GSA-TTS / FAC

GSA's Federal Audit Clearinghouse
Other
20 stars 6 forks source link

FAC - Possible bug: faulty gaap_results information through /general #2608

Closed stucka closed 1 month ago

stucka commented 1 year ago

Describe the bug

I'm not an accountant, but I seem to be seeing that there are several distinct categories for opinions reflected in gaap_results. The system is combining some.

non-GAAP seems to be an additional flag to the categories, so I can see that co-existing. But I'm not sure what a unmodified, adverse opinion would be; from what I can tell, that's a contradiction in terms. From a recent pull of /general:

I see no data for disclaimer_of_opinion entries to evaluate whether they're suffering from any overlap problems.

I don't know that "unmodified opinion" can co-exist with "not_gaap".

Steps to reproduce the bug

Where the environment contains an API key in FED_API_KEY:

import requests
import os
fedapikey = os.environ['FED_API_KEY']
requestheaders = {"X-Api-Key": fedapikey}
r = requests.get(f"https://api.fac.gov/general", headers=requestheaders)
resultsdict = {}
for entry in r.json():
    results = entry['gaap_results']
    if results not in resultsdict:
        resultsdict[results] = 0
    resultsdict[results] += 1
print(resultsdict)

Expected Behavior

https://www.ecfr.gov/current/title-2/subtitle-A/chapter-II/part-200/subpart-F/subject-group-ECFRea73e47c9a286e6/section-200.515

200.515(d)(1)(i) seems to suggest that unmodified, qualified, adverse and disclaimer opinions are exclusive categories within the GAAP-compliant audits.

Screenshots

No response

System setup

No response

Additional context

No response

Code of Conduct

cephillips commented 1 year ago

Hmm, I don't actually recall an unmodified opinion being part of it. I thought it was Unqualified Opinion, Qualified Opinion, Adverse Opinion, Disclaimer of Opinion. Unqualified being fine and Disclaimer being so bad that the auditor couldn't even make an opinion.

jadudm commented 10 months ago

I missed this bug ticket. We need to discuss it, and potentially bring it forward soon.

danswick commented 3 months ago

Quick follow-up here: we've learned that some submitters have been reusing the same finding reference numbers for multiple distinct findings. We think this is either a hold-over practice from Census or reflects a difference between Census' and GSA's data models. We would like to explore stricter validations for these cases soon.

sambodeme commented 2 months ago

Quick follow-up here: we've learned that some submitters have been reusing the same finding reference numbers for multiple distinct findings. We think this is either a hold-over practice from Census or reflects a difference between Census' and GSA's data models. We would like to explore stricter validations for these cases soon.

@danswick This validation has been merged

danswick commented 1 month ago

We implemented this validation and then very quickly learned from the auditor community that our validation was too aggressive and was blocking some valid findings combinations. More detail here, but the issue in this ticket has been both resolved and subsequently reverted.