yenta 0.4.0

Introduction

Yenta is YEt ‘Nother TAskrunner; it executes a pipeline, defined by a series of tasks and dependencies among them. The goal of Yenta is to provide a reasonable feature set while maintaining simplicity and usability. Yenta tasks are simply functions decorated with the @task decorator and complying with a specific format for their arguments.

Yenta is inspired in part by the functional state management pattern used in projects like Redux. Although I would hesitate to call Yenta “functional” in the strict sense, it does use a caching logic according to which identical inputs are assumed to produce identical outputs under default conditions. This means that Yenta will automatically reuse the cached output of a task if nothing about the inputs has changed.

The name “Yenta” is an old Yiddish name; in the American Yiddish theater of the 30s, a character named Yenta was depicted as a busy-body, so the name became a byword for someone who won’t mind their own business.

Features

• Graph-based task execution

• Lightweight

• Idempotent tasks

• Simple, intuitive API

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.

0.1.0 (2020-05-27)

• First release on PyPI.

• Working pipeline execution with caching and reuse of results

0.2.0 (2020-07-01)

• Selector functionality implemented

0.2.1 (2020-07-10)

• Write intermediate pipeline results to a temp file and catch the case where a non-serializable object is returned

0.2.2 (2020-07-10)

• Append running dir to sys.path

0.2.3 (2020-07-10)

• Initialize artifact date_created by default

0.2.4 (2020-07-16)

• FileArtifacts can be directories

• Artifacts now have a meta field to which arbitrary JSON-serializable information can be attached

• If an exception is raised during task execution, the stack trace is printed

• pydot is now used for graph generation

• Artifact equality is checked by comparing location and hash

0.3.0 (2020-10-31)

• Serialization scheme moved from JSON to pickle to improve performance.

• “Values” no longer explicitly exist; anything that can be pickled is a valid value.

• Multiple pipelines with different names can be used in the same project.