windeval/__init__.py

-from . import toolbox, processing, plotting
-from .processing import conversions, diagnostics
-from .io.api import open_product, select, info, plot, report
+from . import processing, plotting
+from .io import api as io
+from .processing import diagnostics, conversions
+from .plotting import plot
+from .io.api import open_product, save_product, select, info, report
 
 __all__ = [
     # modules
-    "toolbox",
+    "api",
+    "io",
     "processing",
     "plotting",
-    # functions
+    # core functions
     "open_product",
+    "save_product",
     "info",
     "select",
     "conversions",

windeval/io/api.py

 """
-Data IO api
+Data IO api.
 """
 
 # flake8: noqa
 
-from .products import open_product, info, select
-from ..plotting import plot
+from .products import open_product, save_product, info, select
 from .reports import report

windeval/io/products.py

 import xarray as xr
+from pathlib import Path
 from typing import Any, Dict
 
 
-def open_product(path1, path2, **kwargs: Dict[str, Any]) -> Dict[str, xr.Dataset]:
-    raise NotImplementedError("Import of data through Intake is not yet implemented.")
+def open_product(
+    path0: str,
+    path1: str,
+    *args: Any,
+    experimental: bool = False,
+    **kwargs: Dict[str, Any]
+) -> Dict[str, xr.Dataset]:
+    if not args:
+        names = [Path(path0).stem, Path(path1).stem]
+    else:
+        names = [*args]
+    if experimental:
+        ds = {names[0]: xr.open_dataset(path0), names[1]: xr.open_dataset(path1)}
+    else:
+        raise NotImplementedError(
+            "Import of data through Intake is not yet implemented."
+        )
+
+    return ds
+
+
+def save_product(
+    ds: Dict[str, xr.Dataset],
+    path: str,
+    experimental: bool = False,
+    **kwargs: Dict[str, Any]
+) -> None:
+    if experimental:
+        for k, v in ds.items():
+            v.to_netcdf(Path(path).joinpath(k + ".cdf"))
+    else:
+        raise NotImplementedError("Export of data is not yet implemented.")
+
+    return None
 
 
 def info(wndpr: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]) -> None:

windeval/io/reports.py

@@ -2,5 +2,5 @@ import xarray as xr
 from typing import Any, Dict
 
 
-def report(wndpr: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]) -> None:
+def report(wnd: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]) -> None:
     raise NotImplementedError("Generation of reports is not yet implemented.")

windeval/plotting.py

 import xarray as xr
-from typing import Dict, Any
+from functools import singledispatch
+from typing import Dict, Any, Union, List, Optional
+import matplotlib.pyplot as plt
 
 
-def plot(wndpr: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]):
-    raise NotImplementedError("Plotting is not yet implemented.")
+class Plot:
+    @staticmethod
+    def power_spectral_density(
+        wnd: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]
+    ) -> None:
+        for wndkey in wnd.keys():
+            plt.semilogx(wnd[wndkey].power_spectral_density)
+
+        plt.legend(wnd.keys())
+        plt.title("PSD: power spectral density")
+        plt.xlabel("Frequency")
+        plt.ylabel("Power")
+        plt.tight_layout()
+        plt.grid()
+
+        return None
+
+
+@singledispatch
+def plot(*args, **kwargs):
+    raise NotImplementedError("Data type not supported for conversion.")
+
+
+@plot.register
+def _(
+    wnddict: dict,
+    var: str,
+    *args: Any,
+    dataset: Optional[Union[str, List[str]]] = None,
+    **kwargs: Dict[str, Any]
+) -> None:
+
+    if dataset is None:
+        dataset = []
+    else:
+        raise NotImplementedError(
+            "Specifing a dataset for plots is not implemented yet."
+        )
+
+    if getattr(Plot, var, None) is not None:
+        getattr(Plot, var)(wnddict, *dataset, *args, **kwargs)
+    else:
+        for wndkey in wnddict.keys():
+            wnddict[wndkey][var].plot()
+
+    return None

windeval/processing.py

-"""Preprocessing module."""
+"""
+Preprocessing module.
+"""
 
 import numpy as np
 import xarray as xr
-from typing import Callable, Union, Optional, Dict, Any
+from functools import singledispatch
+from scipy import signal
+from typing import Callable, Union, Optional, Dict, Any, Iterable
 
 
 class BulkFormula:
@@ -602,13 +606,74 @@ def sverdrup_transport(X: xr.Dataset) -> xr.Dataset:
     return X
 
 
-def conversions(
-    wndpr: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]
-) -> Dict[str, xr.Dataset]:
+class Conversions:
+    def __init__(self):
         raise NotImplementedError("Conversions are not yet implemented.")
 
 
-def diagnostics(
-    wndpr: Dict[str, xr.Dataset], *args: Any, **kwargs: Dict[str, Any]
+@singledispatch
+def conversions(*args, **kwargs):
+    raise NotImplementedError("Data type not supported for conversion.")
+
+
+class Diagnostics:
+    @staticmethod
+    def welch(da: xr.DataArray, *args: Any, **kwargs: Dict[str, Any]) -> xr.Dataset:
+        f, psd = signal.welch(da.values.flat, *args, **kwargs)
+        ds = xr.Dataset(
+            {"power_spectral_density": ("frequency", psd)},
+            coords={"frequency": (["frequency"], f)},
+        )
+
+        return ds
+
+
+@singledispatch
+def diagnostics(*args, **kwargs):
+    raise NotImplementedError("Data type not supported.")
+
+
+@diagnostics.register
+def _(da: xr.DataArray, diag: str, *args: Any, **kwargs: Dict[str, Any]) -> xr.Dataset:
+    if getattr(Diagnostics, diag, None) is not None:
+        y = getattr(Diagnostics, diag)(da, *args, **kwargs)
+    elif getattr(da, diag, None) is not None:
+        y = getattr(da, diag)(*args, **kwargs)
+    else:
+        raise ValueError("Unknown diagnostic.")
+
+    return y
+
+
+@diagnostics.register  # type: ignore
+def _(
+    ds: xr.Dataset, var: str, diag: str, *args: Any, **kwargs: Dict[str, Any]
+) -> xr.Dataset:
+    ds = ds.merge(diagnostics(ds[var], diag, *args, **kwargs))
+
+    return ds
+
+
+@diagnostics.register  # type: ignore
+def _(
+    wnddict: dict,
+    var: str,
+    diag: str,
+    *args: Any,
+    dataset: Optional[Iterable] = None,
+    **kwargs: Dict[str, Any]
 ) -> Dict[str, xr.Dataset]:
-    raise NotImplementedError("Diagnostics are not yet implemented.")
+
+    if dataset is None:
+        dataset = wnddict.keys()
+    else:
+        raise NotImplementedError(
+            "Specifing a dataset for diagnostics is not implemented yet."
+        )
+
+    for wndkey in dataset:
+        wnddict[wndkey] = wnddict[wndkey].merge(
+            diagnostics(wnddict[wndkey][var], diag, *args, **kwargs)
+        )
+
+    return wnddict

windeval/tests/test_plotting.py

 import pytest
-from windeval import plotting
+from windeval import plotting, processing
 
 
-def test_report(X):
+def test_plot(X):
+    plotting.plot({"ds": X}, "eastward_wind")
     with pytest.raises(NotImplementedError):
-        plotting.plot({"ds": X})
+        plotting.plot(False)
+    with pytest.raises(NotImplementedError):
+        plotting.plot({"ds": X}, "eastward_wind", dataset=[])
+    ds = processing.diagnostics({"ds": X}, "eastward_wind", "welch")
+    plotting.plot(ds, "power_spectral_density")

windeval/tests/test_processing.py

@@ -190,5 +190,5 @@ def test_conversions(X):
 
 
 def test_diagnostics(X):
-    with pytest.raises(NotImplementedError):
-        processing.diagnostics({"ds": X})
+    ds = processing.diagnostics({"ds": X}, "eastward_wind", "welch")
+    assert isinstance(ds["ds"]["power_spectral_density"], xr.DataArray)

windeval/wrapper.py

-"""Wrapper."""
-
 import xarray as xr
-from functools import singledispatch
-from typing import Optional, Any, Dict, List
+from typing import Optional
 from . import processing
 
 
-@singledispatch
-def calculate(ds, *args, **kwargs):
-    raise NotImplementedError("Data type not supported.")
-
-
-@calculate.register
-def _(
-    ds: xr.Dataset, var: str, diag: str, *args: Any, **kwargs: Dict[str, Any]
-) -> xr.DataArray:
-    return calculate(ds[var], diag, *args, **kwargs)
-
-
-@calculate.register  # type: ignore
-def _(
-    da: xr.DataArray, diag: str, *args: Any, **kwargs: Dict[str, Any]
-) -> xr.DataArray:
-
-    f = getattr(processing, diag, None)
-    if f is not None:
-        y = f(da, *args, **kwargs)
-    else:
-        y = getattr(da, diag)(*args, **kwargs)
-
-    return y
-
-
-@calculate.register  # type: ignore
-def _(
-    d: dict, keys: List[str], diag: str, *args: Any, **kwargs: Dict[str, Any]
-) -> xr.DataArray:
-
-    f = getattr(processing, diag)
-    return f(*[d[k] for k in keys], *args, **kwargs)
-
-
 def ekman(
     X: xr.Dataset,
     component: str = "eastward",