Skip to content

Releases: BesenbacherLab/lionheart

lionheart 1.1.1

09 Dec 17:36
Compare
Choose a tag to compare
  • Improves CLI documentation for some commands (in --help pages).

lionheart 1.1.0

07 Dec 13:47
Compare
Choose a tag to compare

This release adds multiple CLI commands that:

  1. allow reproducing results from the article and seeing the effect of adding your own datasets:
  • Adds lionheart cross_validate command. Perform nested leave-one-dataset-out cross-validation on your dataset(s) and/or the included features.
  • Adds lionheart validate command. Validate a model on the included external dataset or a custom dataset.
  • Adds lionheart evaluate_univariates command. Evaluate each feature (cell-type) separately on your dataset(s) and/or the included features.
  1. expands what you can do with your own data:
  • Adds lionheart customize_thresholds command. Calculate the ROC curve and probability densities (for deciding probability thresholds) on your data and/or the included features for a custom model or an included model. Allows using probability thresholds suited to your own data when using lionheart predict_sample and lionheart validate.
  • Adds --custom_threshold_dirs argument in lionheart predict_sample. Allows passing the ROC curves and probability densities extracted with lionheart customize_thresholds.

Also:

  • Adds matplotlib as dependency.
  • Bumps generalize dependency requirement to 0.2.1.
  • Bumps utipy dependency requirement to 1.0.3.

lionheart 1.0.2

02 Dec 18:06
Compare
Choose a tag to compare
  • Fixes bug when training model on a single dataset.
  • Adds tests for a subset of the CLI tools.

lionheart 1.0.1

01 Dec 01:43
Compare
Choose a tag to compare
  • Fixed model name.

lionheart 1.0.0

26 Nov 20:37
Compare
Choose a tag to compare

Release for first submission and preprint.

Current CLI functions are:

  • lionheart extract_features
  • lionheart predict_sample
  • lionheart collect
  • lionheart train_model