eexp-engine-utils 0.0.42

Helper utilities for eexp_engine - supporting ProActive and Kubeflow execution environments

eexp_engine_utils

Helper utilities for the ExtremeXP Experimentation Engine that automatically adapt to different execution environments (Kubeflow, ProActive, Local).

Overview

This package provides a transparent proxy that automatically routes function calls to the appropriate implementation based on your execution environment. No more manual environment checking or conditional imports!

Installation

pip install -e exp_engine_utils

Quick Start

from eexp_engine_utils import utils

# Load a dataset - automatically works in any environment
data = utils.load_dataset(variables, resultMap, "input_data")

# Process your data
processed_data = your_processing_function(data)

# Save the result - automatically works in any environment
utils.save_dataset(variables, resultMap, "output_data", processed_data)

That's it! The utils proxy automatically detects your execution environment and routes calls to the correct implementation.

How It Works

The package uses a transparent proxy pattern that automatically determines which utils module to use:

• Kubeflow/Local: Routes to kubeflow_utils (reads EXECUTIONWARE environment variable)
• ProActive: Routes to proactive_utils (reads execution_engine_runtime_config_*.json file)

You write your task code once, and it works everywhere.

Key Features

• ✅ Zero configuration - Works automatically based on runtime environment
• ✅ Clean API - Just import utils and go
• ✅ Portable - Same task code runs on Kubeflow, ProActive, and Local
• ✅ Simple migration - Replace conditional imports with one line

Available Functions

Dataset Management

• load_dataset(variables, resultMap, key) - Load a single dataset
• load_datasets(variables, resultMap, key) - Load multiple datasets
• save_dataset(variables, resultMap, key, value) - Save a single dataset
• save_datasets(variables, resultMap, key, values, file_names) - Save multiple datasets

Helpers

• get_experiment_results(variables) - Get experiment results
• load_dataset_by_path(file_path) - Load from specific path
• load_pickled_dataset_by_path(file_path) - Load pickled data
• And more environment-specific helpers...

Documentation

See USAGE_EXAMPLE.md for detailed usage examples and migration guide.

Requirements

• Python >= 3.8
• requests >= 2.25.0
• fsspec >= 2021.0.0
• s3fs >= 2021.0.0
• minio >= 7.0.0

License

Apache License 2.0

Links

• Homepage: https://github.com/extremexp-HORIZON/extremexp-experimentation-engine
• Issues: https://github.com/extremexp-HORIZON/extremexp-experimentation-engine/issues