Merge pull request #1153 from PCMDI/bug/1152_piControl-multifile_lee1043

Bug/1152 pi control multifile lee1043
PCMDI · Oct 2, 2024 · 388b3ee · 388b3ee
2 parents 776634f + a813521
commit 388b3ee
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 9 deletions.
diff --git a/pcmdi_metrics/variability_mode/lib/lib_variability_mode.py b/pcmdi_metrics/variability_mode/lib/lib_variability_mode.py
@@ -53,7 +53,7 @@ def get_domain_range(mode: str, regions_specs: dict):
 
 
 def read_data_in(
-    path: str,
+    path: Union[str, list],
     var_in_data: str,
     var_to_consider: str,
     syear: Union[str, int, float],

diff --git a/pcmdi_metrics/variability_mode/variability_modes_driver.py b/pcmdi_metrics/variability_mode/variability_modes_driver.py
@@ -83,6 +83,17 @@
     write_nc_output,
 )
 
+
+def search_paths(paths, index1, index2, case_sensitive=False):
+    def split_string(text):
+        return set(re.split(r"[._ /]", text.lower() if not case_sensitive else text))
+
+    index1 = index1 if case_sensitive else index1.lower()
+    index2 = index2 if case_sensitive else index2.lower()
+
+    return [path for path in paths if {index1, index2}.issubset(split_string(path))]
+
+
 # =================================================
 # Collect user defined options
 # -------------------------------------------------
@@ -553,7 +564,7 @@
 
     model_path_list = sort_human(model_path_list)
 
-    debug_print(f"model_path_list: f{model_path_list}", debug)
+    debug_print(f"model_path_list: {model_path_list}", debug)
 
     # Find where run can be gripped from given filename template for modpath
     if realization == "*":
@@ -569,18 +580,27 @@
             ).split("/")[-1],
         ).index(realization)
 
+        runs = [
+            re.split("[._]", model_path.split("/")[-1])[run_in_modpath]
+            for model_path in model_path_list
+        ]
+
+    else:
+        runs = [realization]
+
+    print("runs:", runs)
+
     # -------------------------------------------------
     # Run
     # -------------------------------------------------
-    for model_path in model_path_list:
-        print("model_path:", model_path)
+    for run in runs:
+        print("run:", runs)
         try:
-            if realization == "*":
-                run = re.split("[._]", model_path.split("/")[-1])[run_in_modpath]
-            else:
-                run = realization
             print(" --- ", run, " ---")
 
+            model_run_path = search_paths(model_path_list, model, run)
+            print("model_run_path:", model_run_path)
+
             if run not in result_dict["RESULTS"][model]:
                 result_dict["RESULTS"][model][run] = {}
 
@@ -602,7 +622,7 @@
 
             # read data in
             model_timeseries = read_data_in(
-                model_path,
+                model_run_path,
                 var,
                 var,
                 msyear,