memory_repro.py

from evalml.automl import AutoMLSearch
from evalml.preprocessing import split_data
import pandas as pd

X = pd.read_csv("/Users/freddy.boulton/Downloads/nyc_taxi.csv")
y = X.pop("trip_duration")

X_train, X_validation, y_train, y_validation = split_data(
    X, y, problem_type="regression", test_size=0.5, random_seed=0
)

# Search for best pipelines
automl = AutoMLSearch(
    X_train=X_train,
    y_train=y_train,
    max_batches=1,
    problem_type="regression",
    train_best_pipeline=False,
)
automl.search()

# Train best pipelines on training dataset
pipelines = {
    pipeline_id: automl.get_pipeline(pipeline_id) for pipeline_id in automl.rankings.id
}
trained_pipeline_dic = automl.train_pipelines(pipelines.values())

# Score best pipelines on validation dataset
validation_result = automl.score_pipelines(
    trained_pipeline_dic.values(), X_validation, y_validation, [automl.objective]
)