Problem with Format Regex in Single Entry JSON schema file?

[cew821]

I think the Regex added in #269 for the format field is not matching the "new" Microsoft MIME types, i.e.

application/vnd.openxmlformats-officedocument.spreadsheetml.sheet

Could someone more fluent in regex double check this for me?

[philipashlock]

Seems to be working to me. Here's a test: http://bl.ocks.org/philipashlock/raw/9144111/
Here's the gist of that test: https://gist.github.com/philipashlock/9144111

What makes you think it's not working?

[cew821]

@gbinal recently tried to run the data.gov harvester on Energy's json file. It seemed to throw an error on all of the files with that mime type (see example screenshot below). My understanding is that data.gov's harvester is using this file as the source for its validation. It's possible this issue is with data.gov and not this file, however.

[mhogeweg]

@cew821 do you have the link to the energy data.json file?

[philipashlock]

Yeah, it might not be updated on data.gov harvester just yet, but from what I can tell it's working fine on that data.json. The only obvious recurring problem I see on Energy's data.json is that references is not in an array.

The JSON Schema file also does not yet have a regex for date fields that supports the full range of possible syntax for ISO 8601 dates, so some valid dates might not validate yet, but that's an issue with the JSON Schema file.

[cew821]

Ahhh.. upon closer inspection, the regex referenced by the data.gov harvester is:
^[-\\w]+/[-\\w]+(\\.[-\\w]+)?([+][-\\w]+)?$ vs. the regex in the schema:
^[-\\w]+/[-\\w]+(\\.[-\\w]+)*([+][-\\w]+)?$

Looks like the harvester needs to be updated to use the * (zero or more) instead of the ? (zero or one). Closing this issue.

Re Energy's data.json: I've got a "fixed" version of our file that addresses the references problem, but the harvester was still choking on a few date + format things. Should be addressed shortly.