more unit test

tests/test_classification.py

@@ -5,7 +5,7 @@
 import pytest
 import xarray as xr
 
-from pydepsi.classification import _idx_within_distance, _nad_block, _nmad_block, ps_selection
+from pydepsi.classification import _idx_within_distance, _nad_block, _nmad_block, network_stm_seletcion, ps_selection
 
 # Create a random number generator
 rng = np.random.default_rng(42)
@@ -73,6 +73,97 @@ def test_ps_seletion_not_implemented():
         ps_selection(slcs, 0.5, method="not_implemented", output_chunks=5)
 
 
+def test_network_stm_seletcion_results():
+    stm = xr.Dataset(
+        data_vars={
+            "amplitude": (("space", "time"), np.ones((100, 10))),
+            "pnt_nad": (("space"), np.linspace(0, 1, 100)),
+            "pnt_nmad": (("space"), np.linspace(0, 1, 100)),
+        },
+        coords={"azimuth": (("space"), np.arange(100)), "range": (("space"), np.arange(100)), "time": np.arange(10)},
+    )
+    res_nad = network_stm_seletcion(stm, min_dist=20, sortby_var="pnt_nad", azimuth_spacing=10, range_spacing=10)
+    res_nmad = network_stm_seletcion(stm, min_dist=20, sortby_var="pnt_nmad", azimuth_spacing=10, range_spacing=10)
+    # Fields should remain the same
+    assert "pnt_nad" in res_nad
+    assert "azimuth" in res_nad
+    assert "range" in res_nad
+    assert "space" in res_nad.dims
+    assert "time" in res_nad.dims
+    # Dimensions should be half
+    assert res_nad.sizes["space"] == 50
+    assert res_nad.sizes["time"] == 10
+    assert res_nmad.sizes["space"] == 50
+    assert res_nmad.sizes["time"] == 10
+
+
+def test_network_stm_seletcion_quality():
+    stm = xr.Dataset(
+        data_vars={
+            "amplitude": (("space", "time"), np.ones((5, 10))),
+            "pnt_nad": (("space"), np.array([0.9, 0.01, 0.9, 0.9, 0.01])),
+            "pnt_nmad": (("space"), np.array([0.01, 0.9, 0.9, 0.9, 0.01])),
+        },
+        coords={
+            "space": np.array([3, 1, 2, 5, 7]),  # non monotonic space coords
+            "time": np.arange(10),
+            "azimuth": (("space"), np.arange(5)),
+            "range": (("space"), np.arange(5)),
+        },
+    )
+    res_nad = network_stm_seletcion(stm, min_dist=3, sortby_var="pnt_nad", azimuth_spacing=1, range_spacing=1)
+    res_nmad = network_stm_seletcion(stm, min_dist=3, sortby_var="pnt_nmad", azimuth_spacing=1, range_spacing=1)
+
+    # The two pixels with the lowest NAD should be selected
+    assert np.all(np.isclose(res_nad["pnt_nad"].values, 0.01, rtol=1e-09, atol=1e-09))
+    assert np.all(np.isclose(res_nmad["pnt_nmad"].values, 0.01, rtol=1e-09, atol=1e-09))
+    assert np.all(res_nad["space"].values == np.array([1, 7]))
+    assert np.all(res_nmad["space"].values == np.array([3, 7]))
+
+
+def test_network_stm_seletcion_include_index():
+    stm = xr.Dataset(
+        data_vars={
+            "amplitude": (("space", "time"), np.ones((5, 10))),
+            "pnt_nad": (("space"), np.array([0.01, 0.01, 0.9, 0.9, 0.01])),
+            "pnt_nmad": (("space"), np.array([0.01, 0.01, 0.9, 0.9, 0.01])),
+        },
+        coords={
+            "space": np.array([1, 2, 5, 6, 7]),  # non monotonic space coords
+            "time": np.arange(10),
+            "azimuth": (("space"), np.arange(5)),
+            "range": (("space"), np.arange(5)),
+        },
+    )
+    res_nad = network_stm_seletcion(
+        stm, min_dist=3, include_index=[1], sortby_var="pnt_nad", azimuth_spacing=1, range_spacing=1
+    )
+    res_nmad = network_stm_seletcion(
+        stm, min_dist=3, include_index=[1], sortby_var="pnt_nmad", azimuth_spacing=1, range_spacing=1
+    )
+
+    # The two pixels with the lowest NAD should be selected
+    assert np.all(np.isclose(res_nad["pnt_nad"].values, 0.01, rtol=1e-09, atol=1e-09))
+    assert np.all(np.isclose(res_nmad["pnt_nmad"].values, 0.01, rtol=1e-09, atol=1e-09))
+    assert np.all(res_nad["space"].values == np.array([2, 7]))
+    assert np.all(res_nmad["space"].values == np.array([2, 7]))
+
+
+def test_network_stm_seletcion_wrong_csr():
+    stm = xr.Dataset(
+        data_vars={
+            "amplitude": (("space", "time"), np.ones((100, 10))),
+            "pnt_nad": (("space"), np.linspace(0, 1, 100)),
+        },
+        coords={"azimuth": (("space"), np.arange(100)), "range": (("space"), np.arange(100)), "time": np.arange(10)},
+    )
+    # catch not implemented method
+    with pytest.raises(NotImplementedError):
+        network_stm_seletcion(
+            stm, min_dist=20, sortby_var="pnt_nad", azimuth_spacing=10, range_spacing=10, crs="not_implemented"
+        )
+
+
 def test_nad_block_zero_dispersion():
     """NAD for a constant array should be zero."""
     slcs = xr.DataArray(