fix license diffs to match new data format

[Zach-Johnson]

Fixes data format issues from #191

[Zach-Johnson]

@AyanSinhaMahapatra I'm not sure how to handle the category field -- it looks like this was some kind of high-level identifier of the type of license and its no longer present in the new format. It seems like it was just used for scoring the delta. I'm not sure how you want to handle this now.

[Zach-Johnson]

@AyanSinhaMahapatra I updated the PR to handle the category by pulling the categories out of the license_references field - let me know if that will work. I think this PR is ready to review.

[AyanSinhaMahapatra]

@Zach-Johnson Thanks++ for the PR. ❤️

I updated the PR to handle the category by pulling the categories out of the license_references field - let me know if that will work

Yes this is indeed the correct way to get these categories, as license reference info is now at codebase-level.

Three things though:

1. About the test failures: the test fixtures (SCTK scans) are massively outdated, these are mostly from SCTK v2.x which is 4-5 major versions behind current SCTK and these need to be updated. Ideally we should have had scripts to re-generate these fixtures from latest as default, or atleast links/instructions to regenerate them, but that's not the case here. So I'll write something up to update the output format temporarily (similar to migrations) to pass the tests before we replace these with actual scans later. And then you can merge/rebase on that and we can look at any other test failures remaining. 😅
2. This would not affect the function but should we also change the models like in https://github.com/Zach-Johnson/deltacode/blob/issue-191-license-fixes/src/deltacode/models.py#L244 to reflect more the current data models at https://github.com/nexB/scancode-toolkit/blob/develop/src/licensedcode/models.py so this is cleaner and less confusing?
3. We should add some integration tests which would install deltacode with latest SCTK/commoncode to check various functions

---

1. and 3, is out of scope for this PR and I'll take care of this, otherwise the changes seem to be correct, but we have to bring the tests back to passing and then apply these changes to verify eveything is working as expected before merging this in. Should not be too long, thanks again for the PR.

[Zach-Johnson]

@AyanSinhaMahapatra I updated the models. I will wait on the test updates to rebase!

[pombredanne]

@Zach-Johnson gentle ping... unless @AyanSinhaMahapatra needs to do something first?

[Zach-Johnson]

@Zach-Johnson gentle ping... unless @AyanSinhaMahapatra needs to do something first?

I was under the impression that @AyanSinhaMahapatra was going to work on fixing up all the tests!