Data profiles with only TXT resources should not be harvested

[higorspinto]

Example: https://us-ed-testing.ckan.io/dataset/the-education-innovator-december-7-2010

We need to avoid harvesting data profiles that only have TXT files as resources, as they are very likely to be false positives.

We can deal with this in the sanitize transformer, rather than in the parser / Scrapy pipeline.

Jira Card

Acceptance Criteria

• [x] Have datasets with only txt files excluded after the sanitizing process

[higorspinto]

time spent: 4h