Update geopandas and cluster environment variables for RHEL9 on kestrel

environment.yml

@@ -8,7 +8,7 @@ dependencies:
   - bokeh=3.9
   - click=8.4
   - cmocean=4.0
-  - geopandas=1.1
+  - geopandas=1.2
   - git-lfs=3.7
   - gitpython=3.1
   - h5py=3.16

helpers/interim_report_batch.py

@@ -26,7 +26,7 @@
 parser.add_argument('--batch_name', '-b', type=str, default='', help='Prefix for batch of runs')
 args = parser.parse_args()
 # check if on hpc
-hpc = True if (int(os.environ.get('REEDS_USE_SLURM',0))) else False
+hpc = reeds.slurm.check_slurm()
 
 # get list of cases with matching batch name
 case_list = glob(os.path.join('runs', '*'))

helpers/restart_runs.py

@@ -145,7 +145,7 @@
             f.writelines(line + '\n')
 
 # Check if we are going to run this in parallel or not
-hpc = True if (int(os.environ.get('REEDS_USE_SLURM',0))) else False
+hpc = reeds.slurm.check_slurm()
 if hpc and len(runs_failed) > 1:
     # On HPC with multiple cases 
     cases_per_node = int(input('Number of simultaneous runs per node [integer]: '))

hourlize/run_hourlize.py

@@ -17,7 +17,11 @@
 import sys
 import traceback
 import copy_rev_folders
+from pathlib import Path
 from collections import OrderedDict
+sys.path.append(str(Path(__file__).parent.parent))
+import reeds
+
 
 #%% ===========================================================================
 ### --- FUNCTIONS ---
@@ -26,16 +30,14 @@
 def get_remote_path(local):
     """detect path to supply curve files either on the HPC or nrelnas01"""
     # remote path for supply curves
-    hpc = True if ('NREL_CLUSTER' in os.environ) else False
+    hpc = reeds.slurm.check_slurm()
     if hpc:
         #For running hourlize on the HPC link to shared-projects folder
-        if os.environ.get('NREL_CLUSTER') == 'kestrel':
+        cluster = reeds.slurm.get_cluster()
+        if cluster == 'kestrel':
             remotepath = '/kfs2/shared-projects/reeds'
-        elif os.environ.get('NREL_CLUSTER') == 'eagle':
-            remotepath = '/shared-projects/reeds'
         else:
-            raise Exception(f"Detected {os.environ.get('NREL_CLUSTER')} as NLR Cluster; "
-                            "only 'eagle' and 'kestrel' are supported")
+            raise ValueError(f"Unsupported cluster: {cluster}")
     else:
         # if not on the hpc running link to nrelnas01 and set local to true
         remotepath = os.path.join(
@@ -165,13 +167,9 @@ def launch_batch_file(casename, configpath, outpath, args, copy_from_rev=False):
     # setup script to run each case
     ext = '.sh' if os.name == 'posix' else '.bat'
     with open(os.path.join(outpath, casename + "_run" + ext), 'w+') as OPATH:
-        if os.environ.get('NREL_CLUSTER') == 'kestrel':
-            OPATH.writelines("source /nopt/nrel/apps/env.sh \n")
-            OPATH.writelines("module load anaconda3 \n")
-            OPATH.writelines("conda activate reeds \n\n")
-        elif os.environ.get('NREL_CLUSTER') == 'eagle':
-            OPATH.writelines("module load conda \n")
-            OPATH.writelines("conda activate reeds \n\n")
+        OPATH.writelines("source /nopt/nrel/apps/env.sh \n")
+        OPATH.writelines("module load anaconda3 \n")
+        OPATH.writelines("conda activate reeds \n\n")
         # run hourlize
         OPATH.writelines(f"cd {hourlize_path}\n")
         callstring = f"python {args.mode}.py --config {configpath}"
@@ -182,8 +180,6 @@ def launch_batch_file(casename, configpath, outpath, args, copy_from_rev=False):
         callstring += "\n"
         OPATH.writelines(callstring)
 
-
-
     # launch run locally or submit to hpc
     if args.local:
         if args.nosubmit:

hourlize/write_site_parameters.py

@@ -26,7 +26,7 @@ def check_negative_values(df, cost_cols):
         ) 
 
 #%%### Fixed inputs
-if reeds.io.hpc:
+if reeds.slurm.check_slurm():
     remotepath = '/kfs2/shared-projects/reeds'
     filepaths = {
         'land': os.path.join(

postprocessing/combine_runs/combine_runs.py

@@ -369,7 +369,7 @@ def main(reeds_path, batch_name, folder_name_suffix, runlist, keywords, local, d
         combine_case = cases.loc[cases.scenario == scen].copy()
         combine_case.to_csv(os.path.join(output_path, 'cases_combined.csv'), index=False)
         # check if run should be submitted to HPC or run directly
-        hpc = True if ('NREL_CLUSTER' in os.environ) else False
+        hpc = reeds.slurm.check_slurm()
         if hpc and local:
             print(
                 "Note: you are on the HPC but are running locally on your current node. "

postprocessing/run_pcm.py

@@ -31,20 +31,6 @@
 
 
 # %% Functions
-def check_slurm(forcelocal=False):
-    """Check whether to submit slurm jobs (if on HPC) or run locally"""
-    hpc = (
-        False
-        if forcelocal
-        else (
-            True
-            if int(os.environ.get('REEDS_USE_SLURM', 0)) and ('NREL_CLUSTER' in os.environ)
-            else False
-        )
-    )
-    return hpc
-
-
 def solvestring_pcm(
     batch_case,
     sw,
@@ -367,7 +353,7 @@ def main(casepath, t, switch_mods=switch_mods_default, label='', overwrite=False
     # bigmem = True
 
     # %% Determine whether to submit slurm job
-    hpc = check_slurm(forcelocal=forcelocal)
+    hpc = reeds.slurm.check_slurm(forcelocal=forcelocal)
 
     ### Run it
     if not hpc:

postprocessing/run_reeds2pras.py

@@ -1,22 +1,18 @@
 #%% Imports
 import pandas as pd
 import os
+import sys
 import site
 import json
 from glob import glob
 import subprocess
 ### Local imports
 reeds_path = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+sys.path.append(reeds_path)
+import reeds
 
 
 #%%### Functions
-def check_slurm(forcelocal=False):
-    """Check whether to submit slurm jobs (if on HPC) or run locally"""
-    hpc = True if (int(os.environ.get('REEDS_USE_SLURM',0))) else False
-    hpc = False if forcelocal else hpc
-    return hpc
-
-
 def submit_job(
     case,
     year=0,
@@ -241,7 +237,7 @@ def main(
     # switch_mods = {}
 
     #%% Determine whether to submit SLURM job
-    hpc = check_slurm(forcelocal=local)
+    hpc = reeds.slurm.check_slurm(forcelocal=local)
 
     ### Run it
     if hpc:

reeds/__init__.py

@@ -14,6 +14,7 @@
 from . import remote as remote
 from . import report_utils as report_utils
 from . import results as results
+from . import slurm as slurm
 from . import spatial as spatial
 from . import techs as techs
 from . import timeseries as timeseries

reeds/io.py

@@ -21,7 +21,6 @@
 if 'runs' in reeds_path.split(os.path.sep):
     reeds_path = reeds_path[: reeds_path.index(os.sep + 'runs' + os.sep)]
 
-hpc = True if ('NREL_CLUSTER' in os.environ) else False
 
 #   ########  ########    ###    ########
 #   ##     ## ##         ## ##   ##     ##
@@ -756,7 +755,7 @@ def get_switches(case=None, **kwargs):
     threads = get_param_value(fpath_opt, "threads", dtype=int)
     sw['threads'] = threads
     ## Determine whether run is on HPC
-    sw['hpc'] = True if int(os.environ.get('REEDS_USE_SLURM', 0)) else False
+    sw['hpc'] = reeds.slurm.check_slurm()
     ## Add the run location
     sw['casedir'] = case
     sw['reeds_path'] = reeds_path if case is None else os.path.dirname(os.path.dirname(case))

reeds/slurm.py

@@ -0,0 +1,26 @@
+### Imports
+import os
+
+
+### Functions
+def get_cluster():
+    """
+    Return the name of the HPC cluster if on a cluster.
+    If not on a cluster, return None.
+    """
+    name_fields = ['NREL_CLUSTER', 'NLR_CLUSTER']
+    ## Keep the first match in name_fields
+    for name_field in name_fields:
+        out = os.environ.get(name_field)
+        if out:
+            break
+    return out
+
+
+def check_slurm(forcelocal=False):
+    """Check whether to submit slurm jobs (if on HPC) or run locally"""
+    if forcelocal:
+        hpc = False
+    else:
+        hpc = True if get_cluster() else False
+    return hpc

runreeds.py

@@ -147,8 +147,8 @@ def get_ivt_numclass(reeds_path, casedir, caseSwitches):
 def get_rev_paths(revswitches):
     # Expand on reV path based on where this run is happening
     # when running on the HPC this links to the shared-projects folder
-    hpc = True if (int(os.environ.get('REEDS_USE_SLURM',0))) else False
-    if os.environ.get('NREL_CLUSTER') == 'kestrel':
+    hpc = reeds.slurm.check_slurm()
+    if reeds.slurm.get_cluster() == 'kestrel':
         hpc_path = '/kfs2/shared-projects/reeds/Supply_Curve_Data'
     else:
         hpc_path = '/shared-projects/reeds/Supply_Curve_Data'
@@ -833,15 +833,13 @@ def setupEnvironment(
         cases_suffix = str(input('Case Suffix: '))
 
     #%% Check whether to submit slurm jobs (if on HPC) or run locally
-    hpc = True if (int(os.environ.get('REEDS_USE_SLURM',0))) else False
-    hpc = False if forcelocal else hpc
+    hpc = reeds.slurm.check_slurm(forcelocal=forcelocal)
 
     ### If on NLR HPC but NOT submitting slurm job, ask for confirmation
-    if ('NREL_CLUSTER' in os.environ) and (not hpc):
+    if reeds.slurm.get_cluster() and (not hpc):
         print(
-            "It looks like you're running on the NLR HPC but the REEDS_USE_SLURM environment "
-            "variable is not set to 1, meaning the model will run locally rather than being "
-            "submitted as a slurm job. Are you sure you want to run locally?"
+            "It looks like you're running on the HPC login node instead of submitting a "
+            "slurm job. Are you sure you want to run locally (not recommended)?"
         )
         confirm_local = str(input('Run job locally? y/[n]: ') or 'n')
         if confirm_local not in ['y','Y','yes','Yes','YES']:
@@ -1278,19 +1276,12 @@ def write_batch_script(
             comment('Set up nodal environment for run', OPATH)
             OPATH.writelines(". $HOME/.bashrc \n")
             OPATH.writelines("module purge \n")
-
-            if os.environ.get('NREL_CLUSTER') == 'kestrel':
-                OPATH.writelines("source /nopt/nrel/apps/env.sh \n")
-                OPATH.writelines("module load anaconda3 \n")
-                OPATH.writelines("module use /nopt/nrel/apps/software/gams/modulefiles \n")
-                OPATH.writelines("module load gams \n")
-                OPATH.writelines("module load julia/1.12.1 \n")
-            else:
-                OPATH.writelines("module load conda \n")
-                OPATH.writelines("module load gams \n")
-
+            OPATH.writelines("module load anaconda3 \n")
+            OPATH.writelines("module use /nopt/nrel/apps/software/gams/modulefiles \n")
+            OPATH.writelines("module load gams \n")
+            OPATH.writelines("module load julia/1.12.1 \n")
             OPATH.writelines("conda activate reeds \n")
-            OPATH.writelines('export R_LIBS_USER="$HOME/rlib" \n\n\n')
+            OPATH.writelines('\n\n')
 
         #%% Write the input_processing script calls
         big_comment('Input processing', OPATH)

tests/get_subset_h5.py

@@ -32,7 +32,7 @@
 ]
 
 # series of paths to updated county h5 files
-hpc = True if ('NREL_CLUSTER' in os.environ) else False
+hpc = reeds.slurm.check_slurm()
 if hpc:
     remotepath = '/kfs2/shared-projects/reeds' #kestrel
 else: