Add new "one-to-one" mode for modelling tides at multiple locations w…

…ith unique timesteps (#1142)

* Add analysis mode to model_tides, update tests

* Clarify which models need special definition file

* In progress update of notebook

* Update notebook, fix plot

Tests/dea_tools/test_coastal.py

@@ -131,7 +131,8 @@ def test_model_tides(measured_tides_ds, x, y, crs, method):
 
  # Test that modelled tides contain correct headings and have same
  # number of timesteps
- assert modelled_tides_df.columns.tolist() == ["x", "y", "tide_model", "tide_m"]
+ assert modelled_tides_df.index.names == ["time", "x", "y"]
+ assert modelled_tides_df.columns.tolist() == ["tide_model", "tide_m"]
  assert len(modelled_tides_df.index) == len(measured_tides_ds.time)
 
  # Test that modelled tides meet expected accuracy
@@ -169,7 +170,8 @@ def test_model_tides_multiplemodels(measured_tides_ds, models, output_format):
 
  if output_format == "long":
  # Verify output has correct columns
- assert modelled_tides_df.columns.tolist() == ["x", "y", "tide_model", "tide_m"]
+ assert modelled_tides_df.index.names == ["time", "x", "y"]
+ assert modelled_tides_df.columns.tolist() == ["tide_model", "tide_m"]
 
  # Verify tide model column contains correct values
  assert modelled_tides_df.tide_model.unique().tolist() == models
@@ -180,7 +182,8 @@ def test_model_tides_multiplemodels(measured_tides_ds, models, output_format):
 
  elif output_format == "wide":
  # Verify output has correct columns
- assert modelled_tides_df.columns[2:].tolist() == models
+ assert modelled_tides_df.index.names == ["time", "x", "y"]
+ assert modelled_tides_df.columns.tolist() == models
 
  # Verify output has same length as orginal timesteps
  assert len(modelled_tides_df.index) == len(measured_tides_ds.time)
@@ -209,6 +212,95 @@ def test_model_tides_units(measured_tides_ds, units, expected_range, expected_dt
  assert modelled_tides_df.tide_m.dtype == expected_dtype
 
 
+# Run test for each combination of mode, output format, and one or
+# multiple tide models
+@pytest.mark.parametrize(
+ "mode, models, output_format",
+ [
+ ("one-to-many", ["FES2014"], "long"),
+ ("one-to-one", ["FES2014"], "long"),
+ ("one-to-many", ["FES2014"], "wide"),
+ ("one-to-one", ["FES2014"], "wide"),
+ ("one-to-many", ["FES2014", "HAMTIDE11"], "long"),
+ ("one-to-one", ["FES2014", "HAMTIDE11"], "long"),
+ ("one-to-many", ["FES2014", "HAMTIDE11"], "wide"),
+ ("one-to-one", ["FES2014", "HAMTIDE11"], "wide"),
+ ],
+)
+def test_model_tides_mode(mode, models, output_format):
+ # Input params
+ x = [122.14, 122.30, 122.12]
+ y = [-17.91, -17.92, -18.07]
+ times = pd.date_range("2020", "2021", periods=3)
+
+ # Model tides
+ modelled_tides_df = model_tides(
+ x=x,
+ y=y,
+ time=times,
+ mode=mode,
+ output_format=output_format,
+ model=models,
+ )
+
+ if mode == "one-to-one":
+ if output_format == "wide":
+ # Should have the same number of rows as input x, y, times
+ assert len(modelled_tides_df.index) == len(x)
+ assert len(modelled_tides_df.index) == len(times)
+
+ # Output indexes should match order of input x, y, times
+ assert all(modelled_tides_df.index.get_level_values("time") == times)
+ assert all(modelled_tides_df.index.get_level_values("x") == x)
+ assert all(modelled_tides_df.index.get_level_values("y") == y)
+
+ elif output_format == "long":
+ # In "long" format, the number of x, y points multiplied by
+ # the number of tide models
+ assert len(modelled_tides_df.index) == len(x) * len(models)
+
+ # Verify index values match expected x, y, time order
+ assert all(
+ modelled_tides_df.index.get_level_values("time")
+ == np.tile(times, len(models))
+ )
+ assert all(
+ modelled_tides_df.index.get_level_values("x") == np.tile(x, len(models))
+ )
+ assert all(
+ modelled_tides_df.index.get_level_values("y") == np.tile(y, len(models))
+ )
+
+ if mode == "one-to-many":
+ if output_format == "wide":
+ # In "wide" output format, the number of rows should equal
+ # the number of x, y points multiplied by timesteps
+ assert len(modelled_tides_df.index) == len(x) * len(times)
+
+ # TODO: Work out what order rows should be returned in in
+ # "one-to-many" and "wide" mode
+
+ elif output_format == "long":
+ # In "long" output format, the number of rows should equal
+ # the number of x, y points multiplied by timesteps and
+ # the number of tide models
+ assert len(modelled_tides_df.index) == len(x) * len(times) * len(models)
+
+ # Verify index values match expected x, y, time order
+ assert all(
+ modelled_tides_df.index.get_level_values("time")
+ == np.tile(times, len(x) * len(models))
+ )
+ assert all(
+ modelled_tides_df.index.get_level_values("x")
+ == np.tile(np.repeat(x, len(times)), len(models))
+ )
+ assert all(
+ modelled_tides_df.index.get_level_values("y")
+ == np.tile(np.repeat(y, len(times)), len(models))
+ )
+
+
 # Run tests for default and custom resolutions
 @pytest.mark.parametrize("resolution", [None, "custom"])
 def test_pixel_tides(satellite_ds, measured_tides_ds, resolution):