In-memory chaining of CLI commands

chain/chain.py

@@ -5,6 +5,14 @@
 import xclim
 
 
+"""
+This CLI demonstrates a workflow where commands can be called individually
+or chained together in memory. That is, each individual command accepts and
+returns an xarray.Dataset. The `click.group` is responsible for readingin the
+input data from disk, and writing the output data to disk.
+"""
+
+
 @click.group(chain=True, help="Chained CLI", invoke_without_command=True)
 @click.argument("input", type=click.File("r"))
 @click.argument("output")
@@ -13,15 +21,18 @@ def cli(input, output):
 
 @cli.result_callback()
 def process_pipeline(processors, input, output):
-    click.echo(f"INPUT: {input.name}")
-    click.echo(f"OUTPUT: {output}")
+    """Read the input, execute commands in memory, write output to disk."""
+    # Read the input data into an xarray.Dataset
     ds = xr.open_dataset(input.name, engine="h5netcdf")
 
+    # Execute individual commands - not obvious here, but options are passed
     for processor in processors:
         ds = processor(ds)
 
+    # Write output to disk
     ds.to_netcdf(output, engine="h5netcdf")
 
+
 @cli.command
 @click.option("-p", "--poly", help="Path to the polygon shapefile.")
 @click.option("-s",
@@ -34,7 +45,6 @@ def subset(**kwargs):
     """Subset on polygon"""
     def processor(ds):
         gdf = gpd.GeoDataFrame.from_file(kwargs["poly"])
-        # buffer = kwargs["buffer"]
         return clisops.core.subset_shape(ds=ds, 
                                          shape=gdf, 
                                          start_date=kwargs["start"], 

chain/test.py

@@ -7,24 +7,27 @@
 from chain import cli
 
 
-@pytest.fixture
+@pytest.fixture(scope="module")
 def tas_series():
     """Return mean temperature time series."""
     _tas_series = partial(tt, variable="tas")
     return _tas_series
 
 
-def test_chain(tas_series, tmp_path):
+@pytest.fixture(scope="module")
+def input_file(tas_series, tmpdir_factory):
     # Create input file
+    input_file = str(tmpdir_factory.mktemp("input").join("in.nc"))
     tas = tas_series(np.ones(366) + 271.15, start="1/1/2000")
     tas = tas.expand_dims(dim={"lon": np.linspace(-80, -70, 10), "lat": np.linspace(40, 50, 10)},)
     tas.lon.attrs["standard_name"] = "longitude"
     tas.lat.attrs["standard_name"] = "latitude"
     ds = xr.Dataset(data_vars={"tas": tas})
-
-    input_file = tmp_path / "in.nc"
     ds.to_netcdf(input_file, engine="h5netcdf")
+    return input_file
+
 
+def test_chain(input_file, tmp_path):
     output_file = tmp_path / "out.nc"
     args = [str(input_file), 
             str(output_file),
@@ -38,9 +41,52 @@ def test_chain(tas_series, tmp_path):
     assert results.exit_code == 0
 
     if output_file.exists():
-        print(output_file)
         out = xr.open_dataset(str(output_file), engine="h5netcdf")
         outvar = list(out.data_vars.values())[0]
         np.testing.assert_allclose(outvar[0], 6588.0)
     else:
-        raise FileNotFoundError
+        raise FileNotFoundError
+
+
+def test_hdd(input_file, tmp_path):    
+    output_file = tmp_path / "out.nc"
+    args = [str(input_file), 
+            str(output_file),
+            "hdd", 
+            "--thresh", "17 degC"]
+    runner = CliRunner()
+    results = runner.invoke(cli, args)
+
+    assert results.exit_code == 0
+
+    if output_file.exists():
+        out = xr.open_dataset(str(output_file), engine="h5netcdf")
+        outvar = list(out.data_vars.values())[0]
+        np.testing.assert_allclose(outvar[0], 6588.0)
+    else:
+        raise FileNotFoundError
+
+
+def test_subset(input_file, tmp_path):
+    output_file = tmp_path / "out.nc"
+    args = [str(input_file), 
+            str(output_file),
+            "subset", 
+            "-p", "small_geojson.json",
+            "-s", "2000-06",
+            "-e", "2000-08",]
+    runner = CliRunner()
+    results = runner.invoke(cli, args)
+
+    assert results.exit_code == 0
+
+    if output_file.exists():
+        out = xr.open_dataset(str(output_file), engine="h5netcdf")
+        assert "tas" in out.data_vars
+        assert out.time.isel(time=0) == np.datetime64("2000-06-01T00:00:00")
+        assert out.time.isel(time=-1) == np.datetime64("2000-08-31T00:00:00")
+        assert len(out.tas.lon) == 1
+        assert len(out.tas.lat) == 1
+    else:
+        raise FileNotFoundError
+