Fixing gold standard script to work with circle ci

[chummels]

This should fix issue #132 so the gold standard script now work with circle_ci instead of travis.

[chummels]

Actually, no, this doesn't address everything. This merely fixes the breakage of the script in accessing the travis yml file. But all of the other text in the script that helped you actually build the gold standard was removed, presumably during the switch to circle ci. I'm not sure why. Was there something wrong with the build order that was given to the user described in the docs: https://trident.readthedocs.io/en/latest/testing.html#generating-gold-standard-answer-test-results-for-comparison ? It's useful to be explicit with the user in regenerating thegold standards, and I personally find it helpful when I haven't built the gold standards in a long time to remember the steps.

[coveralls]

Pull Request Test Coverage Report for Build 741

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 75.209%

---

Totals
Change from base Build 674:	0.0%
Covered Lines:	1796
Relevant Lines:	2388

---

💛 - Coveralls

[brittonsmith]

I'm not sure what other text you're referring to. Between this script and the docs you link to, I don't see anything else that one needs to know to run testing. The build script is written to most efficiently use the testing workflow set up by the service, nothing more. It's not meant to be used as documentation. I missed this because it's not tested functionality. I left a comment on how it could be.

[chummels]

@brittonsmith You're right. I had mistakenly thought that the text that was in the docs about how to generate the gold standards was also meant to be printed to STDOUT as part of this script. But I looked back at past versions and that was never the case. Sorry about that.

[chummels]

I have now updated the script to include a test_gold.py file that will ensure that this functionality is tested in the future. Although I don't foresee us changing the testing infrastructure again, but I guess more tests are always better.

[brittonsmith]

Excellent. This looks great. Definitely an unlikely just in case test, but I'm happy it's there. I'll merge this. Thanks for doing it.