Added pm.Data to track data used in model

[brendan-m-murphy]

• Summary of changes

Using pm.Data for external data used in the model will make sure that data is added to the InferenceData returned from sampling, and will apply dimensions and coordinates to that data.

This will make it easier to construct the outputs from the inversion.

I also added a check to make sure min_error is the correct size.

• Please check if the PR fulfills these requirements

• Closes #xxxx (Replace xxxx with the Github issue number)
• Tests added and passing
• Documentation and tutorials updated/added
• Added an entry to the CHANGELOG.md file
• Added any new requirements to requirements.txt

[alexdanjou]

Looks great, thanks Brendan!

[alexdanjou]

My test run failed because min_error was a 0-dim array and not a float. Would suggest to change l.287 to
if isinstance(min_error, float) or min_error.shape==(): (or check why min_error arrives as a 0-d array)