month_grid=True fails when a month has no valid data; specify a separate colour for missing values like np.nan

[FeiYao-Edinburgh]

Hi there,

import numpy as np
import matplotlib.pyplot as plt
import july
from july.utils import date_range

dates = date_range("2020-01-01", "2020-12-31")
data = np.random.rand(len(dates))

data[0:50] = np.nan
data[300:] = np.nan

# we'll at least need one data for each month to enable month_grid...
july.heatmap(
    dates,
    data,
    title='Github Activity',
    cmap=cmaps.WhiteBlueGreenYellowRed,
    month_grid=True,)

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[66], line 13
     10 data[300:] = np.nan
     12 # we'll at least need one data for each month to enable month_grid...
---> 13 july.heatmap(
     14     dates,
     15     data,
     16     title='Github Activity',
     17     cmap=cmaps.WhiteBlueGreenYellowRed,
     18     month_grid=True,)

File ~/scratch/miniconda3/envs/gcadj/lib/python3.10/site-packages/july/plot.py:75, in heatmap(dates, data, horizontal, cmap, value_label, date_label, weekday_label, month_label, year_label, month_grid, month_grid_color, colorbar, frame_on, value_format, title, cmin, cmax, cbar_label_format, ax, **kwargs)
     73 dates_clean, data_clean = preprocess_inputs(dates, data)
     74 cal = date_grid(dates_clean, data_clean, horizontal)
---> 75 ax = cal_heatmap(
     76     cal=cal,
     77     dates=dates_clean,
     78     horizontal=horizontal,
     79     cmap=cmap,
     80     value_label=value_label,
     81     date_label=date_label,
     82     weekday_label=weekday_label,
     83     month_label=month_label,
     84     year_label=year_label,
     85     month_grid=month_grid,
     86     month_grid_color=month_grid_color,
     87     colorbar=colorbar,
     88     frame_on=frame_on,
     89     value_format=value_format,
     90     title=title,
     91     cmin=cmin,
     92     cmax=cmax,
     93     cbar_label_format=cbar_label_format,
     94     ax=ax,
     95 )
     97 return ax

File ~/scratch/miniconda3/envs/gcadj/lib/python3.10/site-packages/july/helpers.py:96, in cal_heatmap(cal, dates, horizontal, cmap, value_label, date_label, weekday_label, month_label, year_label, month_grid, month_grid_color, colorbar, frame_on, value_format, title, cmin, cmax, cbar_label_format, ax)
     94     add_year_label(ax, dates, horizontal)
     95 if month_grid:
---> 96     add_month_grid(ax, dates, cal, horizontal, month_grid_color)
     97 if colorbar:
     98     add_colorbar(pc, fig, ax, bbox, cbar_label_format)

File ~/scratch/miniconda3/envs/gcadj/lib/python3.10/site-packages/july/helpers.py:280, in add_month_grid(ax, dates, month_grid, horizontal, color)
    278 months = set([d.month for d in dates])
    279 for month in months:
--> 280     coords = get_month_outline(
    281         dates, month_grid, horizontal=horizontal, month=month
    282     )
    283     ax.plot(coords[:, 0], coords[:, 1], color=color, linewidth=1)
    285 # Pad axes so plotted line appears uniform also along edges.

File ~/scratch/miniconda3/envs/gcadj/lib/python3.10/site-packages/july/helpers.py:243, in get_month_outline(dates, month_grid, horizontal, month)
    240                 coords_list.append((x, y))
    242 sorted_coords = np.array(coords_list)
--> 243 min_y = sorted_coords[:, 1].min()
    244 max_y = sorted_coords[:, 1].max()
    245 upper_left = sorted_coords[0]

IndexError: too many indices for array: array is 1-dimensional, but 2 were indexed

[FeiYao-Edinburgh]

As related, I believe it would be fantastic if we could specify a separate colour for missing values like np.nan, instead of filling them with zero, as currently done in utils.preprocess_inputs(). Could the community assist with this?