Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Speed Up Ray Image Tests #2828

Merged
merged 1 commit into from
Dec 7, 2022
Merged

Speed Up Ray Image Tests #2828

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 11 additions & 7 deletions tests/integration_tests/test_ray.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,6 +165,9 @@ def run_preprocessing(
dataset_type="parquet",
num_examples_per_split=20,
nan_percent=0.0,
first_row_none=False,
last_row_none=False,
nan_cols=None,
):
# Split the dataset manually to avoid randomness in splitting
split_to_df = {}
Expand All @@ -182,7 +185,8 @@ def run_preprocessing(
split_to_df[split] = dataset_df
full_df_path = os.path.join(tmpdir, "dataset.csv")
pd.concat(split_to_df.values()).to_csv(full_df_path, index=False)
dataset_path = create_data_set_to_use(dataset_type, full_df_path, nan_percent=nan_percent)
dataset = create_data_set_to_use(dataset_type, full_df_path, nan_percent=nan_percent)
dataset = augment_dataset_with_none(dataset, first_row_none, last_row_none, nan_cols)

# Configure ray backend
config = {
Expand All @@ -204,15 +208,15 @@ def run_preprocessing(
ray_model = LudwigModel(config, backend=backend_config)
*ray_datasets, ray_training_set_metadata = ray_model.preprocess(
skip_save_processed_input=False, # Save the processed input to test pyarrow write/read
dataset=dataset_path,
dataset=dataset,
)

# Run preprocessing with local backend using the ray_training_set_metadata to ensure parity of
# token assignments, etc.
local_model = LudwigModel(config, backend=LOCAL_BACKEND)
*local_datasets, _ = local_model.preprocess(
training_set_metadata=ray_training_set_metadata,
dataset=dataset_path,
dataset=dataset,
)

for ray_dataset, local_dataset in zip(ray_datasets, local_datasets):
Expand Down Expand Up @@ -280,7 +284,7 @@ def run_test_with_features(
preprocessing=None,
first_row_none=False,
last_row_none=False,
nan_cols=[],
nan_cols=None,
):
preprocessing = preprocessing or {}
config = {
Expand Down Expand Up @@ -573,12 +577,12 @@ def test_ray_image_with_fill_strategy_edge_cases(tmpdir, settings, ray_cluster_2
output_features = [
binary_feature(),
]
run_test_with_features(
run_preprocessing(
tmpdir,
"dask",
input_features,
output_features,
df_engine="dask",
dataset_type="pandas+numpy_images",
skip_save_processed_input=False,
first_row_none=first_row_none,
last_row_none=last_row_none,
nan_cols=[input_features[0][NAME]],
Expand Down
6 changes: 4 additions & 2 deletions tests/integration_tests/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@
import unittest
import uuid
from distutils.util import strtobool
from typing import List, Union
from typing import List, Optional, Union

import cloudpickle
import numpy as np
Expand Down Expand Up @@ -824,7 +824,7 @@ def to_fwf(df, fname):


def augment_dataset_with_none(
df: pd.DataFrame, first_row_none: bool = False, last_row_none: bool = False, nan_cols: List = []
df: pd.DataFrame, first_row_none: bool = False, last_row_none: bool = False, nan_cols: Optional[List] = None
) -> pd.DataFrame:
"""Optionally sets the first and last rows of nan_cols of the given dataframe to nan.

Expand All @@ -837,6 +837,8 @@ def augment_dataset_with_none(
:param nan_cols: a list of columns in the dataframe to explicitly set the first or last rows to np.nan
:type nan_cols: list
"""
nan_cols = nan_cols if nan_cols is not None else []

if first_row_none:
for col in nan_cols:
df.iloc[0, df.columns.get_loc(col)] = np.nan
Expand Down