Rewrote key functions in C to provide significant performance boost
(4-5x speedup in certain tests)
get_matches and filter_overlap are the
primary performance enhancers
get_matches_base and filter_overlap_base
retain the original base R functionality in case of issues with building
the code
Changed maintainer email address because current one was bouncing
CRAN emails
mgsub 1.7.3
Vignettes were using the qdap package which was not available
causing errors. This update only modifies the vignette to remove the use
of qdap (and microbenchmark) and so these were removed from the
‘suggests’.
mgsub 1.7.2
There was a warning surfacing in newer versions of R around how
certain functions were linked to in documentation. While not technically
a bug, our CICD pipeline is setup to report those as errors. To ensure
full compliance with CRAN Standards, the linking was updated.
Very minor code refactor to apply linting standards to enable easier
contribution, removed some superfulous lines.
mgsub 1.7.1
Added RNG Version control to the tests to support future versions of
R. There are no functional changes to the way the package runs, this is
only to ensure the testing environment continues to work as
expected.
mgsub 1.7
Added the mgsub_censor function which enables fast, simultaneous
censoring of patterns
Trimmed unnecessary logic and helper functions
mgsub 1.6
New method employed leveraging gregexpr to reduce the amount of
computation time spend identifying matches which slightly improves
performance.
Added ‘escapes’ from the normal processing if there’s only one match
to replace to further improve performance.
mgsub 1.5
Removed NSE evalutaion, deprecating the named list (dictionary)
method
Added support for vector inputs
Added NA handling
Slightly improved vectorized performance
Updated license (MIT)
mgsub 1.0.1
Added checks for named input to mgsub
Added mgsub_alt to support alternative function parameterization
(name is temporary)