Update app.py

app.py

@@ -78,31 +78,50 @@ def segment_clothing(img, clothes=["Hat", "Upper-clothes", "Skirt", "Pants", "Dr
 
     return img_with_alpha.convert("RGB"), final_mask, detected_categories
 
-def find_similar_images(query_embedding, collection, top_k=5):
+#def find_similar_images(query_embedding, collection, top_k=5):
     # ChromaDB에서 가장 유사한 이미지들을 쿼리합니다.
+#    results = collection.query(
+#        query_embeddings=query_embedding.reshape(1, -1),  # 2D 배열로 변환
+#        n_results=top_k,
+#        include=['metadatas', 'embeddings']  # 임베딩을 포함하도록 수정
+#    )
+#    
+#    # 메타데이터와 임베딩을 추출합니다.
+#    top_metadatas = results['metadatas'][0]  
+#    top_embeddings = results['embeddings'][0]  # 임베딩 가져오기
+#    
+#    # 유클리드 거리 계산
+#    distances = euclidean_distances(query_embedding.reshape(1, -1), top_embeddings).flatten()
+#
+#    structured_results = []
+#    for metadata, distance in zip(top_metadatas, distances):
+#        structured_results.append({
+#            'info': metadata,
+#            'similarity': 1 / (1 + distance)  # 거리 기반 유사도 (거리가 작을수록 유사도가 높음)
+#        })
+    
+#    return structured_results
+
+def find_similar_images(query_embedding, collection, top_k=5):
+    query_embedding = query_embedding.reshape(1, -1)  # Reshape to 2D array for ChromaDB
     results = collection.query(
-        query_embeddings=query_embedding.reshape(1, -1),  # 2D 배열로 변환
+        query_embeddings=query_embedding,
         n_results=top_k,
-        include=['metadatas', 'embeddings']  # 임베딩을 포함하도록 수정
+        include=['metadatas', 'distances']  
     )
     
-    # 메타데이터와 임베딩을 추출합니다.
     top_metadatas = results['metadatas'][0]  
-    top_embeddings = results['embeddings'][0]  # 임베딩 가져오기
+    top_distances = results['distances'][0]  
     
-    # 유클리드 거리 계산
-    distances = euclidean_distances(query_embedding.reshape(1, -1), top_embeddings).flatten()
-
     structured_results = []
-    for metadata, distance in zip(top_metadatas, distances):
+    for metadata, distance in zip(top_metadatas, top_distances):
         structured_results.append({
             'info': metadata,
-            'similarity': 1 / (1 + distance)  # 거리 기반 유사도 (거리가 작을수록 유사도가 높음)
+            'similarity': 1 - distance  
         })
     
     return structured_results
 
-
 # 세션 상태 초기화
 if 'step' not in st.session_state:
     st.session_state.step = 'input'