Explore automatic generation of videos with graphical/mouse interactions

[melissawm]

🧰 Task

In order to keep the documentation up to date, the videos (webm) we have on our docs pages should be automated whenever possible. This doesn't mean they would be rebuilt at every docs build, but maybe once every release or when we have important gui changes.

The initial proposal is to have a script that is associated with each video and describes the sequence of actions that needs to be executed to capture that output. For example: for this video on the points layer guide, a script would be:

- Click add points icon
- Add three new points
- Click select points icon
- Select two points individually
- Drag mouse to select group of points
- Change face color 
- Change edge color
- Select group of points with different colors
- Use slider to increase point size
- Select another group of points
- Use slider to decrease point size
- With no point selected, click symbol dropdown and select cross
- Use slider to decrease opacity
- Use slider to increase opacity
- Select group of points and click the "delete selected symbol" icon
- Click the add points icon
- Use the face color dropdown to select a different color
- Use the slider to increase point size and add new points

In the future, we can maybe explore tools such as https://pyautogui.readthedocs.io/en/latest/

[psobolewskiPhD]

Oh wow, that's a great idea!

[melissawm]

Another thought: can we add a "review videos and images for outdated content" to the release guide?

The release manager doesn't need to be responsible for this but maybe they can ping the documentation leads. This would make it a soft regular requirement.

[kephale]

I just want to cross reference this issue with this open napari issue about video recording from @GenevieveBuckley: napari/napari-animation#225

[psobolewskiPhD]

I will add to the cross reference:
docs issue: #289
We may want to consider napari-animation in this context too https://github.com/napari/napari-animation