training.py

import warnings
from datetime import datetime

from tabulate import tabulate
warnings.simplefilter(action='ignore', category=FutureWarning)
import pandas as pd

from feast import FeatureStore

# Note: see https://docs.feast.dev/getting-started/concepts/feature-retrieval for 
# more details on how to retrieve for all entities in the offline store instead
entity_df = pd.DataFrame.from_dict(
    {
        # entity's join key -> entity values
        "driver_id": [1001, 1002, 1003],
        # "event_timestamp" (reserved key) -> timestamps
        "event_timestamp": [
            datetime(2021, 4, 12, 10, 59, 42),
            datetime(2021, 4, 12, 8, 12, 10),
            datetime(2021, 4, 12, 16, 40, 26),
        ],
        # (optional) label name -> label values. Feast does not process these
        "label_driver_reported_satisfaction": [1, 5, 3],
        # values we're using for an on-demand transformation
        #"val_to_add": [1, 2, 3],
        #"val_to_add_2": [10, 20, 30],
    }
)

store = FeatureStore(repo_path="feature_repo")

training_df = store.get_historical_features(
    entity_df=entity_df,
    features=[
        "driver_hourly_stats:conv_rate",
        "driver_hourly_stats:acc_rate",
        "driver_hourly_stats:avg_daily_trips",
        #"transformed_conv_rate:conv_rate_plus_val1",
        #"transformed_conv_rate:conv_rate_plus_val2",
    ],
).to_df()

print("\n----- Retrieved features -----\n")
tabulated_result = tabulate(training_df, headers='keys', tablefmt='psql')
print(f"{tabulated_result}")