Merge pull request #169 from ClimateImpactLab/climate_format_update

Climate input files format update

Author: kemccusker

CHANGELOG.md

@@ -15,6 +15,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Dropped optional/unused dependencies `click`, `dask-jobqueue`, `geopandas`, `gurobipy`, `ipywidgets`, `seaborn`. ([PR #99](https://github.com/ClimateImpactLab/dscim/pull/99), [@brews](https://github.com/brews))
 - Switch build system from `setuptools` to `hatchling`. ([PR #128](https://github.com/ClimateImpactLab/dscim/pull/128), [@brews](https://github.com/brews))
 - Clean up unit test for `dscim.utils.utils.c_equivalence`. ([PR #135](https://github.com/ClimateImpactLab/dscim/pull/135), [@brews](https://github.com/brews))
+- Reformat gmst/gmsl pulse files by removing unnecessary dimensions and indices. ([PR #169](https://github.com/ClimateImpactLab/dscim/pull/169), [@JMGilbert](https://github.com/JMGilbert))
 
 ### Fixed
 

src/dscim/menu/simple_storage.py

@@ -131,39 +131,8 @@ def gmsl_anomalies(self):
         """This function takes coastal sector's GMSL relative to 1991-2009.
         No rebasing occurs, as coastal damages are rebased to the same period.
         """
-        df = xr.open_zarr(self.gmsl_fair_path)
-        df = df.chunk(df.dims)
-
-        datasets = []
-
-        # collapse runtype dimension into two variables, and label
-        # each one (control, pulse) for medians and full simulations
-        for var in df.keys():
-            ds = df[var].to_dataset(dim="runtype")
-            ds = ds.rename({k: f"{k}_{var}" for k in ds.keys()})
-            datasets.append(ds)
-
-        anomaly = xr.combine_by_coords(datasets, combine_attrs="override")
-
-        # drop unnecessary coordinates
-        anomaly = anomaly.drop_vars(
-            ["confidence", "kind", "locations", "workflow_src"],
-            errors="ignore",
-        )
-
-        # rename variables
-        if (
-            "pulse_gmsl_median" in anomaly.keys()
-            and "control_gmsl_median" in anomaly.keys()
-        ):
-            anomaly = anomaly.rename(
-                {
-                    "pulse_gmsl_median": "medianparams_pulse_gmsl",
-                    "control_gmsl_median": "medianparams_control_gmsl",
-                }
-            )
-        else:
-            pass
+        anomaly = xr.open_dataset(self.gmsl_fair_path)
+        anomaly = anomaly.chunk(anomaly.dims)
 
         return anomaly
 

tests/conftest.py

@@ -22,7 +22,7 @@ def climate():
         gmsl_path=os.path.join(data_dir, "gmsl.csv"),
         gmst_path=os.path.join(datadir, "gmst.csv"),
         gmst_fair_path=os.path.join(datadir, "fair_temps_sims.nc4"),
-        gmsl_fair_path=os.path.join(datadir, "fair_slr_sims.zarr"),
+        gmsl_fair_path=os.path.join(datadir, "fair_slr_sims.nc4"),
         damages_pulse_conversion_path=os.path.join(
             datadir,
             "scenario_rcp45-rcp85-ssp245-ssp460-ssp370_CO2_Fossil-CH4-N2O_conversion_pulseyears_2020-2100_5yrincrements_v3.0_newformat.nc",

tests/data/climate/fair_slr_sims.nc4

@@ -159,20 +159,22 @@ def test_climate_gmsl_anomalies(tmp_path):
     d.mkdir()
     infile_path = d / "gmsl_fair.zarr"
     gases = ["CO2_Fossil", "CH4", "N2O"]
-    # This struction might appear weird, but matches that used for initial EPA runs.
+
     ds_in = xr.Dataset(
         {
-            "gmsl": (
-                ["runtype", "pulse_year", "simulation", "runid", "year", "gas"],
-                np.ones((2, 1, 1, 2, 1, 3)),
+            "pulse_gmsl": (
+                ["pulse_year", "runid", "year", "gas"],
+                np.ones((1, 2, 1, 3)),
+            ),
+            "control_gmsl": (
+                ["runid", "year"],
+                np.ones((2, 1)),
             ),
         },
         coords={
             "gas": (["gas"], gases),
             "pulse_year": (["pulse_year"], [2020]),
             "runid": (["runid"], [1, 2]),
-            "runtype": (["runtype"], ["control", "pulse"]),
-            "simulation": (["simulation"], [1]),
             "year": (["year"], [2020]),
         },
     )
@@ -181,19 +183,18 @@ def test_climate_gmsl_anomalies(tmp_path):
     expected = xr.Dataset(
         {
             "pulse_gmsl": (
-                ["pulse_year", "simulation", "runid", "year", "gas"],
-                np.ones((1, 1, 2, 1, 3)),
+                ["pulse_year", "runid", "year", "gas"],
+                np.ones((1, 2, 1, 3)),
             ),
             "control_gmsl": (
-                ["pulse_year", "simulation", "runid", "year", "gas"],
-                np.ones((1, 1, 2, 1, 3)),
+                ["runid", "year"],
+                np.ones((2, 1)),
             ),
         },
         coords={
             "gas": (["gas"], gases),
             "pulse_year": (["pulse_year"], [2020]),
             "runid": (["runid"], [1, 2]),
-            "simulation": (["simulation"], [1]),
             "year": (["year"], [2020]),
         },
     )
@@ -214,66 +215,6 @@ def test_climate_gmsl_anomalies(tmp_path):
     xr.testing.assert_equal(actual, expected)
 
 
-@pytest.mark.parametrize(
-    "unnecesary_coord", ["confidence", "kind", "locations", "workflow_src"]
-)
-def test_climate_gmsl_anomalies_dropcoords(tmp_path, unnecesary_coord):
-    """
-    Test that Climate instances give "gmsl_anomalies" with "unnecessary coords" dropped
-    """
-    # Set up input data in temporary directory because Climate needs to read
-    # from a magical Zarr Store.
-    d = tmp_path / "climate"
-    d.mkdir()
-    infile_path = d / "gmsl_fair.zarr"
-    gases = ["CO2_Fossil", "CH4", "N2O"]
-    # This struction might appear weird, but matches that used for initial EPA runs.
-    ds_in = xr.Dataset(
-        {
-            "gmsl": (
-                [
-                    unnecesary_coord,
-                    "runtype",
-                    "pulse_year",
-                    "simulation",
-                    "runid",
-                    "year",
-                    "gas",
-                ],
-                np.ones((1, 2, 1, 1, 2, 1, 3)),
-            ),
-        },
-        coords={
-            "gas": (["gas"], gases),
-            "pulse_year": (["pulse_year"], [2020]),
-            "runid": (["runid"], [1, 2]),
-            "runtype": (["runtype"], ["control", "pulse"]),
-            "simulation": (["simulation"], [1]),
-            "year": (["year"], [2020]),
-            unnecesary_coord: (
-                [unnecesary_coord],
-                ["foobar"],
-            ),  # This one is unnecessary.
-        },
-    )
-    ds_in.to_zarr(infile_path)
-
-    clim = Climate(
-        gmst_path="",
-        gmsl_path="",
-        gmst_fair_path="bacon",
-        gmsl_fair_path=str(infile_path),
-        pulse_year=2020,
-        damages_pulse_conversion_path="bacon",
-        ecs_mask_path=None,
-        emission_scenarios=None,
-        gases=gases,
-    )
-    actual = clim.gmsl_anomalies
-
-    assert unnecesary_coord not in actual.coords
-
-
 def test_climate_gmst_anomalies(tmp_path):
     """
     Test that Climate instances give "gmst_anomalies" from a GMST FAIR NetCDF file path
@@ -418,20 +359,22 @@ def test_climate_anomalies(tmp_path):
     gmsl_infile_path = d / "gmsl_fair.zarr"
     gases_gmsl = ["CO2_Fossil", "CH4", "N2O"]
     gmsl_infile_path = d / "gmsl_fair2.zarr"
-    # As usual, this struction might appear weird, but matches that used for initial EPA runs.
+
     gmsl_in = xr.Dataset(
         {
-            "gmsl": (
-                ["runtype", "pulse_year", "simulation", "runid", "year", "gas"],
-                np.ones((2, 1, 1, 2, 3, 3)),
+            "pulse_gmsl": (
+                ["pulse_year", "runid", "year", "gas"],
+                np.ones((1, 2, 3, 3)),
+            ),
+            "control_gmsl": (
+                ["runid", "year"],
+                np.ones((2, 3)),
             ),
         },
         coords={
             "gas": (["gas"], gases_gmsl),
             "pulse_year": (["pulse_year"], [2002]),
             "runid": (["runid"], [1, 2]),
-            "runtype": (["runtype"], ["control", "pulse"]),
-            "simulation": (["simulation"], [1]),
             "year": (["year"], [2000, 2001, 2002]),
         },
     )
@@ -442,18 +385,20 @@ def test_climate_anomalies(tmp_path):
             "control_temperature": (["gas", "runid", "year"], np.zeros((3, 2, 1))),
             "pulse_temperature": (["gas", "runid", "year"], np.zeros((3, 2, 1))),
             "control_gmsl": (
-                ["simulation", "runid", "year", "gas"],
-                np.ones((1, 2, 1, 3)),
+                [
+                    "runid",
+                    "year",
+                ],
+                np.ones((2, 1)),
             ),
             "pulse_gmsl": (
-                ["simulation", "runid", "year", "gas"],
-                np.ones((1, 2, 1, 3)),
+                ["runid", "year", "gas"],
+                np.ones((2, 1, 3)),
             ),
         },
         coords={
             "year": (["year"], [2002]),
             "gas": (["gas"], gases_gmst),
-            "simulation": (["simulation"], [1]),
             "runid": (["runid"], [1, 2]),
         },
     ).chunk({"year": 11})