Releases: deeptools/HiCMatrix
Releases · deeptools/HiCMatrix
April 2024 Release
December 2023 Release
No change in code but a lot of linting and compatibility with pypi.
Edit: This version is incompatible with python version 3.7
January 2023 Release
Fix compatibility with numpy 1.24
July 2021 release
- Adding write support for hicPro files
- Fixing a bug in for cool files to adopt the reference genome information (deeptools/HiCExplorer#657)
- Moving CI testing from Travis to Microsoft Azure
- Adding a skip triu option to compute z-scores
- Fixing a potential bug computing cut_intervals positions: #39 Thanks @lldelisle
October 2020 release
- Fixing a major bug concerning the application of the correction factors for cooler files. In version 14 it was not applied, version 15 fixes this.
- A test case for this was added and another test case was corrected.
- New function get_chromosome_sizes_real to get the size of a chromosome as it is stored in the matrix i.e. it considers if the start or the end of the chromosome is pruned.
Summer 2020
- Adding scool support
- Add support for cool files to load only the csr_matrix with instances, features and data, bypassing all 'expensive' operation like to create the cut_intervals.
May 2020 release
Bug fix:
- load and store cool file created via hic2cool and additional correction factor application.
- str vs bytes bug. Thanks @dmalzl
- remove internal scaling in cool load/store which was more a workaround for the bug fixed in 1
- Fix #23 , #25
New:
- If one chromosome or region is given cool files can now load interactions only within a distance. This is reducing the amount memory.
- Homer files can now be gzipped too. #9
March 2020 release
- Make
getBinSize
working with only one bin - Remove last indices from
nan_bins
list - change all
exit
toexception
- fix the behaviour if a non-existing region is called.
August 2019
- Fix a bug in cooler if a csr-matrix contains no data
- Remove python2 compatibility layers (six, past, future, builtins)
- Add python 3.7 to travis testing
- Change np.matrix to np.array due to deprecation warnings
Mid 2019 release
- Automatic scaling if corrected value range is not matching original value range (happens usually with cooler balanced matrices)
- Optimized cooler load to use less memory
- Added option to save a cool file with uncorrected raw data and correction factors. Before the data needed to be corrected and then the correction needed to be reversed.