-
Notifications
You must be signed in to change notification settings - Fork 1.3k
Expand file tree
/
Copy pathtest_compute_dag.py
More file actions
224 lines (190 loc) · 6.46 KB
/
Copy pathtest_compute_dag.py
File metadata and controls
224 lines (190 loc) · 6.46 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
from datetime import timedelta
from unittest.mock import MagicMock
import pytest
from pyspark.sql import DataFrame
from tqdm import tqdm
from feast import BatchFeatureView, Field
from feast.aggregation import Aggregation
from feast.infra.common.materialization_job import (
MaterializationJobStatus,
MaterializationTask,
)
from feast.infra.compute_engines.spark.compute import SparkComputeEngine
from feast.infra.offline_stores.contrib.spark_offline_store.spark import (
SparkOfflineStore,
)
from feast.infra.offline_stores.contrib.spark_offline_store.spark_source import (
SparkSource,
)
from feast.types import Float32, Int32, Int64
from tests.component.spark.utils import (
_check_offline_features,
_check_online_features,
create_entity_df,
create_feature_dataset,
create_spark_environment,
driver,
now,
)
def create_base_feature_view(source):
return BatchFeatureView(
name="hourly_driver_stats",
entities=[driver],
schema=[
Field(name="conv_rate", dtype=Float32),
Field(name="acc_rate", dtype=Float32),
Field(name="avg_daily_trips", dtype=Int64),
Field(name="driver_id", dtype=Int32),
],
online=True,
offline=True,
source=source,
)
def create_agg_feature_view(source):
return BatchFeatureView(
name="agg_hourly_driver_stats",
entities=[driver],
schema=[
Field(name="conv_rate", dtype=Float32),
Field(name="acc_rate", dtype=Float32),
Field(name="avg_daily_trips", dtype=Int64),
Field(name="driver_id", dtype=Int32),
],
online=True,
offline=True,
source=source,
aggregations=[
Aggregation(column="conv_rate", function="sum"),
Aggregation(column="acc_rate", function="avg"),
],
)
def create_chained_feature_view(base_fv: BatchFeatureView):
def transform_feature(df: DataFrame) -> DataFrame:
df = df.withColumn("conv_rate", df["conv_rate"] * 2)
df = df.withColumn("acc_rate", df["acc_rate"] * 2)
return df
return BatchFeatureView(
name="daily_driver_stats",
entities=[driver],
udf=transform_feature,
udf_string="transform",
schema=[
Field(name="conv_rate", dtype=Float32),
Field(name="driver_id", dtype=Int32),
],
online=True,
offline=True,
source=base_fv,
sink_source=SparkSource(
name="daily_driver_stats_sink",
path="/tmp/daily_driver_stats_sink",
file_format="parquet",
timestamp_field="event_timestamp",
created_timestamp_column="created",
),
)
@pytest.mark.integration
def test_spark_dag_materialize_recursive_view():
spark_env = create_spark_environment()
fs = spark_env.feature_store
registry = fs.registry
source = create_feature_dataset(spark_env)
base_fv = create_base_feature_view(source)
chained_fv = create_chained_feature_view(base_fv)
def tqdm_builder(length):
return tqdm(total=length, ncols=100)
try:
fs.apply([driver, base_fv, chained_fv])
# 🧪 Materialize top-level view; DAG will include base_fv implicitly
task = MaterializationTask(
project=fs.project,
feature_view=chained_fv,
start_time=now - timedelta(days=2),
end_time=now,
tqdm_builder=tqdm_builder,
)
engine = SparkComputeEngine(
repo_config=spark_env.config,
offline_store=SparkOfflineStore(),
online_store=MagicMock(),
registry=registry,
)
jobs = engine.materialize(registry, task)
# ✅ Validate jobs ran
assert len(jobs) == 1
assert jobs[0].status() == MaterializationJobStatus.SUCCEEDED
_check_online_features(
fs=fs,
driver_id=1001,
feature="daily_driver_stats:conv_rate",
expected_value=1.6,
full_feature_names=True,
)
entity_df = create_entity_df()
_check_offline_features(
fs=fs, feature="hourly_driver_stats:conv_rate", entity_df=entity_df, size=2
)
finally:
spark_env.teardown()
@pytest.mark.integration
def test_spark_dag_materialize_multi_views():
spark_env = create_spark_environment()
fs = spark_env.feature_store
registry = fs.registry
source = create_feature_dataset(spark_env)
base_fv = create_base_feature_view(source)
chained_fv = create_chained_feature_view(base_fv)
multi_view = BatchFeatureView(
name="multi_view",
entities=[driver],
schema=[
Field(name="driver_id", dtype=Int32),
Field(name="daily_driver_stats__conv_rate", dtype=Float32),
Field(name="daily_driver_stats__acc_rate", dtype=Float32),
],
online=True,
offline=True,
source=[base_fv, chained_fv],
sink_source=SparkSource(
name="multi_view_sink",
path="/tmp/multi_view_sink",
file_format="parquet",
timestamp_field="daily_driver_stats__event_timestamp",
created_timestamp_column="daily_driver_stats__created",
),
)
def tqdm_builder(length):
return tqdm(total=length, ncols=100)
try:
fs.apply([driver, base_fv, chained_fv, multi_view])
# 🧪 Materialize multi-view
task = MaterializationTask(
project=fs.project,
feature_view=multi_view,
start_time=now - timedelta(days=2),
end_time=now,
tqdm_builder=tqdm_builder,
)
engine = SparkComputeEngine(
repo_config=spark_env.config,
offline_store=SparkOfflineStore(),
online_store=MagicMock(),
registry=registry,
)
jobs = engine.materialize(registry, task)
# ✅ Validate jobs ran
assert len(jobs) == 1
assert jobs[0].status() == MaterializationJobStatus.SUCCEEDED
_check_online_features(
fs=fs,
driver_id=1001,
feature="multi_view:daily_driver_stats__conv_rate",
expected_value=1.6,
full_feature_names=True,
)
entity_df = create_entity_df()
_check_offline_features(
fs=fs, feature="hourly_driver_stats:conv_rate", entity_df=entity_df, size=2
)
finally:
spark_env.teardown()