Add mksurfdata_esmf system test to test-suite

cime_config/SystemTests/mksurfdataesmf.py

@@ -0,0 +1,131 @@
+"""
+This test passes if mksurfdata_esmf generates an fsurdat (surface dataset)
+and the CTSM completes a simulation with this fsurdat file.
+
+We test res = '10x15' because it uses a lower-res topography file instead of
+the 1-km topography raw dataset. The 1-km file causes the test to run out of
+memory on cheyenne.
+
+Currently casper complains that `git -C` is not a valid option.
+I added -C to the `git describe` in gen_mksurfdata_namelist.py for this
+system test to work.
+"""
+import os
+import sys
+import subprocess
+from datetime import datetime
+from CIME.SystemTests.system_tests_common import SystemTestsCommon
+from CIME.XML.standard_module_setup import *
+from CIME.SystemTests.test_utils.user_nl_utils import append_to_user_nl_files
+
+logger = logging.getLogger(__name__)
+
+class MKSURFDATAESMF(SystemTestsCommon):
+
+    def __init__(self, case):
+        """
+        initialize an object interface to the SMS system test
+        """
+        SystemTestsCommon.__init__(self, case)
+
+        # Paths and strings needed throughout
+        ctsm_root = self._case.get_value('COMP_ROOT_DIR_LND')
+        self._tool_path = os.path.join(ctsm_root, 'tools/mksurfdata_esmf')
+        self._tool_bld = os.path.join(self._get_caseroot(), 'tool_bld')
+        time_stamp = datetime.today().strftime("%y%m%d")
+        self._res = '10x15'  # see important comment in script's docstring
+        self._model_yr = '1850'
+        self._jobscript = os.path.join(self._get_caseroot(),
+            'mksurfdata_jobscript_single')
+        self._fsurdat_namelist = os.path.join(self._get_caseroot(),
+            f'surfdata_{self._res}_hist_78pfts_CMIP6_{self._model_yr}_c{time_stamp}.namelist')
+        self._fsurdat_nc = os.path.join(self._get_caseroot(),
+            f'surfdata_{self._res}_hist_78pfts_CMIP6_{self._model_yr}_c{time_stamp}.nc')
+        self._TestStatus_log_path = os.path.join(self._get_caseroot(),
+            'TestStatus.log')
+
+    def build_phase(self, sharedlib_only=False, model_only=False):
+        """
+        Build executable that will generate fsurdat
+        Generate namelist for generating fsurdat
+        Generate jobscript that runs executable
+        Modify user_nl_clm to point to the generated fsurdat
+        """
+        # build_phase gets called twice:
+        # - once with sharedlib_only = True and
+        # - once with model_only = True
+        # Call the following steps only once during the test but do not skip
+        # if the test stops and gets restarted.
+        if sharedlib_only:
+            # Paths and strings
+            build_script_path = os.path.join(self._tool_path,
+                'gen_mksurfdata_build.sh')
+            nml_script_path = os.path.join(self._tool_path,
+                'gen_mksurfdata_namelist.py')
+            gen_jobscript_path = os.path.join(self._tool_path,
+                'gen_mksurfdata_jobscript_single.py')
+            gen_mksurfdata_namelist = f'{nml_script_path} --res {self._res} --start-year {self._model_yr} --end-year {self._model_yr}'
+            gen_mksurfdata_jobscript = f'{gen_jobscript_path} --number-of-nodes 12 --tasks-per-node 12 --namelist-file {self._fsurdat_namelist}'
+
+            # Rm tool_bld and build executable that will generate fsurdat
+            try:
+                subprocess.check_call(f'rm -rf {self._tool_bld}', shell=True)
+            except subprocess.CalledProcessError as e:
+                sys.exit(f'{e} ERROR REMOVING {self._tool_bld}. DETAILS IN {self._TestStatus_log_path}')
+            try:
+                subprocess.check_call(f'{build_script_path} --blddir {self._tool_bld}', shell=True)
+            except subprocess.CalledProcessError as e:
+                sys.exit(f'{e} ERROR RUNNING {build_script_path}. DETAILS IN {self._TestStatus_log_path}')
+
+            # Generate namelist for generating fsurdat (rm namelist if exists)
+            if os.path.exists(self._fsurdat_namelist):
+                os.remove(self._fsurdat_namelist)
+            try:
+                subprocess.check_call(gen_mksurfdata_namelist, shell=True)
+            except subprocess.CalledProcessError as e:
+                sys.exit(f'{e} ERROR RUNNING {gen_mksurfdata_namelist}. DETAILS IN {self._TestStatus_log_path}')
+
+            # Generate jobscript that will run the executable
+            if os.path.exists(self._jobscript):
+                os.remove(self._jobscript)
+            try:
+                subprocess.check_call(gen_mksurfdata_jobscript, shell=True)
+            except subprocess.CalledProcessError as e:
+                sys.exit(f'{e} ERROR RUNNING {gen_mksurfdata_jobscript}. DETAILS IN {self._TestStatus_log_path}')
+            # Change self._jobscript to an executable file
+            subprocess.check_call(f'chmod a+x {self._jobscript}', shell=True)
+
+        # Call this step only once even if the test stops and gets restarted.
+        if not os.path.exists(os.path.join(self._get_caseroot(),
+            'done_MKSURFDATAESMF_setup.txt')):
+            # Modify user_nl_clm to point to the generated fsurdat
+            self._modify_user_nl()
+            with open('done_MKSURFDATAESMF_setup.txt', 'w') as fp:
+                pass
+
+        self.build_indv(sharedlib_only=sharedlib_only, model_only=model_only)
+
+    def run_phase(self):
+        """
+        Run executable to generate fsurdat
+        Submit CTSM run that uses fsurdat just generated
+        """
+
+        # Run executable to generate fsurdat (rm fsurdat if exists)
+        if os.path.exists(self._fsurdat_nc):
+            os.remove(self._fsurdat_nc)
+        try:
+            subprocess.check_call(self._jobscript, shell=True)
+        except subprocess.CalledProcessError as e:
+            sys.exit(f'{e} ERROR RUNNING {self._jobscript}; details in {self._TestStatus_log_path}')
+
+        # Submit CTSM run that uses fsurdat just generated
+        self.run_indv()
+
+    def _modify_user_nl(self):
+        """
+        Modify user_nl_clm to point to the generated fsurdat
+        """
+        append_to_user_nl_files(caseroot = self._get_caseroot(),
+                                component = "clm",
+                                contents = "fsurdat = '{}'".format(self._fsurdat_nc))

cime_config/config_tests.xml

@@ -15,6 +15,16 @@ This defines various CTSM-specific system tests
     <DOUT_S>FALSE</DOUT_S>
   </test>
 
+  <test NAME="MKSURFDATAESMF">
+    <DESC>Build and run the mksurfdata_esmf tool to generate a new fsurdat; then run the CTSM with this fsurdat</DESC>
+    <INFO_DBUG>1</INFO_DBUG>
+    <DOUT_S>FALSE</DOUT_S>
+    <CONTINUE_RUN>FALSE</CONTINUE_RUN>
+    <REST_OPTION>never</REST_OPTION>
+    <HIST_OPTION>$STOP_OPTION</HIST_OPTION>
+    <HIST_N>$STOP_N</HIST_N>
+  </test>
+
   <test NAME="FSURDATMODIFYCTSM">
     <DESC>Run the CTSM with an fsurdat generated by the fsurdat_modify tool</DESC>
     <INFO_DBUG>1</INFO_DBUG>

cime_config/testdefs/testlist_clm.xml

@@ -2369,6 +2369,23 @@
     </machines>
   </test>
 
+  <test name="MKSURFDATAESMF_P144x1" grid="f10_f10_mg37" compset="I1850Clm50BgcCrop">
+    <machines>
+      <machine name="cheyenne" compiler="intel" category="aux_clm">
+        <options>
+          <option name="wallclock">00:40:00</option>
+          <option name="comment">Smoke test that first runs the mksurfdata_esmf tool and then ensures that the CTSM does not fail using the just-generated fsurdat file. Global BGC CROP case.</option>
+        </options>
+      </machine>
+      <machine name="cheyenne" compiler="intel" category="clm_pymods">
+        <options>
+          <option name="wallclock">00:40:00</option>
+          <option name="comment">This test invokes python code, so it should be run whenever changing python code (in addition to being run as part of aux_clm).</option>
+        </options>
+      </machine>
+    </machines>
+  </test>
+
   <test name="FSURDATMODIFYCTSM_D_Mmpi-serial_Ld1" grid="5x5_amazon" compset="I2000Clm50SpRs">
     <machines>
       <machine name="cheyenne" compiler="intel" category="aux_clm">

tools/mksurfdata_esmf/README

@@ -32,7 +32,7 @@ on your machine you will be able to build the tool there.
 
 To get a list of the machines that have been ported to cime: 
 
-cd ../../cime/scripts
+cd ../../cime/scripts  # assumes we are in tools/mksurfdata_esmf
 ./query_config --machines
 
 NOTE:
@@ -41,16 +41,14 @@ to be referenced with the env variable PIO which will need to be in the porting
 for the machine. Currently an independent PIO library is not available on cime ported machines.
 
 =======================
-building the executable
+building the executable (working in tools/mksurfdata_esmf)
 =======================
 
-> cd tools/mksurfdata_esmf
 > ./gen_mksurfdata_build.sh      # For machines with a cime build
 
 =======================
 running for a single submission:
 =======================
-> cd ../
 # to generate your target namelist:
 > ./gen_mksurfdata_namelist.py --help
 # for example try --res 1.9x2.5 --start-year 1850 --end-year 1850:

tools/mksurfdata_esmf/gen_mksurfdata_build.sh

@@ -1,46 +1,103 @@
 #! /bin/bash -f
 
-hostname=`hostname --short`
+#----------------------------------------------------------------------
+# Usage subroutine
+usage() {
+  echo ""
+  echo "***********************************************************************"
+  echo "usage:"
+  echo "./gen_mksurfdata_build.sh"
+  echo ""
+  echo "valid arguments: "
+  echo "[-h|--help]  "
+  echo "     Displays this help message"
+  echo "[-v|--verbose]  "
+  echo "     Run in verbose mode"
+  echo "[-b|--blddir <blddir>]  "
+  echo "     Overwrites default, which is /bld in the same directory as ./gen_mksurfdata_build.sh"
+  echo "[-m|--machine <machine>]  "
+  echo "     Overwrites default MACH"
+  echo "***********************************************************************"
+}
+
+
+# Current working directory: the location of ./gen_mksurfdata_build.sh
+cwd=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+
+# Default settings
+verbose="No"
+blddir=$cwd/tool_bld  # may overwrite this default with command-line option (below)
 
 # Define what machine to use that's been ported to cime
+# May overwrite this default with command-line option --machine
+hostname=`hostname --short`
 case $hostname in
   ##cheyenne
   cheyenne* | r* )
       export MACH="cheyenne"
       ;;
+  ##casper
+  casper* )
+      export MACH="casper"
+      ;;
   ## Other machines
   *)
       export MACH="$hostname"
       ;;
 esac
 
-# Create /bld directory
+# Parse command-line options
+while [ $# -gt 0 ]; do
+   case $1 in
+       -h|--help )
+           usage
+           exit 0
+           ;;
+       -v|--verbose )
+           verbose="Yes"
+           ;;
+       -b|--blddir )
+           blddir=$2
+           shift
+           ;;
+       -m|--machine )
+           MACH=$2
+           shift
+           ;;
+       * )
+           echo "ERROR:: invalid argument sent in: $2"
+           usage
+           exit 1
+           ;;
+   esac
+   shift
+done
+
+# Create /tool_bld directory
 echo "cime Machine is: $MACH..."
-if [ -d "bld" ]; then
-  echo "A /bld directory exists; remove it to do a clean build..."
+if [ -d "$blddir" ]; then
+  echo "A /tool_bld directory exists; remove it to do a clean build..."
   exit 1
 fi
-cwd=`pwd`
-rm -rf bld
-mkdir bld
-cd bld
+mkdir $blddir
+cd $blddir
 
 # Run the cime configure tool to figure out what modules need to be loaded
 echo "Run cime configure for machine $MACH..."
 # You can specify the non-default compiler and mpi-library by adding --compiler and --mpilib settings
 if [ -z "$COMPILER" ] || [ -z "$MPILIB" ]; then
   echo "configure for the default MPI-library and compiler..."
-  ../../../cime/tools/configure --macros-format CMake --machine $MACH
+  $cwd/../../cime/tools/configure --macros-format CMake --machine $MACH
 else
   echo "configure for the specific MPILIB=$MPILIB and COMPILER=$COMPILER..."
-  ../../../cime/tools/configure --macros-format CMake --machine $MACH --compiler $COMPILER --mpilib $MPILIB
+  $cwd/../../cime/tools/configure --macros-format CMake --machine $MACH --compiler $COMPILER --mpilib $MPILIB
 fi
 
 if [ $? != 0 ]; then
   echo "Error doing configure for machine name: $MACH"
   exit 1
 fi
-. .env_mach_specific.sh
+. ./.env_mach_specific.sh
 echo "COMPILER = $COMPILER, MPILIB = $MPILIB, DEBUG = $DEBUG, OS = $OS"
 if [ -z "$PIO" ]; then
   echo "The PIO directory for the PIO build is required and was not set in the configure"
@@ -50,7 +107,7 @@ fi
 
 # Build the cmake files
 echo "Do the cmake build..."
-CC=mpicc FC=mpif90 cmake -DCMAKE_BUILD_TYPE=debug ../src
+CC=mpicc FC=mpif90 cmake -DCMAKE_BUILD_TYPE=debug $cwd/src
 if [ $? != 0 ]; then
   echo "Error doing cmake for $MACH $MPILIB $COMPILER"
   exit 1