extracter conflicts with a before_index method about ckanext-extractor

[mystycs]

FYI: This happens in the ckan-worker background job

[mystycs]

Looks like it was an issue with how the solr index was updated, it was running a bunch of other stuff like before_index etc. I used this instead in tasks.py to update the search-index line 110

ckanext-extractor/ckanext/extractor/tasks.py

Line 110 in 811b388

index_for('package').update_dict(pkg_dict)

ckan.lib.cli.SearchIndexCommand('SearchIndexCommand').rebuild(res_dict['package_id'])

[torfsen]

@mystycs It is intended that implementations of before_index in all active plugins are run when ckanext-extractor updates Solr's index for a package. So if you have a before_index implementation that works without ckanext-extractor, then it should be called when ckanext-extractor runs and that call shouldn't fail due to something that ckanext-extractor does.

Judging from your stacktrace, data_dict.get('business_area', '[]') doesn't return a string for you. Have you checked what type that value actually has?

[mystycs]

It inputs a multi valued checkbox from https://github.com/ckan/ckanext-scheming

So it looks like ["tag1", "tag2", "tag3" ]. This is so that tags rendered on the dataset page dont look like lists but rather are split up separately.

[mystycs]

But using my search index update method did the trick. And im having no issues. So this may just be a use case for me.