|
| 1 | +import tempfile |
| 2 | +import uuid |
| 3 | +from contextlib import contextmanager |
| 4 | +from dataclasses import dataclass, field |
| 5 | +from datetime import datetime, timedelta |
| 6 | +from pathlib import Path |
| 7 | +from typing import Any, Dict, List, Optional, Type, Union |
| 8 | + |
| 9 | +import pandas as pd |
| 10 | + |
| 11 | +from feast import FeatureStore, FeatureView, RepoConfig, driver_test_data |
| 12 | +from feast.data_source import DataSource |
| 13 | +from tests.integration.feature_repos.universal.data_source_creator import ( |
| 14 | + DataSourceCreator, |
| 15 | +) |
| 16 | +from tests.integration.feature_repos.universal.data_sources.bigquery import ( |
| 17 | + BigQueryDataSourceCreator, |
| 18 | +) |
| 19 | +from tests.integration.feature_repos.universal.data_sources.file import ( |
| 20 | + FileDataSourceCreator, |
| 21 | +) |
| 22 | +from tests.integration.feature_repos.universal.data_sources.redshift import ( |
| 23 | + RedshiftDataSourceCreator, |
| 24 | +) |
| 25 | +from tests.integration.feature_repos.universal.feature_views import ( |
| 26 | + create_customer_daily_profile_feature_view, |
| 27 | + create_driver_hourly_stats_feature_view, |
| 28 | +) |
| 29 | + |
| 30 | + |
| 31 | +@dataclass(frozen=True, repr=True) |
| 32 | +class IntegrationTestRepoConfig: |
| 33 | + """ |
| 34 | + This class should hold all possible parameters that may need to be varied by individual tests. |
| 35 | + """ |
| 36 | + |
| 37 | + provider: str = "local" |
| 38 | + online_store: Union[str, Dict] = "sqlite" |
| 39 | + |
| 40 | + offline_store_creator: Type[DataSourceCreator] = FileDataSourceCreator |
| 41 | + |
| 42 | + full_feature_names: bool = True |
| 43 | + infer_event_timestamp_col: bool = True |
| 44 | + infer_features: bool = False |
| 45 | + |
| 46 | + |
| 47 | +DYNAMO_CONFIG = {"type": "dynamodb", "region": "us-west-2"} |
| 48 | +REDIS_CONFIG = {"type": "redis", "connection_string": "localhost:6379,db=0"} |
| 49 | +FULL_REPO_CONFIGS: List[IntegrationTestRepoConfig] = [ |
| 50 | + # Local configurations |
| 51 | + IntegrationTestRepoConfig(), |
| 52 | + IntegrationTestRepoConfig(online_store=REDIS_CONFIG), |
| 53 | + # GCP configurations |
| 54 | + IntegrationTestRepoConfig( |
| 55 | + provider="gcp", |
| 56 | + offline_store_creator=BigQueryDataSourceCreator, |
| 57 | + online_store="datastore", |
| 58 | + ), |
| 59 | + IntegrationTestRepoConfig( |
| 60 | + provider="gcp", |
| 61 | + offline_store_creator=BigQueryDataSourceCreator, |
| 62 | + online_store=REDIS_CONFIG, |
| 63 | + ), |
| 64 | + # AWS configurations |
| 65 | + IntegrationTestRepoConfig( |
| 66 | + provider="aws", |
| 67 | + offline_store_creator=RedshiftDataSourceCreator, |
| 68 | + online_store=DYNAMO_CONFIG, |
| 69 | + ), |
| 70 | + IntegrationTestRepoConfig( |
| 71 | + provider="aws", |
| 72 | + offline_store_creator=RedshiftDataSourceCreator, |
| 73 | + online_store=REDIS_CONFIG, |
| 74 | + ), |
| 75 | +] |
| 76 | + |
| 77 | + |
| 78 | +def construct_universal_entities() -> Dict[str, List[Any]]: |
| 79 | + return {"customer": list(range(1001, 1110)), "driver": list(range(5001, 5110))} |
| 80 | + |
| 81 | + |
| 82 | +def construct_universal_datasets( |
| 83 | + entities: Dict[str, List[Any]], start_time: datetime, end_time: datetime |
| 84 | +) -> Dict[str, pd.DataFrame]: |
| 85 | + customer_df = driver_test_data.create_customer_daily_profile_df( |
| 86 | + entities["customer"], start_time, end_time |
| 87 | + ) |
| 88 | + driver_df = driver_test_data.create_driver_hourly_stats_df( |
| 89 | + entities["driver"], start_time, end_time |
| 90 | + ) |
| 91 | + orders_df = driver_test_data.create_orders_df( |
| 92 | + customers=entities["customer"], |
| 93 | + drivers=entities["driver"], |
| 94 | + start_date=end_time - timedelta(days=365), |
| 95 | + end_date=end_time + timedelta(days=365), |
| 96 | + order_count=1000, |
| 97 | + ) |
| 98 | + |
| 99 | + return {"customer": customer_df, "driver": driver_df, "orders": orders_df} |
| 100 | + |
| 101 | + |
| 102 | +def construct_universal_data_sources( |
| 103 | + datasets: Dict[str, pd.DataFrame], data_source_creator: DataSourceCreator |
| 104 | +) -> Dict[str, DataSource]: |
| 105 | + customer_ds = data_source_creator.create_data_source( |
| 106 | + datasets["customer"], |
| 107 | + destination_name="customer_profile", |
| 108 | + event_timestamp_column="event_timestamp", |
| 109 | + created_timestamp_column="created", |
| 110 | + ) |
| 111 | + driver_ds = data_source_creator.create_data_source( |
| 112 | + datasets["driver"], |
| 113 | + destination_name="driver_hourly", |
| 114 | + event_timestamp_column="event_timestamp", |
| 115 | + created_timestamp_column="created", |
| 116 | + ) |
| 117 | + orders_ds = data_source_creator.create_data_source( |
| 118 | + datasets["orders"], |
| 119 | + destination_name="orders", |
| 120 | + event_timestamp_column="event_timestamp", |
| 121 | + created_timestamp_column="created", |
| 122 | + ) |
| 123 | + return {"customer": customer_ds, "driver": driver_ds, "orders": orders_ds} |
| 124 | + |
| 125 | + |
| 126 | +def construct_universal_feature_views( |
| 127 | + data_sources: Dict[str, DataSource], |
| 128 | +) -> Dict[str, FeatureView]: |
| 129 | + return { |
| 130 | + "customer": create_customer_daily_profile_feature_view( |
| 131 | + data_sources["customer"] |
| 132 | + ), |
| 133 | + "driver": create_driver_hourly_stats_feature_view(data_sources["driver"]), |
| 134 | + } |
| 135 | + |
| 136 | + |
| 137 | +@dataclass |
| 138 | +class Environment: |
| 139 | + name: str |
| 140 | + test_repo_config: IntegrationTestRepoConfig |
| 141 | + feature_store: FeatureStore |
| 142 | + data_source_creator: DataSourceCreator |
| 143 | + |
| 144 | + end_date: datetime = field( |
| 145 | + default=datetime.now().replace(microsecond=0, second=0, minute=0) |
| 146 | + ) |
| 147 | + |
| 148 | + def __post_init__(self): |
| 149 | + self.start_date: datetime = self.end_date - timedelta(days=7) |
| 150 | + |
| 151 | + |
| 152 | +def table_name_from_data_source(ds: DataSource) -> Optional[str]: |
| 153 | + if hasattr(ds, "table_ref"): |
| 154 | + return ds.table_ref |
| 155 | + elif hasattr(ds, "table"): |
| 156 | + return ds.table |
| 157 | + return None |
| 158 | + |
| 159 | + |
| 160 | +@contextmanager |
| 161 | +def construct_test_environment( |
| 162 | + test_repo_config: IntegrationTestRepoConfig, |
| 163 | + test_suite_name: str = "integration_test", |
| 164 | +) -> Environment: |
| 165 | + project = f"{test_suite_name}_{str(uuid.uuid4()).replace('-', '')[:8]}" |
| 166 | + |
| 167 | + offline_creator: DataSourceCreator = test_repo_config.offline_store_creator(project) |
| 168 | + |
| 169 | + offline_store_config = offline_creator.create_offline_store_config() |
| 170 | + online_store = test_repo_config.online_store |
| 171 | + |
| 172 | + with tempfile.TemporaryDirectory() as repo_dir_name: |
| 173 | + config = RepoConfig( |
| 174 | + registry=str(Path(repo_dir_name) / "registry.db"), |
| 175 | + project=project, |
| 176 | + provider=test_repo_config.provider, |
| 177 | + offline_store=offline_store_config, |
| 178 | + online_store=online_store, |
| 179 | + repo_path=repo_dir_name, |
| 180 | + ) |
| 181 | + fs = FeatureStore(config=config) |
| 182 | + environment = Environment( |
| 183 | + name=project, |
| 184 | + test_repo_config=test_repo_config, |
| 185 | + feature_store=fs, |
| 186 | + data_source_creator=offline_creator, |
| 187 | + ) |
| 188 | + |
| 189 | + try: |
| 190 | + yield environment |
| 191 | + finally: |
| 192 | + fs.teardown() |
0 commit comments