Fix test suite (#1314)

datashader/datatypes.py

@@ -649,6 +649,10 @@ def __array__(self, dtype=None):
         dtype = np.dtype(object) if dtype is None else np.dtype(dtype)
         return np.asarray(self.tolist(), dtype=dtype)
 
+    def duplicated(self, *args, **kwargs):
+        msg = "duplicated is not implemented for RaggedArray"
+        raise NotImplementedError(msg)
+
 
 @jit(nopython=True, nogil=True)
 def _eq_ragged_ragged(start_indices1,

datashader/tests/test_datatypes.py

@@ -835,6 +835,12 @@ def test_sort_values_frame(self):
     def test_where_series(self):
         pass
 
+    @pytest.mark.xfail(reason="not currently supported")
+    def test_duplicated(self, data):
+        # Added in Pandas 2.2
+        # https://github.com/pandas-dev/pandas/pull/55255
+        super().test_duplicated(data)
+
 class TestRaggedPrinting(eb.BasePrintingTests):
     @pytest.mark.skip(reason="Can't autoconvert ragged array to numpy array")
     def test_dataframe_repr(self):
@@ -873,6 +879,11 @@ def test_fillna_no_op_returns_copy(self):
     def test_fillna_series_method(self):
         pass
 
+    @pytest.mark.skip(reason="Can't fill with nested sequences")
+    def test_ffill_limit_area(self):
+        # Added in Pandas 2.2
+        pass
+
 
 class TestRaggedReshaping(eb.BaseReshapingTests):
     @pytest.mark.skip(reason="__setitem__ not supported")
@@ -886,3 +897,15 @@ def test_transpose(self):
     @pytest.mark.skip(reason="transpose with numpy array elements seems not supported")
     def test_transpose_frame(self):
         pass
+
+    @pytest.mark.skipif(
+        Version(pd.__version__) == Version("2.2.0"), reason="Regression in Pandas 2.2"
+    )
+    def test_merge_on_extension_array(self, data):
+        super().test_merge_on_extension_array(data)
+
+    @pytest.mark.skipif(
+        Version(pd.__version__) == Version("2.2.0"), reason="Regression in Pandas 2.2"
+    )
+    def test_merge_on_extension_array_duplicates(self, data):
+        super().test_merge_on_extension_array_duplicates(data)

datashader/tests/test_geopandas.py

@@ -200,7 +200,7 @@ def test_lines_spatialpandas(geom_type, explode, use_boundary, npartitions):
 def test_points_geopandas(geom_type):
     df = geopandas.read_file(geodatasets.get_path("nybb"))
 
-    df["geometry"] = df["geometry"].sample_points(100, seed=93814)  # multipoint
+    df["geometry"] = df["geometry"].sample_points(100, rng=93814)  # multipoint
     if geom_type == "point":
         df = df.explode(index_parts=False)  # Multipoint -> point.
     unique_geom_type = df["geometry"].geom_type.unique()
@@ -218,7 +218,7 @@ def test_points_geopandas(geom_type):
 def test_points_dask_geopandas(geom_type, npartitions):
     df = geopandas.read_file(geodatasets.get_path("nybb"))
 
-    df["geometry"] = df["geometry"].sample_points(100, seed=93814)  # multipoint
+    df["geometry"] = df["geometry"].sample_points(100, rng=93814)  # multipoint
     if geom_type == "point":
         df = df.explode(index_parts=False)  # Multipoint -> point.
     unique_geom_type = df["geometry"].geom_type.unique()
@@ -240,7 +240,7 @@ def test_points_dask_geopandas(geom_type, npartitions):
 def test_points_spatialpandas(geom_type, npartitions):
     df = geopandas.read_file(geodatasets.get_path("nybb"))
 
-    df["geometry"] = df["geometry"].sample_points(100, seed=93814)  # multipoint
+    df["geometry"] = df["geometry"].sample_points(100, rng=93814)  # multipoint
     if geom_type == "point":
         df = df.explode(index_parts=False)  # Multipoint -> point.
     unique_geom_type = df["geometry"].geom_type.unique()

datashader/tests/test_pandas.py

@@ -2734,7 +2734,8 @@ def test_line_antialias():
     numerator = np.nan_to_num(line_antialias_sol_0_intersect) + np.nan_to_num(line_antialias_sol_1)
     denom = np.nan_to_num(line_antialias_sol_count_ignore_aa_0) + \
         np.nan_to_num(line_antialias_sol_count_ignore_aa_1)
-    sol = 3*numerator / denom
+    with np.errstate(invalid='ignore'):  # Dividing by zero is expected
+        sol = 3*numerator / denom
     assert_eq_ndarray(agg.data, sol, close=True)
 
     agg = cvs.line(agg=ds._min_row_index(), **kwargs)

datashader/tests/test_pipeline.py

@@ -9,7 +9,7 @@
 df = pd.DataFrame({'x': np.array(([0.] * 10 + [1] * 10)),
                    'y': np.array(([0.] * 5 + [1] * 5 + [0] * 5 + [1] * 5)),
                    'f64': np.arange(20, dtype='f8')})
-df.f64.iloc[2] = np.nan
+df.loc['f64', 2] = np.nan
 
 cvs = ds.Canvas(plot_width=2, plot_height=2, x_range=(0, 1), y_range=(0, 1))
 cvs10 = ds.Canvas(plot_width=10, plot_height=10, x_range=(0, 1), y_range=(0, 1))

pyproject.toml

@@ -18,3 +18,16 @@ target-version = "py39"
 
 [tool.ruff.per-file-ignores]
 "test_mpl_ext.py" = ["E402"]  # Module level import not at top of file
+
+[tool.pytest.ini_options]
+addopts = ["--pyargs", "--doctest-modules", "--doctest-ignore-import-errors", "--strict-config", "--strict-markers", "--color=yes"]
+norecursedirs = 'doc .git dist build _build .ipynb_checkpoints'
+minversion = "7"
+xfail_strict = true
+log_cli_level = "INFO"
+# skipping any notebooks that require extra deps
+nbsmoke_skip_run = ".*tiling.ipynb$\n.*streaming-aggregation.ipynb$\n.*8_Geography.ipynb$"
+filterwarnings = [
+    "ignore:Passing a (SingleBlockManager|BlockManager) to (Series|GeoSeries|DataFrame|GeoDataFrame) is deprecated:DeprecationWarning",  # https://github.com/holoviz/spatialpandas/issues/137
+    "ignore:Accessing the underlying geometries through the `.data`:DeprecationWarning:dask_geopandas.core",  # https://github.com/geopandas/dask-geopandas/issues/264
+]

setup.py

@@ -44,14 +44,13 @@
     'tests': geopandas + [
         'codecov',
         'geodatasets',
-        'fastparquet',  # optional dependency
         'flake8',
         'nbconvert',
         'nbformat',
         'nbsmoke[verify] >0.5',
         'netcdf4',
         'pyarrow',
-        'pytest',
+        'pytest <8',  # Fails lint with IPynbFile is deprecated
         'pytest-benchmark',
         'pytest-cov',
         'rasterio',

tox.ini

@@ -71,14 +71,6 @@ deps = unit: {[_unit]deps}
        examples_extra: {[_examples_extra]deps}
        all: {[_all]deps}
 
-[pytest]
-addopts = -v --pyargs --doctest-modules --doctest-ignore-import-errors
-norecursedirs = doc .git dist build _build .ipynb_checkpoints
-# skipping any notebooks that require extra deps
-nbsmoke_skip_run = .*tiling.ipynb$
-                   .*streaming-aggregation.ipynb$
-                   .*8_Geography.ipynb$
-
 [flake8]
 include = *.py
 # run_tests.py is generated by conda build, which appears to have a