    uniq = []
     for h in hits:
         u = h["source_url"]
         if u not in url_to_id:
             url_to_id[u] = len(url_to_id) + 1
-        citations.append((url_to_id[u], u))
-    # unique 保持
-    uniq = []
-    seen = set()
-    for cid, u in citations:
-        if u in seen: continue
-        seen.add(u)
-        uniq.append((cid, u))
+            uniq.append((url_to_id[u], u))
     return uniq