* started converting handles
* more changes
* sort of working for Dataset
* yay all the tests are passing for Dataset handle changes
* working for other handle types
* remove debugging logging
* remove unnecessary spaces
* fix null logic
* more NULL
* updates
* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* consolidate steps
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* [R-package] prevent symbol lookup conflicts
* add unit tests
* only run test on Windows
* move to .Call() calls
* fix references
* testing registration for CMake builds
* revert NAMESPACE changes
* revert testing changes
* [R-package] enable use of trees with linear models at leaves (fixes#3319)
* remove problematic pragmas
* fix tests
* try to fix build scripts
* try fixing pragma check
* more pragma checks
* ok fix pragma stuff for real
* empty commit
* regenerate documentation
* try skipping test
* uncomment CI
* add note on missing value types for R
* add tests on saving and re-loading booster
* add explicit return statement to functions in callback.R
* add explicit return statement to functions in lgb.Booster.R
* add explicit return statement to functions in lgb.Dataset.R
* add explicit return statement to functions in lgb.Predictor.R
* add explicit return statement to functions in lgb.cv.R
* add explicit return statement to functions in lgb.interprete.R
* add explicit return statement to functions in lgb.plot.importance.R
* add explicit return statement to functions in saveRDS.lgb.Booster.R
* add explicit return statement to functions in utils.R
* add explicit return statement to functions in lgb.plot.interpretation.R
* add explicit return statement to functions in build_r.R
* fix typo
* return(self) -> return(invisible(self))
* fix some inconsistent indention
* fix test failure
* add another return() statement
* fix linting errors
* Fix LightGBM models locale sensitivity and improve R/W performance.
When Java is used, the default C++ locale is broken. This is true for
Java providers that use the C API or even Python models that require JEP.
This patch solves that issue making the model reads/writes insensitive
to such settings.
To achieve it, within the model read/write codebase:
- C++ streams are imbued with the classic locale
- Calls to functions that are dependent on the locale are replaced
- The default locale is not changed!
This approach means:
- The user's locale is never tampered with, avoiding issues such as
https://github.com/microsoft/LightGBM/issues/2979 with the previous
approach https://github.com/microsoft/LightGBM/pull/2891
- Datasets can still be read according the user's locale
- The model file has a single format independent of locale
Changes:
- Add CommonC namespace which provides faster locale-independent versions of Common's methods
- Model code makes conversions through CommonC
- Cleanup unused Common methods
- Performance improvements. Use fast libraries for locale-agnostic conversion:
- value->string: https://github.com/fmtlib/fmt
- string->double: https://github.com/lemire/fast_double_parser (10x
faster double parsing according to their benchmark)
Bugfixes:
- https://github.com/microsoft/LightGBM/issues/2500
- https://github.com/microsoft/LightGBM/issues/2890
- https://github.com/ninia/jep/issues/205 (as it is related to LGBM as well)
* Align CommonC namespace
* Add new external_libs/ to python setup
* Try fast_double_parser fix#1
Testing commit e09e5aad828bcb16bea7ed0ed8322e019112fdbe
If it works it should fix more LGBM builds
* CMake: Attempt to link fmt without explicit PUBLIC tag
* Exclude external_libs from linting
* Add exernal_libs to MANIFEST.in
* Set dynamic linking option for fmt.
* linting issues
* Try to fix lint includes
* Try to pass fPIC with static fmt lib
* Try CMake P_I_C option with fmt library
* [R-package] Add CMake support for R and CRAN
* Cleanup CMakeLists
* Try fmt hack to remove stdout
* Switch to header-only mode
* Add PRIVATE argument to target_link_libraries
* use fmt in header-only mode
* Remove CMakeLists comment
* Change OpenMP to PUBLIC linking in Mac
* Update fmt submodule to 7.1.2
* Use fmt in header-only-mode
* Remove fmt from CMakeLists.txt
* Upgrade fast_double_parser to v0.2.0
* Revert "Add PRIVATE argument to target_link_libraries"
This reverts commit 3dd45dde7b92531b2530ab54522bb843c56227a7.
* Address James Lamb's comments
* Update R-package/.Rbuildignore
Co-authored-by: James Lamb <jaylamb20@gmail.com>
* Upgrade to fast_double_parser v0.3.0 - Solaris support
* Use legacy code only in Solaris
* Fix lint issues
* Fix comment
* Address StrikerRUS's comments (solaris ifdef).
* Change header guards
Co-authored-by: James Lamb <jaylamb20@gmail.com>
* [R-package] make package installable with CRAN toolchain (fixes#2960)
* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* remove GPU stuff
* use wildcard to find objects to build
* use -lomp
* build configure before moving files
* using wildcard for objects
* Update .github/workflows/main.yml
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* add explicit objects back
* reduce allowed R CMD check NOTEs and catch stderr from build-cran-package on Windows
* fixing things
* pin autoconf version
* show diff
* add automake back
* run less checks
* command was in the wrong place
* fix autoconf version
* change strategy for handling configure
* fix Rbuildignore
* fix NOTEs
* fix notes about unrecognized files
* fixing extra files
* remove USE_R35
* add OpenMP check for Mac CRAN build
* run all checks
* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* suggestions from code review
* undo indenting
* remove 03 from Makevars.win.in
* update language about OpenMP in configure script
* checking if configure.ac check works
* add autoconf back
* remove testing code in configure.ac
* more fixes for CI on configure script
* print git diff
* add VERSION.txt when checking configure
* fix relative paths
* remove git diff
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* [R-package] replaced gendef.exe with R code to create R.def (fixes#3064)
* fix broken things
* trying to add R4.0 docs
* changed make
* fixing make paths
* update notes on environment variables
* fix linting
* fixes to CI
* fixing build_cmd and other stuff
* fix bad substitution
* fix Azure Linux builds
* I am bad at bash
* simplifying
* only testing R
* getting better logs
* mingw32
* docs
* toolchain
* using msys
* fix visual studio condition
* toolchain test
* full CI
* fix if-elses
* bump allowed NOTEs
* search for Rscript
* updates to docs
* use processx
* fix mismatched arguments
* move CI to GitHub Actions
* minor changes
* fix workflow file
* fix templating
* fix Azure DevOps
* debugging windows builds
* dont shQuote file name
* all GitHub Actions jobs
* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* minor cleanup
* remove objdump printing
* make file.remove() invisible
* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* reduce duplicated paths in docs
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
* Added more linters on R code
* started working on implicit integers
* finished style changes to handle implicit integers
* regenned documentation and added concatenation linter
* changed channel for r-lintr
* try building stringi before lintr
* trying to get libicui18n
* trying another thing
* trying conda-forge again
* added re-install of stringi
* uncommented other stages
* Update .ci/test.sh
Co-Authored-By: Nikita Titov <nekit94-08@mail.ru>
* removed apt update and changed lintr version floor
* get lintr from CRAN
* R needs to come before C++ linting
* testing lintr install from CRAN
* trying one more thing
* more verbose
* order might matter
* removed commented code
* cleaned up linting block in test.sh
* grouped conda install calls and fixed a few integer array things
* added R-package docs generation routines
* change theme to be more consistent with sphinx_rtd_theme on main site in terms of color scheme
* placed man folder with old Rd files back
* specify full path to conda and make script more readable by one line - one pkg
* removed commented lines from build_r_site script
* made one line - one argument in build_reference() call
* pin R package versions
* fixed conflict
* Fixed typos in docs
* Fixed inconsistencies in documentation
* Updated strategy for registering routines
* Fixed issues caused by smashing multiple functions into one Rd
* Fixed issues with documentation
* Removed VignetteBuilder and updated Rbuildignore
* Added R build artefacts to gitignore
* Added namespacing on data.table set function. Updated handling of CMakeLists file to get around CRAN check.
* Updated build instructions
* Added R build script
* Removed build_r.sh script and updated R-package install instructions
James Lamb
CuriousCorrelation
Nikita Titov
Guangyu Zeng
Alberto Ferreira
Gao Tao
Laurae
Khoa Nguyen
Zach Kurtz