APFAKE

Have you ever thought "I should record the AP API test that's coming in a half hour" only to find out that it actually concluded a half hour AGO? I've done this several times. And if I've done it before, I have to assume others have done it as well.

Taking the FINAL JSON state file of an AP API election test (e.g., what happens if you hit the test URL after a test has run but before it's zeroed out in the moments before a new test or a live election), apfake will generate a file of zeros and then n intermediate states between the zeros and the final file with a smoothly incrementing set of votes.

In short: If you give it the last file of a test, apfake will generate all the missing files for you.

CAVEATS

• apfake does not call races.
• apfake votes do not properly match actual election night spikes when ballots are counted.
• apfake precincts reporting and precincts reporting percentages are probably wrong.
• apfake does not work on zeros or initialization data. It needs the format &results=ru from the AP API.

USAGE

Installing

apfake is easiest to install straight from PyPi. It requires a Python 3 environment.

pip install apfake

Using

apfake -f 2018-03-20-final.json -n 10 -d /tmp/foo/ -r 2018-03-20

apfake accepts four command-line options, two are required.

• -f, --final-results-path REQUIRED The path to the final results JSON file.
• -n, --number REQUIRED The number of files to generate.
• -r, --racedate OPTIONAL The YYYY-MM-DD racedate of the election. If the file specified in final-results-path has a racedate in the filename, this will be used. Otherwise, apfake will raise a ValueError.
• -d, --data-directory OPTIONAL A directory or set of directories to write the data to. apfake will create these directories if they do not exist, and will also create a directory named for the racedate to hold the output files. Defaults to /tmp/.

DEBUGGING WITCHERY

This is the shape that apfake expects your JSON to have. If you use the format &results=ru it will probably work. If you use a different format, it might work, though YMMV.

"""
results.keys()
dict_keys(['electionDate', 'timestamp', 'races', 'nextrequest'])

results['races'][0].keys()
dict_keys(['test', 'raceID', 'raceType', 'raceTypeID', 'officeID', 'officeName', 'party', 'seatName', 'reportingUnits'])

result['races][0]['reportingUnits'][0].keys()
dict_keys(['statePostal', 'stateName', 'level', 'lastUpdated', 'precinctsReporting', 'precinctsTotal', 'precinctsReportingPct', 'candidates'])

results['races'][0]['reportingUnits'][0]['candidates'][0].keys()
dict_keys(['first', 'last', 'party', 'incumbent', 'candidateID', 'polID', 'ballotOrder', 'polNum', 'voteCount'])
"""