all pre-commit now passing

Author: jedwards4b

.gitignore

@@ -182,9 +182,9 @@ cython_debug/
 .abstra/
 
 # Visual Studio Code
-#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore
 #  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
-#  and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#  and can be added to the global gitignore or merged into this file. However, if you prefer,
 #  you could uncomment the following to ignore the entire vscode folder
 # .vscode/
 

.pylintrc

@@ -15,5 +15,5 @@ disable=
 max-line-length=100
 
 [TYPECHECK]
-ignored-modules=xarray,xesmf,cmor
+ignored-modules=xarray,xesmf,cmor,numpy,geocat.comp
 ignored-classes=cmor

cmip7_prep/__init__.py

@@ -1,3 +1,3 @@
-
 """CMIP7 preparation toolkit: regridding and CMOR writing for CESM outputs."""
+
 __all__ = ["regrid", "cmor_writer"]

cmip7_prep/cli.py

@@ -1,5 +1,5 @@
-
 """Simple CLI entry points for cmip7_prep (argparse-based)."""
+
 from __future__ import annotations
 import argparse
 from pathlib import Path
@@ -9,13 +9,15 @@
 from cmip7_prep.regrid import regrid_to_1deg
 from cmip7_prep.cmor_writer import CmorSession
 
+
 def make_target(out: Path) -> None:
     """Create a canonical 1° grid template with lat/lon and (placeholder) bounds."""
     lat = np.arange(-89.5, 90.5, 1.0)
     lon = np.arange(0.5, 360.5, 1.0)
     ds = xr.Dataset(coords={"lat": ("lat", lat), "lon": ("lon", lon)})
     ds.to_netcdf(out)
 
+
 def prepare(
     *,
     var: str,
@@ -26,43 +28,75 @@ def prepare(
     outdir: Path,
 ) -> None:
     """Regrid (if needed) and CMOR-write a single variable (thin wrapper demo)."""
-    ds = xr.open_mfdataset([str(p) for p in in_files], combine="by_coords", use_cftime=True)
+    ds = xr.open_mfdataset(
+        [str(p) for p in in_files], combine="by_coords", use_cftime=True
+    )
     da = ds[var]
     ds_tmp = xr.Dataset({var: da})
     da1 = regrid_to_1deg(ds_tmp, var)
     ds_out = xr.Dataset({var: da1})
-    with CmorSession(tables_path=str(cmor_tables), dataset_json=str(dataset_json)) as cm:
-        vdef = type("VDef", (), {"name": var, "realm": realm, "units": da.attrs.get("units", "")})()
+    with CmorSession(
+        tables_path=str(cmor_tables), dataset_json=str(dataset_json)
+    ) as cm:
+        vdef = type(
+            "VDef",
+            (),
+            {"name": var, "realm": realm, "units": da.attrs.get("units", "")},
+        )()
         cm.write_variable(ds_out, var, vdef, outdir=outdir)
 
+
 def main(argv: list[str] | None = None) -> int:
     """Entry point for the cmip7_prep command-line interface."""
     p = argparse.ArgumentParser(prog="cmip7-prep")
     sub = p.add_subparsers(dest="cmd", required=True)
 
     p_make = sub.add_parser("make-target", help="Write a simple 1° lat/lon grid file.")
-    p_make.add_argument("out", type=Path, help="Output NetCDF path for the target grid.")
+    p_make.add_argument(
+        "out", type=Path, help="Output NetCDF path for the target grid."
+    )
     p_make.set_defaults(func=lambda a: make_target(a.out))
 
-    p_prep = sub.add_parser("prepare", help="Regrid one variable and write CMOR output.")
+    p_prep = sub.add_parser(
+        "prepare", help="Regrid one variable and write CMOR output."
+    )
     p_prep.add_argument("--var", required=True, help="Variable name in input files.")
-    p_prep.add_argument("--in-file", "-i", action="append", required=True, type=Path, help="Input files (repeatable).")
+    p_prep.add_argument(
+        "--in-file",
+        "-i",
+        action="append",
+        required=True,
+        type=Path,
+        help="Input files (repeatable).",
+    )
     p_prep.add_argument("--realm", required=True, help="CMOR realm/table, e.g., Amon.")
-    p_prep.add_argument("--cmor-tables", required=True, type=Path, help="Path to CMOR Tables directory.")
-    p_prep.add_argument("--dataset-json", required=True, type=Path, help="Path to cmor_dataset.json.")
-    p_prep.add_argument("--outdir", default=Path("out"), type=Path, help="Output directory for CMORized files.")
-    p_prep.set_defaults(func=lambda a: prepare(
-        var=a.var,
-        in_files=a.in_file,
-        realm=a.realm,
-        cmor_tables=a.cmor_tables,
-        dataset_json=a.dataset_json,
-        outdir=a.outdir,
-    ))
+    p_prep.add_argument(
+        "--cmor-tables", required=True, type=Path, help="Path to CMOR Tables directory."
+    )
+    p_prep.add_argument(
+        "--dataset-json", required=True, type=Path, help="Path to cmor_dataset.json."
+    )
+    p_prep.add_argument(
+        "--outdir",
+        default=Path("out"),
+        type=Path,
+        help="Output directory for CMORized files.",
+    )
+    p_prep.set_defaults(
+        func=lambda a: prepare(
+            var=a.var,
+            in_files=a.in_file,
+            realm=a.realm,
+            cmor_tables=a.cmor_tables,
+            dataset_json=a.dataset_json,
+            outdir=a.outdir,
+        )
+    )
 
     args = p.parse_args(argv)
     args.func(args)
     return 0
 
+
 if __name__ == "__main__":
     raise SystemExit(main())

cmip7_prep/cmor_dataset.json

@@ -25,4 +25,4 @@
   "institution": "National Center for Atmospheric Research, Boulder, CO, USA",
   "tracking_id": "",
   "title": "CESM CMIP7 output prepared with cmip7_prep"
-}
+}

cmip7_prep/cmor_writer.py

@@ -1,11 +1,11 @@
-
 """Thin CMOR wrapper used by cmip7_prep.
 
 This module centralizes CMOR session setup and writing so that the rest of the
 pipeline can stay xarray-first. It supports either a dataset JSON file (preferred)
 or directly injected global attributes, and creates axes based on the coordinates
 present in the provided dataset.
 """
+
 from __future__ import annotations
 
 from contextlib import AbstractContextManager
@@ -88,10 +88,14 @@ def _define_axes(self, ds: xr.Dataset, vdef: Any) -> list[int]:
             cal = time.attrs.get("calendar", "standard")
             if np.issubdtype(time.dtype, np.datetime64):
                 # convert to numeric time using CF units if needed
-                tvals = xr.conventions.times.encode_cf_datetime(time.values, t_units, calendar=cal)
+                tvals = xr.conventions.times.encode_cf_datetime(
+                    time.values, t_units, calendar=cal
+                )
             else:
                 tvals = time.values
-            axes.append(cmor.axis(table_entry="time", units=str(t_units), coord_vals=tvals))
+            axes.append(
+                cmor.axis(table_entry="time", units=str(t_units), coord_vals=tvals)
+            )
 
         # Vertical axis (pressure or model levels)
         levels_info = getattr(vdef, "levels", None)
@@ -121,7 +125,9 @@ def _define_axes(self, ds: xr.Dataset, vdef: Any) -> list[int]:
             table_entry = "alev"
             if isinstance(levels_info, dict):
                 table_entry = levels_info.get("axis_entry", table_entry)
-            axes.append(cmor.axis(table_entry=table_entry, units="1", coord_vals=lev.values))
+            axes.append(
+                cmor.axis(table_entry=table_entry, units="1", coord_vals=lev.values)
+            )
 
         # Latitude / Longitude
         if "lat" in ds.coords and "lon" in ds.coords:
@@ -152,7 +158,9 @@ def _define_axes(self, ds: xr.Dataset, vdef: Any) -> list[int]:
     # public API
     # -------------------------
 
-    def write_variable(self, ds: xr.Dataset, varname: str, vdef: Any, outdir: Path) -> None:
+    def write_variable(
+        self, ds: xr.Dataset, varname: str, vdef: Any, outdir: Path
+    ) -> None:
         """Write one variable from `ds` to a CMOR-compliant NetCDF file.
 
         Parameters
@@ -162,8 +170,10 @@ def write_variable(self, ds: xr.Dataset, varname: str, vdef: Any, outdir: Path)
         varname : str
             Name of the variable in `ds` to CMORize.
         vdef : Any
-            An object with fields: ``name``, ``realm``, optional ``units``, ``positive``, and optional
-            ``levels`` dict (see :meth:`_define_axes`). This is typically a light-weight holder.
+            An object with fields: ``name``, ``realm``,
+            optional ``units``, ``positive``, and optional
+            ``levels`` dict (see :meth:`_define_axes`).
+            This is typically a light-weight holder.
         outdir : Path
             Output directory for the CMORized NetCDF file.
         """
@@ -182,8 +192,12 @@ def write_variable(self, ds: xr.Dataset, varname: str, vdef: Any, outdir: Path)
 
         axes_ids = self._define_axes(ds, vdef)
         units = getattr(vdef, "units", "")
-        var_id = cmor.variable(getattr(vdef, "name", varname), units, axes_ids,
-                               positive=getattr(vdef, "positive", None))
+        var_id = cmor.variable(
+            getattr(vdef, "name", varname),
+            units,
+            axes_ids,
+            positive=getattr(vdef, "positive", None),
+        )
 
         # Optional variable attributes (e.g., cell_methods)
         if getattr(vdef, "cell_methods", None):

cmip7_prep/dreq.py

@@ -1,4 +1,3 @@
-
 r"""Utilities for reading and querying a CMIP Data Request export.
 
 Supports Airtable-style JSON (with a top-level ``records`` list) and CSV exports.
@@ -12,6 +11,7 @@
 >>> 'tas' in vars_amon
 True
 """
+
 from __future__ import annotations
 
 import csv
@@ -60,8 +60,24 @@ class DReq:
 
     # Heuristics for common column names across exports
     _DEFAULT_TABLE_KEYS = ["Table", "table", "CMOR table", "CMOR_table"]
-    _DEFAULT_GROUP_KEYS = ["Group", "Groups", "Bucket", "MIP bucket", "Baseline group", "Notes", "Tags"]
-    _DEFAULT_SHORT_NAME_KEYS = ["Short name", "short_name", "Variable", "Name", "name", "CMIP Variable", "Compound name"]
+    _DEFAULT_GROUP_KEYS = [
+        "Group",
+        "Groups",
+        "Bucket",
+        "MIP bucket",
+        "Baseline group",
+        "Notes",
+        "Tags",
+    ]
+    _DEFAULT_SHORT_NAME_KEYS = [
+        "Short name",
+        "short_name",
+        "Variable",
+        "Name",
+        "name",
+        "CMIP Variable",
+        "Compound name",
+    ]
 
     def __init__(
         self,
@@ -76,7 +92,9 @@ def __init__(
         self.rows: List[Dict[str, Any]] = self._read(self.path)
         self.table_keys: List[str] = table_keys or self._DEFAULT_TABLE_KEYS
         self.group_keys: List[str] = group_keys or self._DEFAULT_GROUP_KEYS
-        self.short_name_keys: List[str] = short_name_keys or self._DEFAULT_SHORT_NAME_KEYS
+        self.short_name_keys: List[str] = (
+            short_name_keys or self._DEFAULT_SHORT_NAME_KEYS
+        )
         self.group_patterns = _compile_patterns(group_regexes)
 
     # ---------------------------