Merge branch 'main' into deprecate-calendar

CHANGES.rst

@@ -6,6 +6,10 @@ v0.50.0 (unreleased)
 --------------------
 Contributors to this version: Trevor James Smith (:user:`Zeitsperre`), Éric Dupuis (:user:`coxipi`).
 
+New features and enhancements
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+* New properties: Bivariate Spell Length (``xclim.sdba.properties.bivariate_spell_length``), generalized spell lengths with an argument for `window`, and specific spell lengths with `window` fixed to 1  (``xclim.sdba.propertiies.threshold_count``, ``xclim.sdba.propertiies.bivariate_threshold_count``). (:pull:`1758`).
+
 Breaking changes
 ^^^^^^^^^^^^^^^^
 * `pint` has been pinned below v0.24 until `xclim` can be updated to support the latest version. (:issue:`1771`, :pull:`1772`).

pyproject.toml

@@ -128,7 +128,7 @@ target-version = [
 ]
 
 [tool.bumpversion]
-current_version = "0.49.1-dev.2"
+current_version = "0.49.1-dev.3"
 commit = true
 commit_args = "--no-verify"
 tag = false

tests/test_sdba/test_properties.py

@@ -1,7 +1,10 @@
 from __future__ import annotations
 
 import numpy as np
+import pandas as pd
 import pytest
+import xarray as xr
+from xarray import set_options
 
 from xclim import sdba
 from xclim.core.units import convert_units_to
@@ -107,80 +110,135 @@ def test_quantile(self, open_dataset):
         )
         assert out_season.long_name.startswith("Quantile 0.2")
 
+    # TODO: test theshold_count? it's the same a test_spell_length_distribution
     def test_spell_length_distribution(self, open_dataset):
-        sim = (
+        ds = (
             open_dataset("sdba/CanESM2_1950-2100.nc")
             .sel(time=slice("1950", "1952"), location="Vancouver")
-            .pr
-        ).load()
+            .load()
+        )
 
-        tmean = (
-            sdba.properties.spell_length_distribution(sim, op="<", group="time.month")
+        # test pr, with amount method
+        sim = ds.pr
+        kws = {"op": "<", "group": "time.month"}
+        outd = {
+            stat: sdba.properties.spell_length_distribution(da=sim, **kws, stat=stat)
             .sel(month=1)
             .values
+            for stat in ["mean", "max", "min"]
+        }
+        np.testing.assert_array_almost_equal(
+            [outd[k] for k in ["mean", "max", "min"]], [2.44127, 10, 1]
+        )
+
+        # test tasmax, with quantile method
+        simt = ds.tasmax
+        kws = {"thresh": 0.9, "op": ">=", "method": "quantile", "group": "time.month"}
+        outd = {
+            stat: sdba.properties.spell_length_distribution(
+                da=simt, **kws, stat=stat
+            ).sel(month=6)
+            for stat in ["mean", "max", "min"]
+        }
+        np.testing.assert_array_almost_equal(
+            [outd[k].values for k in ["mean", "max", "min"]], [3.0, 6, 1]
         )
-        tmax = (
-            sdba.properties.spell_length_distribution(
-                sim, op="<", group="time.month", stat="max"
-            )
-            .sel(month=1)
-            .values
+
+        # test varia
+        with pytest.raises(
+            ValueError,
+            match="percentile is not a valid method. Choose 'amount' or 'quantile'.",
+        ):
+            sdba.properties.spell_length_distribution(simt, method="percentile")
+
+        assert (
+            outd["mean"].long_name
+            == "Average of spell length distribution when the variable is >= the quantile 0.9 for 1 consecutive day(s)."
         )
-        tmin = (
-            sdba.properties.spell_length_distribution(
-                sim, op="<", group="time.month", stat="min"
-            )
-            .sel(month=1)
-            .values
+
+    def test_spell_length_distribution_mixed_stat(self, open_dataset):
+
+        time = pd.date_range("2000-01-01", periods=2 * 365, freq="D")
+        tas = xr.DataArray(
+            np.array([0] * 365 + [40] * 365),
+            dims=("time"),
+            coords={"time": time},
+            attrs={"units": "degC"},
         )
 
-        np.testing.assert_array_almost_equal([tmean, tmax, tmin], [2.44127, 10, 1])
+        kws_sum = dict(
+            thresh="30 degC", op=">=", stat="sum", stat_resample="sum", group="time"
+        )
+        out_sum = sdba.properties.spell_length_distribution(tas, **kws_sum).values
+        kws_mixed = dict(
+            thresh="30 degC", op=">=", stat="mean", stat_resample="sum", group="time"
+        )
+        out_mixed = sdba.properties.spell_length_distribution(tas, **kws_mixed).values
 
-        simt = (
-            open_dataset("sdba/CanESM2_1950-2100.nc")
-            .sel(time=slice("1950", "1952"), location="Vancouver")
-            .tasmax
-        ).load()
+        assert out_sum == 365
+        assert out_mixed == 182.5
 
-        tmean = sdba.properties.spell_length_distribution(
-            simt, op=">=", group="time.month", method="quantile", thresh=0.9
-        ).sel(month=6)
-        tmax = (
-            sdba.properties.spell_length_distribution(
-                simt,
-                op=">=",
-                group="time.month",
-                stat="max",
-                method="quantile",
-                thresh=0.9,
+    @pytest.mark.parametrize(
+        "window,expected_amount,expected_quantile",
+        [
+            (1, [2.333333, 4, 1], [3, 6, 1]),
+            (3, [1.333333, 4, 0], [2, 6, 0]),
+        ],
+    )
+    def test_bivariate_spell_length_distribution(
+        self, open_dataset, window, expected_amount, expected_quantile
+    ):
+        ds = (
+            open_dataset("sdba/CanESM2_1950-2100.nc").sel(
+                time=slice("1950", "1952"), location="Vancouver"
             )
-            .sel(month=6)
+        ).load()
+        tx = ds.tasmax
+        with set_options(keep_attrs=True):
+            tn = tx - 5
+
+        # test with amount method
+        kws = {
+            "thresh1": "0 degC",
+            "thresh2": "0 degC",
+            "op1": ">",
+            "op2": "<=",
+            "group": "time.month",
+            "window": window,
+        }
+        outd = {
+            stat: sdba.properties.bivariate_spell_length_distribution(
+                da1=tx, da2=tn, **kws, stat=stat
+            )
+            .sel(month=1)
             .values
-        )
-        tmin = (
-            sdba.properties.spell_length_distribution(
-                simt,
-                op=">=",
-                group="time.month",
-                stat="min",
-                method="quantile",
-                thresh=0.9,
+            for stat in ["mean", "max", "min"]
+        }
+        np.testing.assert_array_almost_equal(
+            [outd[k] for k in ["mean", "max", "min"]], expected_amount
+        )
+
+        # test with quantile method
+        kws = {
+            "thresh1": 0.9,
+            "thresh2": 0.9,
+            "op1": ">",
+            "op2": ">",
+            "method1": "quantile",
+            "method2": "quantile",
+            "group": "time.month",
+            "window": window,
+        }
+        outd = {
+            stat: sdba.properties.bivariate_spell_length_distribution(
+                da1=tx, da2=tn, **kws, stat=stat
             )
             .sel(month=6)
             .values
-        )
-
-        np.testing.assert_array_almost_equal([tmean.values, tmax, tmin], [3.0, 6, 1])
-
-        with pytest.raises(
-            ValueError,
-            match="percentile is not a valid method. Choose 'amount' or 'quantile'.",
-        ):
-            sdba.properties.spell_length_distribution(simt, method="percentile")
-
-        assert (
-            tmean.long_name
-            == "Average of spell length distribution when the variable is >= the quantile 0.9."
+            for stat in ["mean", "max", "min"]
+        }
+        np.testing.assert_array_almost_equal(
+            [outd[k] for k in ["mean", "max", "min"]], expected_quantile
         )
 
     def test_acf(self, open_dataset):

xclim/__init__.py

@@ -13,7 +13,7 @@
 
 __author__ = """Travis Logan"""
 __email__ = "logan.travis@ouranos.ca"
-__version__ = "0.49.1-dev.2"
+__version__ = "0.49.1-dev.3"
 
 
 with _resources.as_file(_resources.files("xclim.data")) as _module_data: