magenpy 0.1.5

v0.1.5

on Python PyPI

Changed

• Updated behavior of .load of LDMatrix. Now, by default it loads
  an LDLinearOperator object.
  • Now the cached loaded data is in the form of a LDLinearOperator object.
• Moved a lot of the functionality of converting LD data to CSR format
  to the LDLinearOperator.
• Removed printing where possible in the package and changed it to use
  the logging module.
• Resolved some issues in how the pandas-plink genotype matrix is
  handled in case of splitting by chromosomes/variants.
• Removed fill_na from the standardize method in stats.transforms.genotype.
• Fixed how the package interfaces with tempfile to properly cleanup
  temporary files/directories.
• Made the tests for LDMatrix a bit more comprehensive.

Added

• Added preliminary tests to the CLI scripts (magenpy_ld and magenpy_simulate)
• Support for block iterator for the LDMatrix.
• rank_one_update for the LDLinearOperator class.
• Unified method to map variants to genomic blocks map_variants_to_genomic_blocks.
• Added summary method to LDMatrix to provide a summary of the LD matrix.
• Added __repr__ and __repr_html__ methods to LDMatrix.
• Added functionality to allow slicing of LDLinearOperator and outputting
  subsets of the data as a numpy array directly.
• Added implementation of the PUMAS procedure for sampling summary data
  conditional on the LD matrix. Relevant functions:
  • sumstats_train_test_split
  • multivariate_normal_conditional_sampling
• Added a faster intersection implementation intersect_multiple_arrays.
• Added preliminary bedReaderGenotypeMatrix to support using the bed-reader
  package as a backend (still needs more development and testing).
• Added convenience method setup_logger to set up logging in the package.