Validation issues with scheming multiple text

[ChristianF88]

Hi there,

I am updating an old CKAN system, that has been heavily customized. I am in the course of updating ckanext-scheming to version 3.0. Previously an older version was used with ckanext-repeating.

We're using repeating text in our dataset schema twice. Here is the author field's section:

    {
      "default_value": "context_fullname_email",
      "display_snippet": "eaw_repeating_text.html",
      "modal_text": "author_help",
      "form_snippet": "eaw_schema_repeating_text_js.html",
      "label": "Author",
      "output_validators": "repeating_text_output",
      "validators": "scheming_multiple_text eaw_schema_json_not_empty",
      "form_placeholder": "Bach, Johann <joe@eawag.ch>",
      "field_name": "author",
      "display_property": "dc:creator",
      "form_blanks": 3,
      "preset": "multiple_text",
      "required": true

    }

To migrate from ckanext-repeating to the latest version of scheming I changed the validators from repeating_text to scheming_multiple_text and I added the preset: multiple_text. Unfortunately I couldn't find any more information on how to do the migration anywhere.

Now once I test this schema, I have issue submitting the "new dataset form". Somehow the authors I enter into the repeating text fields are not recognized, as I get a missing author warning.

This is caused by the scheming_multiple_text validator of ckanext-scheming. Visualizing the validators data argument I found that the repeating text fields we use end up in a spot than expected. The repeating text fields end up in __extras. But the validator is checking for author.

{('embargo',): '',
 ('geographic_name',): '',
 ('has_part',): '',
 ('id_external',): '',
 ('is_part_of',): '',
 ('maintainer',): 'ckadm',
 ('name',): 'ghfg',
 ('notes',): 'fdg',
 ('notes-2',): '',
 ('open_data',): 'true',
 ('owner_org',): '87689fbc-3142-40de-89ab-82c760bb34f3',
 ('private',): 'False',
 ('publicationlink',): '',
 ('review_level',): 'none',
 ('reviewed_by',): '',
 ('save',): '',
 ('spatial',): '{}',
 ('status',): 'incomplete',
 ('substances',): '',
 ('substances_generic',): '',
 ('systems',): '',
 ('tags_string',): 'federal',
 ('taxa',): '',
 ('taxa_generic',): '',
 ('title',): 'ghfg',
 ('usage_contact',): 'ckadm',
 ('variables',): ['alkalinity', ''],
 ('state',): 'draft',
 ('type',): 'dataset',
 ('__extras',): {'_ckan_phase': 'dataset_new_1',
  'author-1': 'foerster, christian "<christian.foerster@eawag.ch>"',
  'author-2': '',
  'author-3': '',
  'pkg_name': '',
  'timerange-1': '*',
  'timerange-2': ''},
 ('id',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('author',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('author_email',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('maintainer_email',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('license_id',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('url',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('version',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('log_message',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('tag_string',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('return_to',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('separator_domain_metadata',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('ispublication',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('islongterm',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>',
 ('timerange',): '<ckan.lib.navl.dictization_functions.Missing object at 0x7efc9814fc70>'}

I am not sure how the data is passed and why it ends up in __extras instead of author. Does anyone have some advice on how to correct this?

Cheers!

[ChristianF88]

This problem seems to be related to this issue: ckan/ckan#6875.