INTERNAL
src/query_all.cppDOCUMENTATION
chr1 vs 1.BUG FIXES
Fix htslib_version() to return the version as a string instead of simply
printing it to allow programmatic checks for libdeflate presence.
iscream now informs users whether they have libdeflate-enabled htslib on package load.
BUG FIXES
INTERNAL
inherits(x, "data.frame") instead of "data.frame" %in% class(x)"BREAKING CHANGES to summarize_regions() and summarize_meth_regions():
The summary output now has position as chromosome, start, and end columns
instead of just a feature column with the position string. This allows
direct conversion to GRanges objects without having to use rownames. It is
still possible to use chromosome names as the input query regions ("chr1"),
the start and end in those cases will be NA.
The feature column is now populated only if the input regions are named, if
a vector, or the feature_col argument is set to a column for
data.frame/GRanges inputs.
The count column from summarize_meth_regions() is now named count instead
of cpg_count.
These functions now return a data.table instead of a data.frame to allow
for faster in-place modifications and consistency with tabix() output. Use
data.table::setDF() to convert to a data.frame in-place.
BUG FIXES
summarize_regions() or
summarize_meth_regions() output using the input region strings as these
were repeating and not unique. It was only possible because they were set
from the C++ scope instead of R.ENHANCEMENTS
summarize_regions() and summarize_meth_regions() now support the following
new functions, inspired by bedtools map:
pstddev)first)last)mode)antimode)absmin)absmax)count_distinct)get_granges_string() can now extract names from its mcols using a
feature_col argument as in get_df_string() - names were
previously pulled only from names(). This means GRanges inputs to
summarize_regions() can have the feature_col in its mcols rather than
just as its names.
DOCUMENTATION
genes.bed file used in
vignette("performance") now contains the gene names that the vignette
references.INTERNAL
Refactored summarize_regions to collect similar validation and regions
parsing functions
tabix() writes the input regions of interest to disk only once instead of
doing it for every file
BUG FIX
DOCUMENTATION
tabix()tabix()'s' aligner and col.names documentationBUG FIXES
DOCUMENTATION
Vignettes suppress long startup messages
Updated run-times and plots in tabix vs Rsamtools vignette
BUG FIXES
tabix(), converting GRanges inputs of width 1 now returns them as
chr:start-start instead of chr:start which was invalid for the htslib API
and caused a block of records to be returned instead of just the locus.BREAKING CHANGES
DOCUMENTATION
bsseq creation now runs for the make_mat_bsseq exampleBREAKING CHANGES
make_bsseq_mat changed to make_mat_bsseq for consistency with other matrix
creating functionsBUG FIXES
Fixed MacOS and clang issues (version checking, using the correct linkage tools) in configuration script
Explicitly coerce path to string for bedfile prefix extraction in
Cpp_summarize_regions
ENHANCEMENTS
Added windows as supported platform but without multithreading support
pbapply progress bar added to tabix() functions
INTERNAL
Explicitly return result from tabix.single helper function.
Removed unused C++ variables
file.path used instead of pasting paths
DOCUMENTATION
bsseq added to Suggests to run in examples and vignettes
Changed all \dontrun to \donttest
All vignettes set to BiocStyle
BiocFileCache now used for all vignette file downloads
Reading the genes file in the "performance" vignette fixed to read only three columns
Added more information on bitpacking limits for make_mat_bsseq
Allowed creation of BSseq objects in data structures vignette
Package description updated to add information about methylation support
Replaced all references to CpG with more general methylation
Added CpH as supported in documentation
INTERNAL
SystemRequirements