Added provisional docstring to infer_reversal_potential

[MichaelClerx]

Questions:

• Why is there a plot argument? Setting it to True gives an error if no ax or fig path is set.
• Can we stick to a single order for t, v, i throughout? This one is especially annoying as it goes i, t, segment_info, v

[MichaelClerx]

Should also be using detect_ramp_bounds instead of replicating that code?

[joeyshuttleworth]

These suggestions sound very sensible.

• "Plot" flag is bugged , but I don't think has been exposed because we must be setting it to True all the time?
• Happy to reorder the parameters. We should do a separate pull request for that and choose one ordering
• Detect ramp bounds should work here, you need to pick the last ramp, i.e. ramp_no=-1. It's not immediately obvious what this one-liner is doing without being familiar with how voltage_segments is setup. So this change would make it much more readable. We also would have an exception being thrown if there isn't a ramp (definitely could happen if you use some other protocol).

A bit tangentially, maybe "ramp_no" isn't the best parameter name for detect_ramp_bounds. Perhaps "ramp_index" is clearer

[joeyshuttleworth]

Plotting options seem a bit confused in the original code. I think I got confused in trying to give a few options: allowing the user to provide an axis, allowing the user to save a plot to disc without providing an axis, and not plotting anything at all. It might be simpler to only deal with an "ax" parameter, and not deal with any figure creation.

In any case, the plot parameter isn't really needed, and we should just get rid of it.

[MichaelClerx]

@joeyshuttleworth OK to merge this?