Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[R] Check warnings explicitly for model compatibility tests (#6114)

Browse Source 
* [R] Check warnings explicitly for model compatibility tests

* Address reviewer's feedback
This commit is contained in:
Philip Hyunsu Cho 2020-09-15 10:49:48 -07:00 committed by GitHub
parent 33577ef5d3
commit 9e955fb9b0
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 30 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
R-package/tests/testthat/test_model_compatibility.R
View File

@ -61,7 +61,7 @@ test_that("Models from previous versions of XGBoost can be loaded", {
zipfile <- file.path(getwd(), file_name) zipfile <- file.path(getwd(), file_name)
model_dir <- file.path(getwd(), 'models') model_dir <- file.path(getwd(), 'models')
download.file(paste('https://', bucket, '.s3-', region, '.amazonaws.com/', file_name, sep = ''), download.file(paste('https://', bucket, '.s3-', region, '.amazonaws.com/', file_name, sep = ''),
destfile = zipfile, mode = 'wb') destfile = zipfile, mode = 'wb', quiet = TRUE)
unzip(zipfile, overwrite = TRUE) unzip(zipfile, overwrite = TRUE)
pred_data <- xgb.DMatrix(matrix(c(0, 0, 0, 0), nrow = 1, ncol = 4)) pred_data <- xgb.DMatrix(matrix(c(0, 0, 0, 0), nrow = 1, ncol = 4))
@ -72,13 +72,34 @@ test_that("Models from previous versions of XGBoost can be loaded", {
m <- regmatches(model_file, m)[[1]] m <- regmatches(model_file, m)[[1]]
model_xgb_ver <- m[2] model_xgb_ver <- m[2]
name <- m[3] name <- m[3]
is_rds <- endsWith(model_file, '.rds')
if (endsWith(model_file, '.rds')) { cpp_warning <- capture.output({
# Expect an R warning when a model is loaded from RDS and it was generated by version < 1.1.x
if (is_rds && compareVersion(model_xgb_ver, '1.1.1.1') < 0) {
booster <- readRDS(model_file)
expect_warning(predict(booster, newdata = pred_data))
expect_warning(run_booster_check(booster, name))
} else {
if (is_rds) {
booster <- readRDS(model_file) booster <- readRDS(model_file)
} else { } else {
booster <- xgb.load(model_file) booster <- xgb.load(model_file)
} }
predict(booster, newdata = pred_data) predict(booster, newdata = pred_data)
run_booster_check(booster, name) run_booster_check(booster, name)
}
})
if (compareVersion(model_xgb_ver, '1.0.0.0') < 0) {
# Expect a C++ warning when a model was generated in version < 1.0.x
m <- grepl(paste0('.*Loading model from XGBoost < 1\\.0\\.0, consider saving it again for ',
'improved compatibility.*'), cpp_warning, perl = TRUE)
expect_true(length(m) > 0 && all(m))
} else if (is_rds && model_xgb_ver == '1.1.1.1') {
# Expect a C++ warning when a model is loaded from RDS and it was generated by version 1.1.x
m <- grepl(paste0('.*Attempted to load internal configuration for a model file that was ',
'generated by a previous version of XGBoost.*'), cpp_warning, perl = TRUE)
expect_true(length(m) > 0 && all(m))
}
}) })
}) })

3
src/tree/updater_quantile_hist.h
View File

@ -132,7 +132,8 @@ class QuantileHistMaker: public TreeUpdater {
// We add this compatibility check because it's just recently that we (developers) began // We add this compatibility check because it's just recently that we (developers) began
// persuade R users away from using saveRDS() for model serialization. Hopefully, one day, // persuade R users away from using saveRDS() for model serialization. Hopefully, one day,
// everyone will be using xgb.save(). // everyone will be using xgb.save().
LOG(WARNING) << "Attempted to load interal configuration for a model file that was generated " LOG(WARNING)
<< "Attempted to load internal configuration for a model file that was generated "
<< "by a previous version of XGBoost. A likely cause for this warning is that the model " << "by a previous version of XGBoost. A likely cause for this warning is that the model "
<< "was saved with saveRDS() in R or pickle.dump() in Python. We strongly ADVISE AGAINST " << "was saved with saveRDS() in R or pickle.dump() in Python. We strongly ADVISE AGAINST "
<< "using saveRDS() or pickle.dump() so that the model remains accessible in current and " << "using saveRDS() or pickle.dump() so that the model remains accessible in current and "