Use draw_idle instead of draw in remake_ticks

[jzuhone]

Description

xija_gui_fit is prone to crashing, often without error messages, or only cryptic error messages such as this one:

(ska) acis-110: xija_gui_fit ./dpa_spec_PROD.json --days 365 --stop 2022:105:00:00:00.00 &
Fetching msid: 1dpamzt over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
info:numexpr.utils:NumExpr defaulting to 4 threads.
Fetching msid: sim_z over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
Fetching msid: pitch over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
Fetching msid: roll over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
Fetching msid: aoeclips over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
Getting kadi commanded states over 2021:105:00:04:30.816 to 2022:104:23:51:42.816
Fetching msid: dp_dpa_power over 2021:104:23:37:10.816 to 2022:105:00:19:02.816
Adding plot  1dpamzt data__time
Adding plot  1dpamzt resid__time
Traceback (most recent call last):
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backend_bases.py", line 3031, in _wait_cursor_for_draw_cm
    yield
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py", line 439, in draw
    super().draw()
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backends/backend_qt.py", line 424, in draw
    self.update()
RuntimeError: wrapped C/C++ object of type MplCanvas has been deleted

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/cbook/__init__.py", line 287, in process
    func(*args, **kwargs)
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/Ska/Matplotlib/core.py", line 97, in remake_ticks
    ax.figure.canvas.draw()
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py", line 439, in draw
    super().draw()
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/contextlib.py", line 131, in __exit__
    self.gen.throw(type, value, traceback)
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backend_bases.py", line 3033, in _wait_cursor_for_draw_cm
    self.canvas.set_cursor(self._lastCursor)
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backends/backend_qt.py", line 259, in set_cursor
    self.setCursor(_api.check_getitem(cursord, cursor=cursor))
RuntimeError: wrapped C/C++ object of type MplCanvas has been deleted
Traceback (most recent call last):
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/cbook/__init__.py", line 287, in process
    func(*args, **kwargs)
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backend_bases.py", line 3077, in _zoom_pan_handler
    self.release_zoom(event)
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backend_bases.py", line 3232, in release_zoom
    self.push_current()
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backend_bases.py", line 3243, in push_current
    self.set_history_buttons()
  File "/data/acis/miniconda3/envs/ska/lib/python3.8/site-packages/matplotlib/backends/backend_qt.py", line 797, in set_history_buttons
    self._actions['back'].setEnabled(can_backward)
RuntimeError: wrapped C/C++ object of type QAction has been deleted

This can be resolved if the call to ax.figure.canvas.draw is changed to ax.figure.canvas.draw_idle in remake_ticks, which is what this PR does.

From the Matplotlib documentation:

"In almost all cases, we recommend using backend_bases.FigureCanvasBase.draw_idle over backend_bases.FigureCanvasBase.draw. draw forces a rendering of the figure whereas draw_idle schedules a rendering the next time the GUI window is going to re-paint the screen."

Testing

• Changing this line fixes crashes in xija_gui_fit.
• I checked that the iplot functionality in cheta works correctly with this change.

[taldcroft]

@jzuhone - thanks, looks reasonable. This probably also needs some testing of the original standalone functionality when running interactively in an IPython terminal session.

[jzuhone]

@taldcroft good call--I did check and the interactive plotting does work correctly.

[jzuhone]

any reason not to merge this PR?