Clean up var names in unit tests, avoid MP API access in GitHub workflow

.github/workflows/test.yml

@@ -18,7 +18,6 @@ jobs:
     with:
       os: ${{ matrix.os }}
       python-version: "3.10"
-    secrets: inherit
 
   find-scripts:
     runs-on: ubuntu-latest
@@ -55,5 +54,3 @@ jobs:
 
       - name: Run script
         run: python ${{ matrix.script }}
-        env:
-          MP_API_KEY: ${{ secrets.MP_API_KEY }}

examples/make_assets/structure_viz.py

@@ -1,4 +1,5 @@
 # %%
+import os
 from typing import cast
 
 import matplotlib.pyplot as plt
@@ -8,10 +9,9 @@
 
 import pymatviz as pmv
 from pymatviz.enums import ElemColorScheme, Key
+from pymatviz.utils import TEST_FILES
 
 
-struct: Structure  # for type hinting
-
 df_steels = load_dataset("matbench_steels")
 df_phonons = load_dataset("matbench_phonons")
 
@@ -31,12 +31,26 @@
 
 
 # %% Plot some disordered structures in 2D
-disordered_structs = {
-    mp_id: MPRester().get_structure_by_material_id(mp_id, conventional_unit_cell=True)
-    for mp_id in ["mp-19017", "mp-12712"]
-}
+struct_mp_ids = ("mp-19017", "mp-12712")
+structure_dir = f"{TEST_FILES}/structures"
+
+os.makedirs(structure_dir, exist_ok=True)
+for mp_id in struct_mp_ids:
+    struct_file = f"{structure_dir}/{mp_id}.json.gz"
+    if not os.path.isfile(struct_file):
+        if os.getenv("CI"):
+            raise FileNotFoundError(
+                f"structure for {mp_id} not found, run this script locally to fetch it."
+            )
+
+        struct: Structure = MPRester().get_structure_by_material_id(
+            mp_id, conventional_unit_cell=True
+        )
+        struct.to_file(struct_file)
+
+    else:
+        struct = Structure.from_file(f"{structure_dir}/{mp_id}.json.gz")
 
-for mp_id, struct in disordered_structs.items():
     for site in struct:  # disorder structures in-place
         if "Fe" in site.species:
             site.species = {"Fe": 0.4, "C": 0.4, "Mn": 0.2}

tests/conftest.py

@@ -17,9 +17,9 @@
     from collections.abc import Generator
 
 
-# if platform is windows, set matplotlib backend to "Agg" to fix
-# _tkinter.TclError: Can't find a usable init.tcl in the following directories
-# https://github.com/orgs/community/discussions/26434
+# If platform is Windows, set matplotlib backend to "Agg" to fix:
+# "_tkinter.TclError: Can't find a usable init.tcl in the following directories"
+# See: https://github.com/orgs/community/discussions/26434
 if platform.system() == "Windows":
     import matplotlib as mpl
 
@@ -52,10 +52,10 @@ def df_or_arrays(request: pytest.FixtureRequest) -> DfOrArrays:
 
 @pytest.fixture(autouse=True)
 def _run_around_tests() -> Generator[None, None, None]:
-    """Ensure matplotlib plots are closed after each test so as not to leak state
-    between tests."""
+    """Ensure matplotlib plots are closed after each test
+    so as not to leak state between tests.
+    """
     # runs before each test
-
     yield
 
     # runs after each test
@@ -121,7 +121,7 @@ def matplotlib_scatter() -> plt.Figure:
 
 @pytest.fixture
 def glass_formulas() -> list[str]:
-    """First 20 materials in the Matbench glass dataset.
+    """First 20 materials in the MatBench glass dataset.
 
     from matminer.datasets import load_dataset
 

tests/test_bar.py

@@ -9,7 +9,6 @@
 
 from pymatviz.bar import spacegroup_bar
 from pymatviz.utils import BACKENDS, MATPLOTLIB, PLOTLY
-from tests.conftest import y_pred, y_true
 
 
 if TYPE_CHECKING:
@@ -20,9 +19,6 @@
     from pymatviz.utils import Backend
 
 
-y_std_mock = y_true - y_pred
-
-
 @pytest.mark.parametrize("backend", BACKENDS)
 @pytest.mark.parametrize(
     ("xticks", "show_counts", "show_empty_bins", "log"),

tests/test_io.py

@@ -9,7 +9,6 @@
 from unittest.mock import patch
 from xml.etree import ElementTree as ET
 
-import pandas as pd
 import plotly.graph_objects as go
 import pytest
 from matplotlib import pyplot as plt

tests/test_scatter.py

@@ -26,8 +26,8 @@
     from tests.conftest import DfOrArrays
 
 
-x_col, y_col, *_ = df_regr
-df_tips = px.data.tips()
+X_COL, Y_COL, *_ = df_regr
+DF_TIPS = px.data.tips()
 
 
 @pytest.mark.parametrize("log_density", [True, False])
@@ -166,7 +166,7 @@ def test_density_scatter_plotly(
 
 
 def test_density_scatter_plotly_hover_template() -> None:
-    fig = density_scatter_plotly(df=df_regr, x=x_col, y=y_col, log_density=True)
+    fig = density_scatter_plotly(df=df_regr, x=X_COL, y=Y_COL, log_density=True)
     hover_template = fig.data[0].hovertemplate
     assert "Point Density" in hover_template
     assert "color" not in hover_template  # Ensure log-count values are not displayed
@@ -175,18 +175,18 @@ def test_density_scatter_plotly_hover_template() -> None:
 @pytest.mark.parametrize("stats", [1, (1,), "foo"])
 def test_density_scatter_plotly_raises_on_bad_stats_type(stats: Any) -> None:
     with pytest.raises(TypeError, match="stats must be bool or dict"):
-        density_scatter_plotly(df=df_regr, x=x_col, y=y_col, stats=stats)
+        density_scatter_plotly(df=df_regr, x=X_COL, y=Y_COL, stats=stats)
 
 
 def test_density_scatter_plotly_empty_dataframe() -> None:
-    empty_df = pd.DataFrame({x_col: [], y_col: []})
+    empty_df = pd.DataFrame({X_COL: [], Y_COL: []})
     with pytest.raises(ValueError, match="input should have multiple elements"):
-        density_scatter_plotly(df=empty_df, x=x_col, y=y_col)
+        density_scatter_plotly(df=empty_df, x=X_COL, y=Y_COL)
 
 
 def test_density_scatter_plotly_facet() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker"
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker"
     )
 
     assert isinstance(fig, go.Figure)
@@ -196,7 +196,7 @@ def test_density_scatter_plotly_facet() -> None:
 
 def test_density_scatter_plotly_facet_log_density() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", log_density=True
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", log_density=True
     )
 
     assert fig.layout.coloraxis.colorbar.ticktext is not None
@@ -205,7 +205,7 @@ def test_density_scatter_plotly_facet_log_density() -> None:
 
 def test_density_scatter_plotly_facet_stats() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", stats=True
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", stats=True
     )
 
     # Check there are at least 2 annotations (could be more due to facet labels)
@@ -217,7 +217,7 @@ def test_density_scatter_plotly_facet_stats() -> None:
 
 def test_density_scatter_plotly_facet_best_fit_line() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", best_fit_line=True
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", best_fit_line=True
     )
 
     # Check there are at least 4 shapes (2 identity lines, 2 best fit lines)
@@ -231,18 +231,18 @@ def test_density_scatter_plotly_facet_best_fit_line() -> None:
 
 def test_density_scatter_plotly_facet_custom_bins() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", n_bins=10
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", n_bins=10
     )
 
     # Check that binning has been applied (number of points should be reduced)
-    smoker_count = df_tips["smoker"].value_counts()
+    smoker_count = DF_TIPS["smoker"].value_counts()
     assert len(fig.data[0].x) < smoker_count["No"]
     assert len(fig.data[1].x) < smoker_count["Yes"]
 
 
 def test_density_scatter_plotly_facet_custom_color() -> None:
     fig = density_scatter_plotly(
-        df=df_tips,
+        df=DF_TIPS,
         x="total_bill",
         y="tip",
         facet_col="smoker",
@@ -260,7 +260,7 @@ def test_density_scatter_plotly_facet_density_methods(
     density: Literal["kde", "empirical"],
 ) -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", density=density
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", density=density
     )
 
     assert isinstance(fig, go.Figure)
@@ -269,30 +269,30 @@ def test_density_scatter_plotly_facet_density_methods(
 
 def test_density_scatter_plotly_facet_size() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", size="size", facet_col="smoker"
+        df=DF_TIPS, x="total_bill", y="tip", size="size", facet_col="smoker"
     )
 
     assert "marker.size" in fig.data[0]
     assert "marker.size" in fig.data[1]
 
 
 def test_density_scatter_plotly_facet_multiple_categories() -> None:
-    fig = density_scatter_plotly(df=df_tips, x="total_bill", y="tip", facet_col="day")
+    fig = density_scatter_plotly(df=DF_TIPS, x="total_bill", y="tip", facet_col="day")
 
-    assert len(fig.data) == df_tips["day"].nunique()
+    assert len(fig.data) == DF_TIPS["day"].nunique()
 
 
 def test_density_scatter_plotly_facet_identity_line() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker", identity_line=True
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker", identity_line=True
     )
 
     assert len(fig.layout.shapes) == 2  # Two identity lines, one for each facet
 
 
 def test_density_scatter_plotly_facet_hover_template() -> None:
     fig = density_scatter_plotly(
-        df=df_tips, x="total_bill", y="tip", facet_col="smoker"
+        df=DF_TIPS, x="total_bill", y="tip", facet_col="smoker"
     )
 
     for trace in fig.data: