Merge the latest updates from PR geoschem#200 into the branch for PR g…

…eoschem#204

This brings some later fixes for issues with the creation of
text saying whether or not 2 version have identical results in
the table output.

Signed-off-by: Bob Yantosca <[email protected]>

gcpy/benchmark.py

@@ -188,11 +188,7 @@ def create_total_emissions_table(
 
     # Write a placeholder to the file that denotes where
     # the list of species with differences will be written
-    placeholder = "@%% insert diff_list here %%@"
-    if "Inv" in template:
-        print(f"Inventories that differ btw {refstr} and {devstr}:", file=f)
-    else:
-        print(f"Species that differ btw {refstr} and {devstr}:", file=f)
+    placeholder = "@%% insert diff status here %%@"
     print(f"{placeholder}\n\n", file=f)
 
     # Define a list for differences
@@ -359,7 +355,10 @@ def create_total_emissions_table(
     util.insert_text_into_file(
         filename=outfilename,
         search_text=placeholder,
-        replace_text=diff_list_to_text(diff_list),
+        replace_text=diff_list_to_text(
+            refstr,
+            devstr,
+            diff_list),
         width=90
     )
 
@@ -472,18 +471,16 @@ def create_global_mass_table(
     if trop_only:
         title1 = f"### Global mass (Gg) {label} (Trop only)"
     title2 = f"### Ref = {refstr}; Dev = {devstr}"
-    title3 = f"### Species that differ btw {refstr} and {devstr}:"
 
     # Write a placeholder to the file that denotes where
     # the list of species with differences will be written
-    placeholder = "@%% insert diff_list here %%@"
+    placeholder = "@%% insert diff status here %%@"
 
     # Print header to file
     print("#" * 89, file=f)
     print(f"{title1 : <86}{'###'}", file=f)
     print(f"{title2 : <86}{'###'}", file=f)
     print(f"{'###'  : <86}{'###'}", file=f)
-    print(f"{title3 : <86}{'###'}", file=f)
     print(f"{placeholder}", file=f)
     print("#" * 89, file=f)
 
@@ -585,6 +582,8 @@ def create_global_mass_table(
         filename=outfilename,
         search_text=placeholder,
         replace_text=diff_list_to_text(
+            refstr,
+            devstr,
             diff_list,
             fancy_format=True
         ),
@@ -4509,8 +4508,7 @@ def create_benchmark_summary_table(
 
     # Variables to skip
     skip_vars = gcon.skip_these_vars
-    skip_vars.append("corner_lats")
-    skip_vars.append("corner_lons")
+    skip_vars.append("AREA")
 
     # Pick the proper function to read the data
     reader = util.dataset_reader(
@@ -4562,9 +4560,14 @@ def create_benchmark_summary_table(
         diff_list = []
 
         # Keep track of which variables are different
-        # Loop over the common variables
+        # NOTE: Use 32-point float for comparisons since this is
+        # the precision used for History diagnostics.
         for v in vardict["commonvarsData"]:
-            if not util.array_equals(refdata[v], devdata[v]):
+            if not util.array_equals(
+                    refdata[v],
+                    devdata[v],
+                    dtype=np.float32
+            ):
                 diff_list.append(v)
 
         # Drop duplicate values from diff_list
@@ -4592,6 +4595,8 @@ def create_benchmark_summary_table(
 
 
 def diff_list_to_text(
+        refstr,
+        devstr,
         diff_list,
         fancy_format=False
 ):
@@ -4617,11 +4622,19 @@ def diff_list_to_text(
     # Strip out duplicates from diff_list
     # Prepare a message about species differences (or alternate msg)
     diff_list = util.unique_values(diff_list, drop=[None])
-    diff_text = util.wrap_text(diff_list, width=85)
-    if len(diff_text) > 85:
-        diff_text = "... Too many diffs to print (see below for details)"
 
+    # Print the text
+    n_diff = len(diff_list)
+    if n_diff > 0:
+        diff_text = f"{devstr} and {refstr} show {n_diff} differences"
+    else:
+        diff_text = f"{devstr} and {refstr} are identical"
+    diff_text = util.wrap_text(
+        diff_text,
+        width=83
+    )
+
     if fancy_format:
         diff_text = f"### {diff_text : <82}{'###'}"
-
+        
     return diff_text.strip()

gcpy/util.py

@@ -846,7 +846,8 @@ def compare_varnames(
                              refdata and devdata, and that have lat,
                              lon, and level dimensions.
             commonvarsData   List of all commmon 2D or 3D data variables,
-                             excluding index variables.
+                             excluding index variables.  This is the
+                             list of "plottable" variables.
             refonly          List of 2D or 3D variables that are only
                              present in refdata.
             devonly          List of 2D or 3D variables that are only
@@ -858,37 +859,29 @@ def compare_varnames(
     refonly = [v for v in refvars if v not in devvars]
     devonly = [v for v in devvars if v not in refvars]
     dimmismatch = [v for v in commonvars if refdata[v].ndim != devdata[v].ndim]
-    commonvarsOther = [
+    # Assume plottable data has lon and lat
+    # This is OK for purposes of benchmarking
+    #  -- Bob Yantosca (09 Feb 2023)
+    commonvarsData = [
         v for v in commonvars if (
-          (
-            ("lat" not in refdata[v].dims or "Xdim" not in refdata[v].dims)
-            and
-            ("lon" not in refdata[v].dims or "Ydim" not in refdata[v].dims)
+            ("lat" in refdata[v].dims or "Ydim" in refdata[v].dims)
             and
-            ("lev" not in refdata[v].dims)
-          )
-          or
-          (
-            ("hyam" in v or "hybm" in v)  # Omit these from plottable data
-          )
+            ("lon" in refdata[v].dims or "Xdim" in refdata[v].dims)
         )
+    ]        
+    commonvarsOther = [
+        v for v in commonvars if (
+           v not in commonvarsData
+        )    
     ]
     commonvars2D = [
         v for v in commonvars if (
-          ("lat" in refdata[v].dims or "Xdim" in refdata[v].dims)
-          and
-          ("lon" in refdata[v].dims or "Ydim" in refdata[v].dims)
-          and
-          ("lev" not in refdata[v].dims)
+            (v in commonvarsData) and ("lev" not in refdata[v].dims)
         )
     ]
     commonvars3D = [
         v for v in commonvars if (
-          ("lat" in refdata[v].dims or "Xdim" in refdata[v].dims)
-          and
-          ("lon" in refdata[v].dims or "Ydim" in refdata[v].dims)
-          and
-          ("lev" in refdata[v].dims)
+            (v in commonvarsData) and ("lev" in refdata[v].dims)
         )
     ]
 
@@ -2280,7 +2273,8 @@ def insert_text_into_file(
 
 def array_equals(
         refdata,
-        devdata
+        devdata,
+        dtype=np.float64
 ):
     """
     Tests two arrays for equality.  Useful for checking which
@@ -2292,6 +2286,9 @@ def array_equals(
         The first array to be checked.
     devdata: xarray DataArray or numpy ndarray
         The second array to be checked.
+    dtype : np.float32 or np.float64
+        The precision that will be used to make the evaluation.
+        Default: np.float64
 
     Returns:
     --------
@@ -2314,9 +2311,9 @@ def array_equals(
 
     # This method will work if the arrays hve different dimensions
     # but an element-by-element search will not!
-    refsum = np.sum(refdata, dtype=np.float64)
-    devsum = np.sum(devdata, dtype=np.float64)
-    return np.abs(devsum - refsum) > np.float64(0.0)
+    refsum = np.nansum(refdata, dtype=dtype)
+    devsum = np.nansum(devdata, dtype=dtype)
+    return (not np.abs(devsum - refsum) > dtype(0.0))
 
 
 def make_directory(