Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Browse files
Browse the repository at this point in the history
* Add tweedie package * fix pyarrow typo * Increase the DEFAULT_IMAGE_VERSION_TAG in scripts/jenkins/groovy/buildConfig.groovy * Add missing quotes to test-requirements.txt * Increase DEFAULT_IMAGE_VERSION_TAG * Fix more typos * specify python versions for packages we had issues in the build process * compile python3.6 Changed version of wheel as 0.38.0 was yanked out of pypi due to circular dependency * Add htmlTable for Hmisc in R * Install old pkgs * Update Dockerfile-build-base * fix issues with nodesource setup script in python and jdk-other images * Extract NODE_MAJOR env var ouside of the complicated RUN statements * Add --no-tty to gpg call to resolve issue with non-existing /dev/tty * Enable Tweedie test * Use make install instead of make altinstall when installing python 3.6 to solve issue with virtualenv * Increase default build python version to 3.7 * Update R library path due to new OS version * Install chromium directly (apt now installs just a file that tells you to install it using snap which requires running systemd in the docker) * Update python 3.6 pkgs to resolve statsmodels and scipy version incompatibility * update R tests * Add new error message for cacert test * Deal with lifelines >= 0.27 incompatibility with py < 3.9 * Add more debugging info to the Tweedie test * Fix runit_pubdev_8683_glm_gamma_dispersion_factor_large.R * fix matplotlib issue in python demos and add more debugging info to tweedie ml dispersion estimation * Fix h2o-py/demos/uplift_random_forest_compare_causalml.ipynb by removing the treatment from test set during prediction * Make use of datatable and polars opt-in not opt-out * Make r tweedie dispersion test more reproducible * Fix logical error in an exceptional case in Tweedie dispersion estimation * log more information about Tweedie dispersion estimation * Increase tolerance in pyunit_PUBDEV_6117_xgboost_compare.py * Disable Newton's methon in dispersion estimation in GLM Tweedie * Fix pubdev_8685_tweedie_dispersion_factor_exceed2 * Increase tolerance in pyunit_PUBDEV_6117_xgboost_compare.py to 1e-6 * Fix models.R comments (roxygen complains) * Fix tweedie dispersion estimation for variance power >= 2 with response == 0 * Fix long line in R comment for CRAN checks * Move model_pearson back where it used to be in tweedie dispersion factor test * Remove an unnecessary space * Fix R3.5.3 init java 11 test * Make runit_pubdev_8683_glm_gamma_dispersion_factor_large.R more robust - if R GLM fails to converge don't fail the test * Remove packages for python < 3.5 from test-requirements.txt * Try to fix R issues when building on jenkins Basically, some files are not used in the final package probably due to some non-deterministic issue in roxygen2 which fills our NAMESPACE file and creates our R man pages. * Fix 'The process cannot access the file because it is being used by another process' on windows in python * Make sure R images contain all the suggested packages * Update roxygen and rlang in R images * Update pkgdown * Add mypy and specify the explicit version of already present typing_extensions * Increase timeout in stages that are too often aborted The R medium stages didn't run tests that contained missing package so it's understandable that the runtime there increased. Java JUnit tests also are aborted very often in master so this PR likely doesn't cause the issue.
- Loading branch information