Skip to content

Latest commit

 

History

History
177 lines (133 loc) · 5.63 KB

CHANGELOG.md

File metadata and controls

177 lines (133 loc) · 5.63 KB

% Changelog

Unreleased

Added

  • Tests for plot module

Fixed

  • Import pandas in plot module
  • Use type pathlib.Path for --out-{dendrogram,clusters}
  • Parameter use_names in make_label_mapper

0.6.0 - 2023/03/25

Added

  • Add basic summary statistics for pairwise distances
    • Calculate basic summary statistics for pairwise distances, dump to file
  • Add __all__ to __init__.py
  • Add _version.py, _plot.py
  • Add tests for CLI entry point, etc: test_basic.py, test_cli.py, test_clustering.py, test_rwrtoolkit.py, test_sample_data.py
  • Split up functional_partitioning module: _cluster.py, _datasets.py, _metrics.py, _plot.py, _rwrtoolkit.py, _utils.py, _version.py, functional_partitioning.py -> cli.py
  • Makefile
  • get_scores_vs_ranks_curve function for use with get_elbow
  • --no-clusters option
  • Examples:
    • Add example of --cut-method and --cut-threshold
    • Add example using RWRtoolkit

Changed

  • Allow passing --cut-method=none to return all clusterings
  • Clarify setting out_{clusters,dendrogram} variables
  • Move calc_chi function to _metrics.py
  • Move calc_threshold function to _cluster.py
  • Move get_elbow function to _metrics.py
  • Move plotting functions to _plot.py
  • Rename functional_partitioning.py to cli.py
  • Rename get_top_ranked function
  • Restructure package: prefix modules with underscore
  • Replace --threshold with --cut-threshold and --cut-method
  • Use __version__ as string instead of get_version
  • Wrap _plot_dendrogram_{polar,rectangular} w/ draw_dendrogram
  • Create the output directory (but not its parents)
  • Attempt to clarify installation instructions
  • Refactor tests (use class unittest.TestCase)
  • Include dynamicTreeCut in environment.yml for conda installations
  • Rank transformation of scores: use method='first'
  • Require dissimilarities as a condensed distance matrix
  • Adjust dendrogram figsize depending on orientation
  • Pass optimal_ordering to scipy's hierarchy.linkage fxn
  • Set min_cluster_size_ to 1 instead of 3
    • Dynamic tree cut tries to assign genes that really should be singletons to clusters. This minimizes that behavior, but it's still a problem.
  • Use __version__ as string instead of get_version
  • Wrap _plot_dendrogram_{polar,rectangular} w/ draw_dendrogram
  • Clarify setting out_{clusters,dendrogram} variables
  • Move calc_threshold fxn to cluster
  • Move calc_chi fxn to metrics
  • Move get_elbow fxn to metrics
  • Use the HierarchicalClustering class with dynamicTreeCut (cutreeHybrid) as the default clustering method.

Removed

  • Remove get_top_ranked fxn

Fixed

  • Get version from _version.__version__
  • Install seaborn
  • import dynamicTreeCut from dynamicTreeCut
  • Adhere to standard test name conventions
  • Restructure package: prefix modules with underscore to avoid 'flat-layout' error: 
    error: Multiple top-level modules discovered in a flat-layout:
['metrics', 'datasets', 'cluster', 'functional_partitioning', 'rwrtoolkit'].
    • Import modules with leading underscores
  • Calculate the elbow using mean of scores-vs-ranks vectors
    • Bug in elbow calculation introduced during recent refactoring.
    • Sort mean scores before calculating elbow
  • Modify package structure for install
  • Write clusters to clusters.tsv file

0.5.1 - 2023/03/08

Added

  • metric parameter was added in sklearn AgglomerativeClustering (upstream).
  • dynamicTreeCut dependency for HierarchicalClustering
  • check_symmetry function to verify square (uncondensed) distance matrix

Changed

  • Require scikit-learn>=1.2.0
  • HierarchicalClustering : Use scipy.cluster.hierarchy.{linkage,cut_tree} to create labels_.
  • HierarchicalClustering : Use metric parameter instead of affinity.
  • HierarchicalClustering : Use dynamicTreeCut.cutreeHybrid as default method for creating clusters.

Deprecated

  • affinity parameter was decrepated in sklearn AgglomerativeClustering (upstream).

0.4.0 - 2023/01/09

Initial release.

Added

  • CHANGELOG.md
  • cluster.HierarchicalClustering
  • RWRtoolkit wrapper functions (rwrtoolkit.rwr_singletons)

Changed

  • Split loading 'fullranks' and converting it to 'X' matrix into separate functions

See also: