[ci] speed up conda setup for macOS and Linux jobs (#5743)

.ci/test.sh

@@ -93,14 +93,9 @@ if [[ $TASK == "lint" ]]; then
     exit 0
 fi
 
-conda create -q -y -n $CONDA_ENV "${CONDA_PYTHON_REQUIREMENT}"
-source activate $CONDA_ENV
-
-cd $BUILD_DIRECTORY
-
 if [[ $TASK == "check-docs" ]] || [[ $TASK == "check-links" ]]; then
     cd $BUILD_DIRECTORY/docs
-    conda env update \
+    conda env create \
         -n $CONDA_ENV \
         --file ./env.yml || exit -1
     conda install \
@@ -109,6 +104,7 @@ if [[ $TASK == "check-docs" ]] || [[ $TASK == "check-links" ]]; then
         -n $CONDA_ENV \
             doxygen \
             'rstcheck>=6.0.0' || exit -1
+    source activate $CONDA_ENV
     # check reStructuredText formatting
     cd $BUILD_DIRECTORY/python-package
     rstcheck --report-level warning $(find . -type f -name "*.rst") || exit -1
@@ -131,8 +127,8 @@ if [[ $TASK == "check-docs" ]] || [[ $TASK == "check-links" ]]; then
     exit 0
 fi
 
-# re-including python=version[build=*cpython] to ensure that conda doesn't fall back to pypy
-conda install -q -y -n $CONDA_ENV \
+# including python=version[build=*cpython] to ensure that conda doesn't fall back to pypy
+conda create -q -y -n $CONDA_ENV \
     cloudpickle \
     dask-core \
     distributed \
@@ -147,6 +143,10 @@ conda install -q -y -n $CONDA_ENV \
     scikit-learn \
     scipy || exit -1
 
+source activate $CONDA_ENV
+
+cd $BUILD_DIRECTORY
+
 if [[ $OS_NAME == "macos" ]] && [[ $COMPILER == "clang" ]]; then
     # fix "OMP: Error #15: Initializing libiomp5.dylib, but found libomp.dylib already initialized." (OpenMP library conflict due to conda's MKL)
     for LIBOMP_ALIAS in libgomp.dylib libiomp5.dylib libomp.dylib; do sudo ln -sf "$(brew --cellar libomp)"/*/lib/libomp.dylib $CONDA_PREFIX/lib/$LIBOMP_ALIAS || exit -1; done