More sensible time axis and tick labels for 2D plots

[rcomer]

🚀 Pull Request

Description

1. Removed the date2num conversions from datetimes to floats in plot.py. I'm not clear why these were there, but I assume were an ancient workaround for something. They had been there since the initial commit to github. Without them, the time axis ticks get labelled with datetime information instead of floats (see the updated images in the test data PR).
2. Pulled in the axis label change from removed units for time coords due to not matching autoticks #5484. As discussed at Quickplot Automatic Time Axis Label Completely Incorrect #5426, we only want to label these axes as "time".
3. Reverted the quickplot change from Remove quickplot unit label for time reference coordinates. Closes #615. #616, as [2] is a more general version of this.

Closes #5484 and closes #5426

---

Consult Iris pull request check list

[rcomer]

Test data PR: SciTools/iris-test-data#87

[rcomer]

Note that users who specified a DateFormatter and DateLocator should not see any difference in their tick labels. E.g. this plot is the same with main and with my branch:

import matplotlib.pyplot as plt
import matplotlib.dates as mdates

import iris
import iris.plot as iplt
import iris.tests.stock

cube = iris.tests.stock.realistic_4d()
cube = cube[:, 0, 0, :]

fig, _ = plt.subplots(layout="constrained")

iplt.contourf(cube)
ax = plt.gca()
ax.yaxis.set_major_locator(mdates.MinuteLocator(interval=10))
ax.yaxis.set_major_formatter(mdates.DateFormatter("%M minutes after %H"))

plt.savefig("test.png")

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (cec0e09) 89.44% compared to head (62a92b7) 89.44%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #5561   +/-   ##
=======================================
  Coverage   89.44%   89.44%           
=======================================
  Files          89       89           
  Lines       22597    22583   -14     
  Branches     5390     5386    -4     
=======================================
- Hits        20211    20199   -12     
+ Misses       1639     1638    -1     
+ Partials      747      746    -1     

Files	Coverage Δ
lib/iris/plot.py	86.35% <ø> (-0.10%)	⬇️
lib/iris/quickplot.py	90.37% <100.00%> (+0.78%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.