enchance: refine group_by_field and provide group_by example(#1809)

Signed-off-by: MrPresent-Han <[email protected]>

examples/group_by.py

@@ -0,0 +1,63 @@
+from pymilvus import CollectionSchema, FieldSchema, Collection, connections, DataType, Partition, utility
+import random
+import numpy as np
+import secrets
+
+
+def generate_random_hex_string(length):
+    return secrets.token_hex(length // 2)
+
+
+IP = "localhost"
+connections.connect("default", host=IP, port="19530")
+
+dim = 128
+clean_exist = False
+prepare_data = True
+
+fields = [
+    FieldSchema(name="pk", dtype=DataType.INT64, is_primary=True),
+    FieldSchema(name="int64", dtype=DataType.INT64),
+    FieldSchema(name="float", dtype=DataType.FLOAT),
+    FieldSchema(name="float_vector", dtype=DataType.FLOAT_VECTOR, dim=dim),
+    FieldSchema(name="bool", dtype=DataType.BOOL),
+    FieldSchema(name="string", dtype=DataType.VARCHAR, max_length=512)
+]
+schema = CollectionSchema(fields=fields)
+collection_name = 'test_group_by_' + generate_random_hex_string(24)
+if clean_exist and utility.has_collection(collection_name):
+    utility.drop_collection(collection_name)
+
+collection = Collection(collection_name, schema=schema)
+nb = 1500
+batch_num = 3
+vectors = [[random.random() for _ in range(dim)] for _ in range(nb)]
+# insert data
+if prepare_data:
+    for i in range(batch_num):
+        data = [
+            [i for i in range(nb * i, nb * (i + 1))],
+            [i % 33 for i in range(nb)],
+            [np.float32(i) for i in range(nb)],
+            vectors,
+            [bool(random.randrange(2)) for i in range(nb)],
+            [str(i % 44) for i in range(nb * i, nb * (i + 1))],
+        ]
+        collection.insert(data)
+        print("insert data done")
+        collection.flush()
+    collection.create_index("float_vector", {"metric_type": "COSINE"})
+
+# create collection and load
+collection.load()
+batch_size = 100
+search_params = {"metric_type": "COSINE"}
+result = collection.search(vectors[:3], "float_vector", search_params, limit=batch_size, timeout=600,
+                           output_fields=["int64"]) #set up group_by_field
+
+for i in range(len(result)):
+    resultI = result[i]
+    print(f"---result{i}_size:{len(result[i])}-------------------------")
+    for j in range(len(resultI)):
+        print(resultI[j])
+    print("----------------------------")

pymilvus/client/prepare.py

@@ -622,12 +622,13 @@ def search_requests_with_expr(
                 raise ParamError(message=f"wrong type for offset, expect int, got {type(offset)}")
             search_params["offset"] = offset
 
+        group_by_field = kwargs.get(GROUP_BY_FIELD)
+        if group_by_field is not None:
+            search_params[GROUP_BY_FIELD] = group_by_field
+
         if param.get("metric_type", None) is not None:
             search_params["metric_type"] = param["metric_type"]
 
-        if param.get(GROUP_BY_FIELD, None) is not None:
-            search_params[GROUP_BY_FIELD] = param[GROUP_BY_FIELD]
-
         if anns_field:
             search_params["anns_field"] = anns_field