MCF version validation before running pygeometa about pygeometa HOT 10 CLOSED

Fixed thanks @RousseauLambertLP good catch!

from pygeometa.

Thanks @alexandreleroux. Questions:

• MCF version is a required value, so if it's missing it's an error
• how do we manage MCF versions? Suggest we support with pygeometa versions major versions
• what's the nature of the behaviour we're looking for? i.e.:
  • pygeometa manages versions in an n-1 fashion, i.e. its current version and the previous major version?
  • if a version is not supported do we throw error or warning and parse as best we can?

from pygeometa.

@tomkralidis good questions

• I confirm: if an MCF file doesn't state its version, then an error is provided
• Management of versions, when a new MCF version is created:
  • the sample.yml should be updated accordingly
  • pygeometa documentation should be updated accordingly
  • pygeometa should be tested against this new MCF version
  • if tests succeed, the next pygeometa release must support the new MCF version at the same time that the new MCF version specification is released
• if pygeometa is being ran against an MCF version that is not supported, I agree with throwing an error and trying to parse nonetheless

It's still unclear how MCF versions will evolve. Maybe it will be worth later on to have a script that automatically update MCF files from one version to the latest one. We'll see with usage.

Makes sense?

from pygeometa.

+1, will have something soon.

from pygeometa.

Implemented in #87. Documentation needs to be updated.

from pygeometa.

Looking at the code, mcf_versions is initialized twice (L154 and L224) and used only in line 226.

I think we could remove one of the line where it's initialized. I tested it without line 154 and it works fine.

from pygeometa.

I did a short version section in the MCF doc.

See:

https://github.com/RousseauLambertLP/pygeometa/blob/mcf-version-doc/doc/MCF_Reference.md#version

from pygeometa.

thanks @RousseauLambertLP , good work :-) Comments:

• typo: pygeometa doesn't have a capital P
• why do we state that pygeometa will support "previous major MCF versions"? I'm not certain we want that! :-)
• typo: sentence ends with a , instead of a .

from pygeometa.

I though from the previous comments that we wanted to support the previous (n-1) major version. We could removed it for now since we are at version 1...

from pygeometa.

Closing the issue, pygeometa validates MCF version and the doc was updated accordingly.

from pygeometa.