Update app.py

app.py

@@ -20,14 +20,7 @@ from fastembed_ext import FastEmbedEmbeddingsLc
 from fastembed.sparse.splade_pp import supported_splade_models
 from fastembed import SparseTextEmbedding, SparseEmbedding
 from unstructured.partition.auto import partition
-from qdrant_client.models import (
-    NamedSparseVector,
-    NamedVector,
-    SparseVector,
-    PointStruct,
-    SearchRequest,
-    ScoredPoint,
-)
+from pymilvus import MilvusClient
 
 def make_points(texts: List[str], metadatas: List[dict], dense: List[ndarray], sparse: List[SparseEmbedding])-> List[PointStruct]:
     points = []
@@ -186,25 +179,38 @@ def load_models_and_documents():
             n_gpu_layers=32
         )
 
-        provider = ['CUDAExecutionProvider']
-
-        dense_model = FastEmbedEncoder(
+        dense_model = HuggingfaceEncoder(
             name='mixedbread-ai/mxbai-embed-large-v1',
-            providers=provider,
-            cache_dir=os.getenv('HF_HOME')
+            device='cuda'
         )
         
         sparse_model = SparseTextEmbedding(
             'Qdrant/all_miniLM_L6_v2_with_attentions',
-            cache_dir=os.getenv('HF_HOME'),
-            providers=provider
+            cache_dir=os.getenv('HF_HOME')
         )
 
         nltk.download('punkt')
         nltk.download('averaged_perceptron_tagger')
 
-    client = QdrantClient(':memory:')
+    client = MilvusClient('https://' + os.getenv('SPACE_HOST') + ':' + str(19530))
     collection_name = 'collection_demo'
+
+    fields = [
+        FieldSchema(name="pk", dtype=DataType.VARCHAR, is_primary=True, auto_id=True, max_length=100),
+        FieldSchema(name="text", dtype=DataType.VARCHAR, max_length=8192),
+        FieldSchema(name="sparse_vector", dtype=DataType.SPARSE_FLOAT_VECTOR),
+        FieldSchema(name="dense_vector", dtype=DataType.FLOAT_VECTOR, dim=1024)
+    ]
+    
+    schema = CollectionSchema(fields, "")
+    col = Collection(collection_name, schema)
+
+    sparse_index = {"index_type": "SPARSE_INVERTED_INDEX", "metric_type": "IP"}
+    dense_index = {"index_type": "FLAT", "metric_type": "COSINE"}
+    col.create_index("sparse_vector", sparse_index)
+    col.create_index("dense_vector", dense_index)
+
+    client.insert(collection_name)
     
     client.create_collection(
         collection_name,