Consistency between CLI and Python interfaces.

[oerc0122]

Currently, there are some cases (e.g. temp_max, temp_min vs. temp_start, temp_end) where python keywords do not align with CLI. These should be made uniform.

In order to rectify this, depending on how we feel about API breaking changes at this stage, we can do the following.

If CLI name adopted in Python:

• rename the positional argument to the new name
• Add the alias (old) term to KW only block
• Raise DeprecationWarning on using old kw.

If Python name adopted in CLI:

• Add new term as option
• Move the alias (old) term to a new argument group (c.f. Group CLI options #484) (Deprecated)
• Raise DeprecationWarning on using old term.

If we're not that concerned (code considered beta?) we can just rename the terms as is and update the tutorials accordingly.

[ElliottKasoar]

If we're not that concerned (code considered beta?) we can just rename the terms as is and update the tutorials accordingly.

I would still consider the code to be in beta, but given the number of people who do seem to be using it practically, it's probably worth deprecating arguments where possible, as long as it doesn't inhibit the changes we want to make.

This isn't quite the focus of the issue, but while we're at it, I also wonder if we want to change from --model-path just to --model, since in a lot of cases it's not a path to a model, but a model name etc.

This also makes it clearer that you shouldn't need to pass model names via calc-kwargs (as we still do in various places).

[alinelena]

+1 for model instead of model-path.

[ElliottKasoar]

While we're breaking things, we should consider improving the name of filter_func, as discussed in #311, which doesn't need to be tied to #313, although it's worth considering the future inclusion of constraints.