Adopt NumPy and Pandas linting rules

.github/workflows/full_test.yml

@@ -18,7 +18,7 @@ jobs:
 
       - uses: astral-sh/ruff-action@v2
         with:
-          version: 0.6.2
+          version: 0.14.5
           src: src
 
       - name: Set up uv

pyproject.toml

@@ -45,7 +45,7 @@ classifiers = [
 dev = [
     "pytest",
     "plotly >= 4.5",
-    "ruff==0.6.2",
+    "ruff==0.14.5",
     "quarto-cli==1.5.57",
     "quartodoc==0.11.1",
     "netCDF4",
@@ -70,8 +70,8 @@ notebooks = ["nbformat", "nbconvert", "jupyter", "plotly", "shapely", "seaborn"]
 
 
 [tool.ruff.lint]
-ignore = ["E501"]
-select = ["E4", "E7", "E9", "F", "D200", "D205"]
+ignore = ["E501", "PD011"]
+select = ["E4", "E7", "E9", "F", "D200", "D205", "NPY", "PD"]
 
 [tool.mypy]
 python_version = "3.10"

src/modelskill/comparison/_collection_plotter.py

@@ -72,6 +72,7 @@ def scatter(
         xlabel: Optional[str] = None,
         ylabel: Optional[str] = None,
         skill_table: Optional[Union[str, List[str], Mapping[str, str], bool]] = None,
+        random_state: Optional[int] = None,
         ax: Optional[Axes] = None,
         **kwargs,
     ) -> Axes | list[Axes]:
@@ -138,6 +139,8 @@ def scatter(
             This kword adds a box at the right of the scatter plot.
             mapping can be used to rename the metrics in the table.
             by default False
+        random_state : int, optional
+            random seed for point sampling, by default None
         ax : matplotlib axes, optional
             axes to plot on, by default None
         **kwargs
@@ -176,6 +179,7 @@ def scatter(
                 xlabel=xlabel,
                 ylabel=ylabel,
                 skill_table=skill_table,
+                random_state=random_state,
                 ax=ax,
                 **kwargs,
             )
@@ -201,6 +205,7 @@ def _scatter_one_model(
         xlabel: Optional[str],
         ylabel: Optional[str],
         skill_table: Optional[Union[str, List[str], Mapping[str, str], bool]],
+        random_state: Optional[int] = None,
         ax,
         **kwargs,
     ):
@@ -268,6 +273,7 @@ def _scatter_one_model(
             skill_scores=skill_scores,
             skill_score_unit=skill_score_unit,
             ax=ax,
+            random_state=random_state,
             **kwargs,
         )
 

src/modelskill/comparison/_comparer_plotter.py

@@ -476,6 +476,7 @@ def scatter(
         xlabel: Optional[str] = None,
         ylabel: Optional[str] = None,
         skill_table: Optional[Union[str, List[str], Mapping[str, str], bool]] = None,
+        random_state: Optional[int] = None,
         ax: Optional[matplotlib.axes.Axes] = None,
         **kwargs,
     ) -> matplotlib.axes.Axes | list[matplotlib.axes.Axes]:
@@ -542,6 +543,8 @@ def scatter(
             modelskill.options.metrics.list. This kword adds a box at the
             right of the scatter plot, by default False
             mapping can be used to rename the metrics in the table.
+        random_state : int, optional
+            random seed for point sampling, by default None
         ax : matplotlib.axes.Axes, optional
             axes to plot on, by default None
         **kwargs
@@ -581,6 +584,7 @@ def scatter(
                 ylabel=ylabel,
                 skill_table=skill_table,
                 ax=ax,
+                random_state=random_state,
                 **kwargs,
             )
             axes.append(ax_mod)
@@ -746,7 +750,14 @@ def taylor(
         df = df.rename(columns={"_std_obs": "obs_std", "_std_mod": "std"})
 
         pts = [
-            TaylorPoint(name=r.model, obs_std=r.obs_std, std=r.std, cc=r.cc, marker=marker, marker_size=marker_size)
+            TaylorPoint(
+                name=r.model,
+                obs_std=r.obs_std,
+                std=r.std,
+                cc=r.cc,
+                marker=marker,
+                marker_size=marker_size,
+            )
             for r in df.itertuples()
         ]
 

src/modelskill/comparison/_comparison.py

@@ -56,7 +56,7 @@ def _parse_dataset(data: xr.Dataset) -> xr.Dataset:
     assert "Observation" in data.data_vars
 
     # no missing values allowed in Observation
-    if data["Observation"].isnull().any():
+    if data["Observation"].isnull().any():  # noqa: PD003
         raise ValueError("Observation data must not contain missing values.")
 
     # coordinates

src/modelskill/metrics.py

@@ -632,7 +632,7 @@ def peak_ratio(
         }
     )
     df_filter["Maximum"] = df_filter.max(axis=1)
-    df_filter.sort_values(by="Maximum", ascending=False, inplace=True)
+    df_filter = df_filter.sort_values(by="Maximum", ascending=False)
     # Finally we do the selection of the N- largest peaks from either model or measured
     df_filter = df_filter.iloc[0:top_n_peaks, :]
     # Rename to avoid further refactoring

src/modelskill/model/dfsu.py

@@ -219,7 +219,7 @@ def _extract_point(
 
         # TODO not sure why we rename here
         assert self.name is not None
-        ds_model.rename({ds_model.items[0].name: self.name}, inplace=True)
+        ds_model = ds_model.rename({ds_model.items[0].name: self.name})
 
         return PointModelResult(
             data=ds_model,
@@ -268,7 +268,7 @@ def _extract_track(
 
         if isinstance(self.data, mikeio.DataArray):
             ds_model = self.data.extract_track(track=track, method=method)
-            ds_model.rename({self.data.name: self.name}, inplace=True)
+            ds_model = ds_model.rename({self.data.name: self.name})
             aux_items = None
         else:
             if isinstance(self.data, mikeio.dfsu.Dfsu2DH):
@@ -279,7 +279,7 @@ def _extract_track(
                 ds_model = self.data[self.sel_items.all].extract_track(
                     track=track, method=method
                 )
-            ds_model.rename({self.sel_items.values: self.name}, inplace=True)
+            ds_model = ds_model.rename({self.sel_items.values: self.name})
             aux_items = self.sel_items.aux
 
         item_names = [i.name for i in ds_model.items]

src/modelskill/plotting/_misc.py

@@ -53,25 +53,11 @@ def _xyticks(n_sectors=8, lim=None):
 
 
 def sample_points(
-    x: np.ndarray, y: np.ndarray, show_points: bool | int | float | None = None
+    x: np.ndarray,
+    y: np.ndarray,
+    show_points: bool | int | float | None = None,
+    random_state: int | None = None,
 ) -> Tuple[np.ndarray, np.ndarray]:
-    """Sample points to be plotted
-
-    Parameters
-    ----------
-    x: np.ndarray, 1d
-    y: np.ndarray, 1d
-    include: bool, int or float, optional
-        default is subset the data to 50k points
-
-    Returns
-    -------
-    np.ndarray, np.ndarray
-        x and y arrays with sampled points
-    """
-
-    assert len(x) == len(y), "x and y must have same length"
-
     if show_points is True:
         return x, y
 
@@ -104,8 +90,8 @@ def sample_points(
             show_points = len(x)
         n_samples = show_points
 
-    np.random.seed(20)  # TODO should this be a parameter?
-    ran_index = np.random.choice(range(len(x)), n_samples, replace=False)
+    rng = np.random.default_rng(seed=random_state)
+    ran_index = rng.choice(range(len(x)), n_samples, replace=False)
     x_sample = x[ran_index]
     y_sample = y[ran_index]
 

src/modelskill/plotting/_scatter.py

@@ -44,6 +44,7 @@ def scatter(
     skill_scores: Mapping[str, float] | None = None,
     skill_score_unit: Optional[str] = "",
     ax: Optional[Axes] = None,
+    random_state: Optional[int] = None,
     **kwargs,
 ) -> Axes:
     """Scatter plot tailored for model skill comparison.
@@ -120,6 +121,8 @@ def scatter(
         unit for skill_scores, by default None
     ax : matplotlib.axes.Axes, optional
         axes to plot on, by default None
+    random_state : int, optional
+        random seed for point sampling, by default None
     **kwargs
 
     Returns
@@ -155,7 +158,7 @@ def scatter(
     if norm is None:
         norm = colors.PowerNorm(vmin=1, gamma=0.5)
 
-    x_sample, y_sample = sample_points(x, y, show_points)
+    x_sample, y_sample = sample_points(x, y, show_points, random_state=random_state)
     show_points = len(x_sample) > 0
     xq, yq = quantiles_xy(x, y, quantiles)
 

src/modelskill/skill.py

@@ -51,7 +51,7 @@ def _name_to_title_in_kwargs(self, kwargs: Any) -> None:
     def _get_plot_df(self, level: int | str = 0) -> pd.DataFrame:
         ser = self.skillarray._ser
         if isinstance(ser.index, pd.MultiIndex):
-            df = ser.unstack(level=level)
+            df = ser.unstack(level=level)  # noqa: PD010
         else:
             df = ser.to_frame()
         return df
@@ -226,7 +226,7 @@ def grid(
             # TODO raise error?
             return None
             # df = self.df[field]    TODO: at_least_2d...
-        df = ser.unstack()
+        df = ser.unstack()  # noqa: PD010
 
         vmin = None
         vmax = None

tests/plot/test_plot.py

@@ -122,9 +122,9 @@ def test_format_skill_table():
 
 @pytest.fixture
 def x_y():
-    np.random.seed(42)
-    x = np.random.rand(100000)
-    y = np.random.rand(100000)
+    rng = np.random.default_rng(seed=42)
+    x = rng.random(100000)
+    y = rng.random(100000)
     return x, y
 
 

tests/test_comparercollection.py

@@ -455,7 +455,7 @@ def test_save_and_load_preserves_raw_model_data(cc, tmp_path):
 
 
 def test_plot_scatter(cc):
-    ax = cc.plot.scatter(skill_table=True)
+    ax = cc.plot.scatter(skill_table=True, random_state=20)
     assert ax is not None