Validation tests

[mariagould]

In addition to json schema validation, we will use the following tests to validate ROR metadata records:

• Check that language code is iso639 compliant and two characters long
• Check country code is iso3166 compliant and two characters long
• Check that geonames data is the same as the data from geonames
• Check that relationship ids mentioned in the record that’s being checked exists in the rest of the record files, i.e. if the record being checked has a relationship labeled “related” and is pointing to id: 2, record with id: 2 is also pointing back to the relationship of the file being checked. Similarly, check that parent has a child relationship and vice versa in the record corpus being indexed.
• Check that links and wikipedia url are urls and do resolve
• Check that established is between 2 and 5 characters