Merge pull request #20 from kthyng/simplify_demo

Simplify demo

.github/workflows/test.yaml

@@ -8,8 +8,8 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: ["ubuntu-latest"]
-        python-version: ["3.7", "3.8"]
+        os: ["macos-latest", "ubuntu-latest", "windows-latest"]
+        python-version: ["3.7", "3.8", "3.9"]
     steps:
       - uses: actions/checkout@v2
       - name: Cache conda
@@ -22,9 +22,9 @@ jobs:
           key: ${{ runner.os }}-conda-${{ env.CACHE_NUMBER }}-${{ hashFiles('ci/environment-py${{ matrix.python-version }}.yml') }}
       - uses: conda-incubator/setup-miniconda@v2
         with:
-          # mamba-version: "*" # activate this to build with mamba.
-          # channels: conda-forge, defaults # These need to be specified to use mamba
-          # channel-priority: true
+          mamba-version: "*" # activate this to build with mamba.
+          channels: conda-forge, defaults # These need to be specified to use mamba
+          channel-priority: true
           environment-file: ci/environment-py${{ matrix.python-version }}.yml
 
           activate-environment: test_env_model_assessor

.pre-commit-config.yaml

@@ -41,7 +41,7 @@ repos:
     - id: seed-isort-config
 
 - repo: https://github.com/psf/black
-  rev: stable
+  rev: 22.3.0  # https://github.com/psf/black/issues/2964
   hooks:
   - id: black
     language_version: python3

.readthedocs.yml

@@ -0,0 +1,12 @@
+version: 2
+
+build:
+  os: "ubuntu-20.04"
+  tools:
+    python: "mambaforge-4.10"
+
+conda:
+    environment: docs/environment.yml
+
+sphinx:
+  configuration: docs/conf.py

ci/environment-py3.7.yml

@@ -7,7 +7,7 @@ dependencies:
   - cartopy
   - cf_xarray
   - extract_model
-  - matplotlib
+  - matplotlib-base
   - numpy
   - ocean_data_gateway
   - pandas

ci/environment-py3.8.yml

@@ -7,7 +7,7 @@ dependencies:
   - cartopy
   - cf_xarray
   - extract_model
-  - matplotlib
+  - matplotlib-base
   - numpy
   - ocean_data_gateway
   - pandas

ci/environment-py3.9.yml

@@ -0,0 +1,21 @@
+name: test_env_model_assessor
+channels:
+  - conda-forge
+dependencies:
+  - python=3.9
+  ############## These will have to be adjusted to your specific project
+  - cartopy
+  - cf_xarray
+  - extract_model
+  - matplotlib-base
+  - numpy
+  - ocean_data_gateway
+  - pandas
+  - scipy
+  - xarray
+  ##############
+  - pytest
+  - pip:
+    - codecov
+    - pytest-cov
+    - coverage[toml]

docs/environment.yml

@@ -8,14 +8,13 @@ dependencies:
    # - cartopy
    # - matplotlib
    # - ipython
-   - jupyter
+   # - jupyter
    # - jupyterlab
    - ocean_model_skill_assessor
    # - xarray
    # These are needed for the docs themselves
    - numpydoc
-   # https://stackoverflow.com/questions/68642540/nbsphinx-causes-build-to-fail-when-building-jupyter-notebooks
-   - sphinx==4.0.2
+   - sphinx
    - sphinx_rtd_theme
    - ipython
    - nbconvert

docs/index.rst

@@ -9,6 +9,7 @@ Welcome to ocean-model-skill-assessor's documentation!
 .. toctree::
    :maxdepth: 3
 
+   Demo
    Demo_workflows
    api
 

ocean_model_skill_assessor/main.py

@@ -203,6 +203,7 @@ def run(
     erddap=None,
     figname_map=None,
     figname_data_prefix="",
+    horizontal_interp=False,
     local=None,
     only_search=False,
     only_searchplot=False,
@@ -238,6 +239,9 @@ def run(
         Figure name for map showing data locations.
     figname_data_prefix : str, optional
         Prefix for figures for dataset-model comparisons.
+    horizontal_interp : bool, optional
+        If True, use `em.select()` to interpolate to the data location horizontally.
+        If False, use `em.sel2d()` to use the nearest grid point to the data location.
     local : dict, optional
         Inputs for local reader.
     only_search : boolean, optional
@@ -361,18 +365,39 @@ def run(
             continue
 
         for variable in variables:
-            kwargs = dict(
-                da=dsm.cf[variable]
-                .cf.isel(Z=0)
-                .cf.sel(lon=slice(lon - 5, lon + 5), lat=slice(lat - 5, lat + 5)),
-                longitude=lon,
-                latitude=lat,
-                T=T,
-                iZ=Z,
-                locstream=True,
-            )
 
-            model_var = em.select(**kwargs).to_dataset()
+            if horizontal_interp:
+                kwargs = dict(
+                    da=dsm.cf[variable]
+                    .cf.isel(Z=0)
+                    .cf.sel(lon=slice(lon - 5, lon + 5), lat=slice(lat - 5, lat + 5)),
+                    longitude=lon,
+                    latitude=lat,
+                    T=T,
+                    iZ=Z,
+                    locstream=True,
+                )
+
+                model_var = em.select(**kwargs).to_dataset()
+
+            else:
+                kwargs = dict(
+                    longitude=lon,
+                    latitude=lat,
+                    #                     T=T,
+                    Z=0,
+                    method="nearest",
+                )
+                if T is not None:
+                    kwargs["T"] = T
+
+                # xoak doesn't work for 1D lon/lat coords
+                if dsm.cf["longitude"].ndim == dsm.cf["latitude"].ndim == 1:
+                    model_var = dsm.cf[variable].cf.sel(**kwargs).to_dataset()
+
+                elif dsm.cf["longitude"].ndim == dsm.cf["latitude"].ndim == 2:
+                    model_var = dsm.cf[variable].em.sel2dcf(**kwargs).to_dataset()
+
             # Combine and align the two time series of variable
             df = omsa.stats._align(data.cf[variable], model_var.cf[variable])
             stats = df.omsa.compute_stats

ocean_model_skill_assessor/stats.py

@@ -115,7 +115,7 @@ def compute_mean_square_error(
     error = obs - model
     if centered:
         error += -obs.mean() + model.mean()
-    return (error ** 2).mean()
+    return (error**2).mean()
 
 
 def compute_murphy_skill_score(

tests/test_plot.py

@@ -7,12 +7,13 @@
 def test_time_series():
     ref_times = pd.date_range(start="2000-12-30", end="2001-01-03", freq="6H")
     reference = pd.DataFrame(
-        {"reference": np.sin(ref_times.view(int))}, index=ref_times
+        {"reference": np.sin(ref_times.values.astype("float32"))}, index=ref_times
     )
 
     sample_times = pd.date_range(start="2000-12-28", end="2001-01-04", freq="D")
     sample = pd.DataFrame(
-        {"FAKE_SAMPLES": np.sin(sample_times.view(int))}, index=sample_times
+        {"FAKE_SAMPLES": np.sin(sample_times.values.astype("float32"))},
+        index=sample_times,
     )
 
     time_series.plot(reference, sample, "test")

tests/test_stats.py

@@ -8,10 +8,12 @@
 
 class TestStats:
     ref_times = pd.date_range(start="2000-12-30", end="2001-01-03", freq="6H")
-    obs = pd.DataFrame({"obs": np.sin(ref_times.view(int))}, index=ref_times)
+    obs = pd.DataFrame(
+        {"obs": np.sin(ref_times.values.astype("float32"))}, index=ref_times
+    )
 
     model_times = pd.date_range(start="2000-12-28", end="2001-01-04", freq="D")
-    data = 1.25 * np.sin(model_times.view(int) + 2)
+    data = 1.25 * np.sin(model_times.values.astype("float32") + 2)
     model = pd.DataFrame({"model": data}, index=model_times)
 
     aligned_signals = stats._align(obs, model)
@@ -23,20 +25,20 @@ def test_align(self):
 
         assert isinstance(self.aligned_signals, pd.DataFrame)
         assert self.aligned_signals.shape == (5, 2)
-        assert np.isclose(self.aligned_signals["model"].mean(), 0.16391766802943322)
-        assert np.isclose(self.aligned_signals["obs"].mean(), 0.13113413442354657)
+        assert np.isclose(self.aligned_signals["model"].mean(), -0.35140932)
+        assert np.isclose(self.aligned_signals["obs"].mean(), -0.28112745)
 
     def test_align_xr(self):
 
         assert isinstance(self.aligned_signals_xr, pd.DataFrame)
         assert self.aligned_signals_xr.shape == (5, 2)
-        assert np.isclose(self.aligned_signals_xr["model"].mean(), 0.16391766802943322)
-        assert np.isclose(self.aligned_signals_xr["obs"].mean(), 0.13113413442354657)
+        assert np.isclose(self.aligned_signals_xr["model"].mean(), -0.35140932)
+        assert np.isclose(self.aligned_signals_xr["obs"].mean(), -0.28112745)
 
     def test_bias(self):
         bias = stats.compute_bias(self.obs, self.model)
 
-        assert np.isclose(bias, 0.032783533605886636)
+        assert np.isclose(bias, -0.07028185)
 
     def test_correlation_coefficient(self):
         corr_coef = stats.compute_correlation_coefficient(self.obs, self.model)
@@ -46,35 +48,35 @@ def test_correlation_coefficient(self):
     def test_index_of_agreement(self):
         ioa = stats.compute_index_of_agreement(self.obs, self.model)
 
-        assert np.isclose(ioa, 0.9872247131609245)
+        assert np.isclose(ioa, 0.9845252502709627)
 
     def test_mean_square_error(self):
         mse = stats.compute_mean_square_error(self.obs, self.model, centered=False)
 
-        assert np.isclose(mse, 0.03156566684130121)
+        assert np.isclose(mse, 0.024080737)
 
     def test_mean_square_error_centered(self):
         mse = stats.compute_mean_square_error(self.obs, self.model, centered=True)
 
-        assert np.isclose(mse, 0.03049090676561291)
+        assert np.isclose(mse, 0.019141199)
 
     def test_murphy_skill_score(self):
         mss = stats.compute_murphy_skill_score(self.obs, self.model)
 
-        assert np.isclose(mss, 0.935296965985732)
+        assert np.isclose(mss, 0.9213713780045509)
 
     def test_root_mean_square_error(self):
         rmse = stats.compute_root_mean_square_error(self.obs, self.model)
 
-        assert np.isclose(rmse, 0.17766729254790034)
+        assert np.isclose(rmse, 0.1551797)
 
     def test_descriptive_statistics(self):
         max, min, mean, std = stats.compute_descriptive_statistics(self.model, ddof=0)
 
-        assert np.isclose(max, 0.97866)
-        assert np.isclose(min, -0.906447)
-        assert np.isclose(mean, 0.039614)
-        assert np.isclose(std, 0.863681)
+        assert np.isclose(max, 0.882148)
+        assert np.isclose(min, -1.247736)
+        assert np.isclose(mean, -0.301843)
+        assert np.isclose(std, 0.757591)
 
     def test_stats(self):
         stats_output = stats.compute_stats(self.obs, self.model)