Better missing value handling strategy for ptable heatmap

[DanielYang59]

• Currently missing value would either be replaced by zero or mean, perhaps it's better to add an option to keep missing value as is (shown as -).
• User cannot control the missing value strategy from ptable_heatmap, which default to mean, leading to erroneous values and statistics.
  

  pymatviz/pymatviz/ptable/_process_data.py

  Line 82 in 25f90bd

  missing_strategy: Literal["zero", "mean"] = "mean",

from matminer.datasets import load_dataset

import pymatviz as pmv
from pymatviz.enums import Key


df_expt_gap = load_dataset("matbench_expt_gap")
counted_data = pmv.count_elements(df_expt_gap[Key.composition], fill_value=None)

ax = pmv.ptable_heatmap(
    counted_data,
    value_show_mode="percent",
)
pmv.io.save_and_compress_svg(ax, "./ptable-heatmap.svg")

Should be: