test: add test case for bulkwriter

tests/python_client/common/bulk_insert_data.py

@@ -24,13 +24,13 @@
 class DataField:
     pk_field = "uid"
     vec_field = "vectors"
-    float_vec_field = "float_vectors"
+    float_vec_field = "float32_vectors"
     sparse_vec_field = "sparse_vectors"
     image_float_vec_field = "image_float_vec_field"
     text_float_vec_field = "text_float_vec_field"
     binary_vec_field = "binary_vec_field"
-    bf16_vec_field = "bf16_vec_field"
-    fp16_vec_field = "fp16_vec_field"
+    bf16_vec_field = "brain_float16_vec_field"
+    fp16_vec_field = "float16_vec_field"
     int_field = "int_scalar"
     string_field = "string_scalar"
     bool_field = "bool_scalar"
@@ -504,16 +504,16 @@ def gen_data_by_data_field(data_field, rows, start=0, float_vector=True, dim=128
     data = []
     if rows > 0:
         if "vec" in data_field:
-            if "float" in data_field:
+            if "float" in data_field and "16" not in data_field:
                 data = gen_vectors(float_vector=True, rows=rows, dim=dim)
                 data = pd.Series([np.array(x, dtype=np.dtype("float32")) for x in data])
             elif "sparse" in data_field:
                 data = gen_sparse_vectors(rows, sparse_format=sparse_format)
                 data = pd.Series([json.dumps(x) for x in data], dtype=np.dtype("str"))
-            elif "fp16" in data_field:
+            elif "float16" in data_field:
                 data = gen_fp16_vectors(rows, dim)[1]
                 data = pd.Series([np.array(x, dtype=np.dtype("uint8")) for x in data])
-            elif "bf16" in data_field:
+            elif "brain_float16" in data_field:
                 data = gen_bf16_vectors(rows, dim)[1]
                 data = pd.Series([np.array(x, dtype=np.dtype("uint8")) for x in data])
             elif "binary" in data_field:
@@ -758,10 +758,10 @@ def gen_npy_files(float_vector, rows, dim, data_fields, file_size=None, file_num
                 if "binary" in data_field:
                     float_vector = False
                     vector_type = "binary"
-                if "bf16" in data_field:
+                if "brain_float16" in data_field:
                     float_vector = True
                     vector_type = "bf16"
-                if "fp16" in data_field:
+                if "float16" in data_field:
                     float_vector = True
                     vector_type = "fp16"
 

tests/python_client/common/common_func.py

@@ -2114,19 +2114,19 @@ def gen_fp16_vectors(num, dim):
     return raw_vectors, fp16_vectors
 
 
-def gen_sparse_vectors(nb, dim):
-    """
-    generate sparse vector data
-    return sparse_vectors
-    """
+def gen_sparse_vectors(nb, dim=1000, sparse_format="dok"):
+    # default sparse format is dok, dict of keys
+    # another option is coo, coordinate List
+
     rng = np.random.default_rng()
-    entities = [
-        {
-         d: rng.random() for d in random.sample(range(dim), random.randint(1, 1))
-        }
-        for _ in range(nb)
-    ]
-    return entities
+    vectors = [{
+        d: rng.random() for d in random.sample(range(dim), random.randint(20, 30))
+    } for _ in range(nb)]
+    if sparse_format == "coo":
+        vectors = [
+            {"indices": list(x.keys()), "values": list(x.values())} for x in vectors
+        ]
+    return vectors
 
 
 def gen_vectors_based_on_vector_type(num, dim, vector_data_type):

tests/python_client/requirements.txt

@@ -12,8 +12,8 @@ allure-pytest==2.7.0
 pytest-print==0.2.1
 pytest-level==0.1.1
 pytest-xdist==2.5.0
-pymilvus==2.5.0rc31
-pymilvus[bulk_writer]==2.5.0rc31
+pymilvus==2.5.0rc45
+pymilvus[bulk_writer]==2.5.0rc45
 pytest-rerunfailures==9.1.1
 git+https://github.com/Projectplace/pytest-tags
 ndg-httpsclient