Update appStore/search.py

appStore/search.py

@@ -1,48 +1,49 @@
 from appStore.prep_utils import get_client
-from langchain_qdrant import FastEmbedSparse, RetrievalMode
+from langchain_qdrant import FastEmbedSparse
 from torch import cuda
 from qdrant_client.http import models
 from langchain_huggingface import HuggingFaceEmbeddings
-# get the device to be used eithe gpu or cpu
-device = 'cuda' if cuda.is_available() else 'cpu'
 
+device = 'cuda' if cuda.is_available() else 'cpu'
 
-def hybrid_search(client, query, collection_name):
+def hybrid_search(client, query, collection_name, limit=300):
     embeddings = HuggingFaceEmbeddings(
-        model_kwargs = {'device': device},
-        encode_kwargs = {'normalize_embeddings': True},
-        model_name='BAAI/bge-m3'
+        model_name='BAAI/bge-m3',
+        model_kwargs={'device': device},
+        encode_kwargs={'normalize_embeddings': True}
     )
-
     sparse_embeddings = FastEmbedSparse(model_name="Qdrant/bm25")
 
-    # embed query
+    # 1) Embed the query
     q_dense = embeddings.embed_query(query)
     q_sparse = sparse_embeddings.embed_query(query)
 
-    results = client.search_batch(collection_name=collection_name,
-                        requests=[
-                            models.SearchRequest(
-                                vector=models.NamedVector(
-                                    name="text-dense",
-                                    vector=q_dense,
-                                ),
-                                limit=limit,  # was 10, now uses the parameter
-                                with_payload = True,
-                            ),
-                            models.SearchRequest(
-                                vector=models.NamedSparseVector(
-                                    name="text-sparse",
-                                    vector=models.SparseVector(
-                                        indices=q_sparse.indices,
-                                        values=q_sparse.values,
-                                    ),
-                                ),
-                                limit=10,
-                                with_payload = True,
-                            ),
-                        ],)
-
+    # 2) Request more than 10 items
+    results = client.search_batch(
+        collection_name=collection_name,
+        requests=[
+            # Dense request
+            models.SearchRequest(
+                vector=models.NamedVector(
+                    name="text-dense",
+                    vector=q_dense,
+                ),
+                limit=limit,  # was 10, now uses the parameter
+                with_payload=True,
+            ),
+            # Sparse request
+            models.SearchRequest(
+                vector=models.NamedSparseVector(
+                    name="text-sparse",
+                    vector=models.SparseVector(
+                        indices=q_sparse.indices,
+                        values=q_sparse.values,
+                    ),
+                ),
+                limit=limit,  # was 10, now uses the parameter
+                with_payload=True,
+            ),
+        ]
+    )
 
     return results
-