add keras skips (#1303)

Author: EmilHvitfeldt

.github/workflows/R-CMD-check.yaml

@@ -59,20 +59,19 @@ jobs:
           extra-packages: any::rcmdcheck
           needs: check
 
-      - name: Install reticulate
-        run: pak::pkg_install('reticulate')
+      - name: Install dev reticulate
+        run: pak::pkg_install('rstudio/reticulate')
         shell: Rscript {0}
 
-      - name: Install Miniconda
-        # conda can fail at downgrading python, so we specify python version in advance
-        env:
-          RETICULATE_MINICONDA_PYTHON_VERSION: "3.8"
-        run: reticulate::install_miniconda() # creates r-reticulate conda env by default
-        shell: Rscript {0}
+      - uses: actions/setup-python@v4
+        with:
+          python-version: 3.11
 
       - name: Install TensorFlow
         run: |
-          tensorflow::install_tensorflow(version='2.13', conda_python_version = NULL)
+          reticulate::virtualenv_create('r-reticulate', python='3.11')
+          reticulate::use_virtualenv('r-reticulate')
+          tensorflow::install_tensorflow(version='2.16')
         shell: Rscript {0}
 
       - uses: r-lib/actions/check-r-package@v2

.github/workflows/pkgdown.yaml

@@ -37,15 +37,14 @@ jobs:
           needs: website
 
       - name: Install Miniconda
-        # conda can fail at downgrading python, so we specify python version in advance
-        env:
-          RETICULATE_MINICONDA_PYTHON_VERSION: "3.8"
-        run: reticulate::install_miniconda() # creates r-reticulate conda env by default
+        run: |
+          reticulate::install_miniconda()
         shell: Rscript {0}
-
+ 
       - name: Install TensorFlow
         run: |
-          tensorflow::install_tensorflow(version='2.13', conda_python_version = NULL)
+          reticulate::conda_create('r-reticulate', packages = c('python==3.11'))
+          tensorflow::install_tensorflow(version='2.16')
         shell: Rscript {0}
 
       - name: Install Torch

.github/workflows/test-coverage.yaml

@@ -32,15 +32,14 @@ jobs:
         shell: Rscript {0}
 
       - name: Install Miniconda
-        # conda can fail at downgrading python, so we specify python version in advance
-        env:
-          RETICULATE_MINICONDA_PYTHON_VERSION: "3.8"
-        run: reticulate::install_miniconda() # creates r-reticulate conda env by default
+        run: |
+          reticulate::install_miniconda()
         shell: Rscript {0}
-
+ 
       - name: Install TensorFlow
         run: |
-          tensorflow::install_tensorflow(version='2.13', conda_python_version = NULL)
+          reticulate::conda_create('r-reticulate', packages = c('python==3.11'))
+          tensorflow::install_tensorflow(version='2.16')
         shell: Rscript {0}
 
       - name: Test coverage

tests/testthat/test-linear_reg_keras.R

@@ -1,3 +1,4 @@
+skip("waiting for keras3")
 skip_if_not_installed("modeldata")
 
 hpc <- hpc_data[1:150, c(2:5, 8)]

tests/testthat/test-logistic_reg_keras.R

@@ -1,3 +1,4 @@
+skip("waiting for keras3")
 skip_if_not_installed("modeldata")
 
 # ------------------------------------------------------------------------------

tests/testthat/test-mlp_keras.R

@@ -1,3 +1,4 @@
+skip("waiting for keras3")
 skip_if_not_installed("modeldata")
 
 hpc <- hpc_data[1:150, c(2:5, 8)]

tests/testthat/test-multinom_reg_keras.R

@@ -1,3 +1,4 @@
+skip("waiting for keras3")
 skip_if_not_installed("modeldata")
 
 hpc <- hpc_data[1:150, c(2:5, 8)]

vignettes/articles/Examples.Rmd

@@ -553,6 +553,7 @@ The following examples use consistent data sets throughout. For regression, we u
   Now we create the model fit object:
 
   ```{r}
+  #| eval: false
   set.seed(1)
   linreg_reg_fit <- linreg_reg_spec |> fit(ridership ~ ., data = Chicago_train)
   linreg_reg_fit
@@ -561,6 +562,7 @@ The following examples use consistent data sets throughout. For regression, we u
   The holdout data can be predicted:
 
   ```{r}
+  #| eval: false
   predict(linreg_reg_fit, Chicago_test)
   ```
 
@@ -791,6 +793,7 @@ The following examples use consistent data sets throughout. For regression, we u
   Now we create the model fit object:
 
   ```{r}
+  #| eval: false
   set.seed(1)
   logreg_cls_fit <- logreg_cls_spec |> fit(Class ~ ., data = data_train)
   logreg_cls_fit
@@ -799,6 +802,7 @@ The following examples use consistent data sets throughout. For regression, we u
   The holdout data can be predicted for both hard class predictions and probabilities. We'll bind these together into one tibble:
 
   ```{r}
+  #| eval: false
   bind_cols(
     predict(logreg_cls_fit, data_test),
     predict(logreg_cls_fit, data_test, type = "prob")
@@ -1103,6 +1107,7 @@ The following examples use consistent data sets throughout. For regression, we u
   Now we create the model fit object:
 
   ```{r}
+  #| eval: false
   set.seed(1)
   mlp_reg_fit <- mlp_reg_spec |> fit(ridership ~ ., data = Chicago_train)
   mlp_reg_fit
@@ -1111,6 +1116,7 @@ The following examples use consistent data sets throughout. For regression, we u
   The holdout data can be predicted:
 
   ```{r}
+  #| eval: false
   predict(mlp_reg_fit, Chicago_test)
   ```
 
@@ -1136,6 +1142,7 @@ The following examples use consistent data sets throughout. For regression, we u
   Now we create the model fit object:
 
   ```{r}
+  #| eval: false
   set.seed(1)
   mlp_cls_fit <- mlp_cls_spec |> fit(Class ~ ., data = data_train)
   mlp_cls_fit
@@ -1144,6 +1151,7 @@ The following examples use consistent data sets throughout. For regression, we u
   The holdout data can be predicted for both hard class predictions and probabilities. We'll bind these together into one tibble:
 
   ```{r}
+  #| eval: false
   bind_cols(
     predict(mlp_cls_fit, data_test),
     predict(mlp_cls_fit, data_test, type = "prob")
@@ -1365,6 +1373,7 @@ The following examples use consistent data sets throughout. For regression, we u
   Now we create the model fit object:
 
   ```{r}
+  #| eval: false
   set.seed(1)
   mr_cls_fit <- mr_cls_spec |> fit(island ~ ., data = penguins_train)
   mr_cls_fit
@@ -1373,6 +1382,7 @@ The following examples use consistent data sets throughout. For regression, we u
   The holdout data can be predicted for both hard class predictions and probabilities. We'll bind these together into one tibble:
 
   ```{r}
+  #| eval: false
   bind_cols(
     predict(mr_cls_fit, penguins_test),
     predict(mr_cls_fit, penguins_test, type = "prob")