Spec compliance

[simleo]

In our data package examples we should ensure compliance with the specs at https://specs.frictionlessdata.io/data-package. For instance, "name": "objectsTable" is not valid because "name" cannot contain uppercase characters, so the value might be changed to "objects_table". Validity can be checked with:

import datapackage
datapackage.DataPackage("datapackage.json").validate()

And here's the error for the above case:

ValidationError: u'objectsTable' does not match u'^([a-z0-9._-])+$'

Failed validating u'pattern' in schema[u'properties'][u'resources'][u'items'][u'properties'][u'name']:
    {u'description': u"An identifier for this package. Lower case characters with '.', '_' and '-' are allowed.",
     u'pattern': u'^([a-z0-9._-])+$',
     u'title': u'Name',
     u'type': u'string'}

On instance[u'resources'][0][u'name']:
    u'objectsTable'

Ensuring that generated data packages will be compliant is trickier, because we'd need to perform the same checks that datapackage does on configuration files etc. Rather, it should be the user's responsibility to check that any setting that will end up in a restricted data package field is compliant with the corresponding specs.

[danfowler]

@simleo there's an open discussion on this: frictionlessdata/datapackage#463

[simleo]

@danfowler thanks! Another thing I'm wondering about is whether there is a way to know which version of the specs a given version of the Python API is implementing (and if older specs are still available).

[danfowler]

@simleo thanks for raising the issue: frictionlessdata/datapackage-py#149

@roll is the best one to answer this