refactor from matplotlib to plotly in EDA scripts

- eda_mp_trj.py replace matplotlib with Plotly for element count heatmaps and ratios
- plot_structure_perturbation.py utilize Plotly for histogram and structure visualizations
- wbm_umap_projection.py use Plotly for UMAP scatter plots, improving interactivity and aesthetics

data/mp/eda_mp_trj.py

@@ -10,7 +10,6 @@
 import pandas as pd
 import plotly.express as px
 import pymatviz as pmv
-from matplotlib.colors import SymLogNorm
 from pymatgen.core import Composition
 from pymatgen.core.tensors import Tensor
 from pymatviz.enums import Key
@@ -243,42 +242,39 @@ def info_dict_to_id(info: dict[str, int | str]) -> str:
 
 excl_elems = "He Ne Ar Kr Xe".split() if (excl_noble := False) else ()
 
-ax_ptable = pmv.ptable_heatmap(  # matplotlib version looks better for SI
+fig = pmv.ptable_heatmap_plotly(
     trj_elem_counts,
-    # zero_color="#efefef",
-    log=(log := SymLogNorm(linthresh=10_000)),
     exclude_elements=excl_elems,  # drop noble gases
-    # cbar_range=None if excl_noble else (10_000, None),
-    show_values=(show_vals := True),
-    # label_font_size=17 if show_vals else 25,
-    # value_font_size=14,
-    cbar_title="MPtrj Element Counts",
+    log=(log := True),
+    colorbar=dict(title="MPtrj Element Counts"),
 )
 
 img_name = f"mp-trj-element-counts-by-{count_mode}"
-if log:
-    img_name += "-symlog" if isinstance(log, SymLogNorm) else "-log"
 if excl_noble:
     img_name += "-excl-noble"
-pmv.save_fig(ax_ptable, f"{PDF_FIGS}/{img_name}.pdf")
+if log:
+    img_name += "-log"
+pmv.save_fig(fig, f"{PDF_FIGS}/{img_name}.pdf")
+fig.show()
 
 
 # %%
 normalized = True
-ax_ptable = pmv.ptable_heatmap_ratio(
-    trj_elem_counts / (len(df_mp_trj) if normalized else 1),
-    mp_occu_counts / (len(df_mp) if normalized else 1),
-    zero_color="#efefef",
-    fmt=".2f",
-    not_in_denominator=None,
-    not_in_numerator=None,
-    not_in_either=None,
+fig = pmv.ptable_heatmap_plotly(
+    {
+        elem: (trj_count / 1_580_395) / (mp_count / len(df_mp))
+        for elem, trj_count in trj_elem_counts.items()
+        if elem in mp_occu_counts
+        for mp_count in [mp_occu_counts[elem]]  # clever way to get mp_count in scope
+    },
+    colorbar=dict(title="MPtrj/MP Element Count Ratio"),
 )
 
 img_name = "mp-trj-mp-ratio-element-counts-by-occurrence"
 if normalized:
     img_name += "-normalized"
-pmv.save_fig(ax_ptable, f"{PDF_FIGS}/{img_name}.pdf")
+pmv.save_fig(fig, f"{PDF_FIGS}/{img_name}.pdf")
+fig.show()
 
 
 # %% plot formation energy per atom distribution

models/cgcnn/plot_structure_perturbation.py

@@ -1,5 +1,4 @@
 # %%
-import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
 import pymatviz as pmv
@@ -11,28 +10,30 @@
 __date__ = "2022-12-02"
 
 rng = np.random.default_rng(0)
+pmv.set_plotly_template("pymatviz_dark")
 
 
 # %%
-ax = pd.Series(rng.weibull(1.5, 100_000)).hist(bins=100)
+fig = pd.Series(rng.weibull(1.5, 100_000)).hist(bins=100, backend="plotly")
 title = "Distribution of perturbation magnitudes"
-ax.set(xlabel="magnitude of perturbation", ylabel="count", title=title)
+fig.layout.update(xaxis_title="Perturbation Magnitude", title=title)
+fig.show()
 
 
 # %%
 struct = Structure(
     lattice=Lattice.cubic(5),
-    species=("Fe", "O"),
+    species=("Fe", "Fe"),
     coords=((0, 0, 0), (0.5, 0.5, 0.5)),
 )
 
-ax = pmv.structure_2d(struct)
-ax.set(title=f"Original structure: {struct.formula}")
-ax.set_aspect("equal")
+fig = pmv.structure_2d_plotly(struct)
+fig.layout.update(title=f"Original structure: {struct.formula}")
+fig.show()
 
 
 # %%
-fig, axs = plt.subplots(3, 4, figsize=(12, 10))
-for idx, ax in enumerate(axs.flat, start=1):
-    pmv.structure_2d(perturb_structure(struct), ax=ax)
-    ax.set(title=f"perturbation {idx}")
+pmv.structure_2d_plotly(
+    [perturb_structure(struct) for _ in range(12)],
+    subplot_title=lambda _struct, idx: f"perturbation {idx}",
+)

scripts/wbm_umap_projection.py

@@ -12,7 +12,6 @@
 # %%
 import os
 
-import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
 import pymatviz as pmv
@@ -189,18 +188,20 @@ def features_to_drop(df_in: pd.DataFrame, threshold: float = 0.95) -> list[str]:
 umap_cols = list(df_umap)
 if umap_cols != ["UMAP 1", "UMAP 2"]:
     raise ValueError(f"Unexpected {umap_cols=}")
-min_step, max_step = df_umap.index.min(), df_umap.index.max()
-ax = df_umap.plot.scatter(
-    *umap_cols, c=df_umap.index, cmap="Spectral", s=5, figsize=(6, 4), colorbar=False
+
+fig = df_umap.plot.scatter(
+    x="UMAP 1",
+    y="UMAP 2",
+    color=df_umap.index.astype(str).str.replace("0", "MP original structures"),
+    backend="plotly",
+    template="pymatviz_white",
 )
-cbar = ax.figure.colorbar(
-    ax.collections[0],
-    boundaries=np.arange(min_step, max_step + 2) - 0.5,
-    ticks=range(min_step, max_step + 1),
+fig.layout.legend.update(
+    title="WBM step:", orientation="h", y=1.1, itemsizing="constant"
 )
-cbar.ax.set_title("WBM step (0 = MP)", rotation=90, y=0.5, x=3, va="center")
+fig.show()
 
 
 # %%
-plt.tight_layout()
-pmv.save_fig(ax, f"{PDF_FIGS}/wbm-final-struct-matminer-features-2d-umap.png", dpi=300)
+img_path = f"{PDF_FIGS}/wbm-final-struct-matminer-features-2d-umap.png"
+pmv.save_fig(fig, img_path, scale=3)