NEWS

lightr 1.8.0

lr_parse_irr8(), lr_parse_rfl8() and lr_parse_raw8() now error if you provide a file not produced by AvaSoft 8.2 as they have not been tested properly with other versions of AvaSoft 8.

lightr now depends on R >= 4.0.0, following the tidyverse recommendation
the future package is now explicitly listed as a dependency, removing an R CMD check NOTE and thus removing the strain on CRAN reviewers. The future package was already a indirect dependency via the future.apply package.

this project now uses lintr to ensure the code is always following the current best coding practices in the R community
ensure floating precision issues are not causing tests to fail on CRAN (@Bisaloo, #156)

Errors in low-level parsers are now passed as warnings in high-level lr_get_XXX() functions instead of being completely silenced.
IRR8 (irradiance files produced by AvaSoft 8) are now explicitly supported by lr_get_spec(). An alias has been added for the low-level parser: lr_parse_irr8().

The date returned in metadata (by lr_get_metadata(), lr_convert_tocsv(metadata = TRUE) and lr_parse_XXXX()) is now a datetime (with UTC as timezone and format %Y-%m-%d %H:%M:%S). The column name in lr_get_metadata() has been updated from date to datetime to reflect this. Thanks to Giancarlo Chiarenza for the report, and to Hao Ye, Laura DeCicco and Elin Waring for helpful comments about timezones and datetime formatting for reproducibility.

datetime parsing is supported for more formats
Files produced by OceanInsight software (such as SpectraSuite) in Spanish can now be parsed)

the cores argument in lr_get_spec(), lr_get_metadata(), and lr_convert_tocsv() has been completely removed. It was already deprecated since lightr 1.0 (released on CRAN on 2020-01-27)
lr_convert_tocsv() gains a new metadata argument (defaults to TRUE) to determine if metadata should be exported in a csv file as well alongside the spectral data.

Output of the low-level parsers lr_parse_XXXX() is now a named list with elements data and metadata
lr_parse_raw8() and lr_parse_rfl8() now explicitly mention the number of spectra in the warnings instead of the generic "multiple spectra"
lr_parse_generic() now makes sure that the data is ordered by increasing wavelengths, which fixes a bug reported by @itamshab

(mostly internal) compute_processed() function is now named lr_compute_processed()
disable hash tests on Solaris (the output is still checked by other tests)

fixed tests on platform with no long-doubles ('noLD')
restored tests on 32bits machines
spec_ID extraction from Avantes exported files (ttt and trt) is now more robust, meaning it should work for more files.

date column in metadata is now always formatted as ISO 8601.
lightr can now import AvaSoft8 files (test files provided by M.D. Shawkey and L. Swierk), via the functions lr_parse_rfl8()/lr_parse_raw8().
lightr can correctly imports TRM files from AvaSoft 6.0 (previously it only supported files from AvaSoft 7.0).
lightr can now import binary .spc files (via the lr_parse_spc() parser). This format is used by OceanInsight and CRAIC.

new test suite on a different locale (in this case fr_FR.UTF-8) to ensure parsing is locale-independent.
warnings on CRAN build system for platforms that don't support markdown 2 have been fixed.
new, stricter tests for various file formats.
jdx files saved in a locale that uses , as the decimal separator are now parsed correctly.
Avantes exported files in non-English locales (ttt and trt files) are now parsed correctly again (this was a regression compared to pavo's pavo::getspec()). Thanks to A. Fargevieille for reporting the issue and providing a test file.

parallel processing now relies on the future package, which offers windows and high performance computing (HPC) environments support. The progress bar is produced by the progressr package and can be customised as well.
lr_parse_generic() (and thus lr_get_spec()) now works with non-UTF8 files (this was a regression compared to pavo's pavo::getspec()).
lr_get_spec() now has a new interpolate argument to determine whether you want your data interpolated and pruned at every nm or not.

readBin() (in the binary Avantes parser) now has an explicit endian value, making this package portable to platform that use big endians.
documentation has been updated to follow the rebranding of OceanOptics into OceanInsight.
vignette chunks that use pavo now only run if the package is available.
savetime field is now extracted from converted avantes files (when available).