evalutils 0.1.11

evalutils helps challenge administrators easily create evaluation containers for grand-challenge.org.

evalutils

evalutils helps challenge administrators easily create evaluation containers for grand-challenge.org.

• Free software: MIT license

• Documentation: https://evalutils.readthedocs.io.

Features

• Generation your challenge evaluation project boilerplate using Cookiecutter

• Scripts to build, test and export your generated docker container for grand-challenge.org

• Loading of CSV, ITK and Pillow compatible prediction files

• Validation of submitted predictions

• Interface to SciKit-Learn metrics and Pandas aggregations

• Bounding box annotations with Intersection Over Union calculations

Getting Started

evalutils requires Python 3.6, and can be installed from pip. Please see the Getting Started documentation for more details.

Credits

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

History

0.1.11 (2018-05-15)

• Breaking change: Renamed Evaluation to ClassificationEvaluation

• Adds support for DetectionEvaluation

0.1.10 (2018-04-19)

• Simplifies the example template

0.1.9 (2018-04-19)

• Breaking change: Renamed bb2 to other in BoundingBox()

• ground_truth_path is no longer a required argument

0.1.8 (2018-04-18)

• Fixes template folder in distribution

0.1.7 (2018-04-18)

• Adds cookiecutter templating for generating new projects

• Adds equality check for BoundingBox

0.1.6 (2018-03-30)

• Improves pandas csv handling

0.1.5 (2018-03-30)

• Corrects loading of some CSV files

• Adds logging and more tests

• Adds referencing to _ground_truth and _prediction in joined pandas tables

0.1.3 (2018-03-29)

• Adds basic implementation with

  • Full Evaluation workflow

  • CSV, SimpleITK, and ImageIO loaders

  • BoundingBox annotations with intersection, union and intersection over union metrics

  • Unique File Indices, Unique Image, Expected Column Names and Number of cases validators

0.1.0 (2018-03-22)

• First release on PyPI.