Accessing Validation History

Any time you run validate, the comparison results are cached on the file system.

There are a few functions that provide access to previous comparisons:

from greenlight import last_comparison, all_comparisons, all_comparisons_dict

last_comparison(product_code)

Retrieve the last-known comparison for a given product.

This returns the last comparison that took place.

This contains the file-level comparisons, table-level comparisons, and any custom facts set.

prev_results = last_comparison('ff.all') # That's it!

all_comparisons(product_code)

Return a list of historical comparisons.

This returns a list of all of the known comparisons (from earliest to latest).

You can retrieve the same comparison results by date/timestamp by using all_comparisons_dict.

all_comparisons_dict(product_code)

Return a collection of historical comparisons as a dictionary.

Each key is the validation date/timestamp of the previous run.

>>> historical_comps = all_comparisons_dict('ff.all')

>>> historical_comps.keys()
    [u'20181008_20_44',
     u'20181010_12_59',
     u'20181010_13_02',
     u'20181010_16_01',
     u'20181010_16_17',
     u'20181010_16_31',
     u'20181010_17_23',
     u'20181010_17_27',
     u'20181011_16_01',
     u'20181011_16_30',
     u'20181011_16_56',
     u'20181011_17_14',
     u'20181011_17_24',
     u'20181012_15_36',
     u'20181015_13_43',
     u'20181015_13_48',
     u'20181015_14_01']

This way you can retrieve specific results.

# You can pass this to test_run()
prev_result = historical_comps['20181010_16_31']

test_run(prev_results, run_default_tests=True, custom_tests=None, log_dir=None)

You can see how a given results set responds to tests by using the test_run function.

Simulate a validation run using a previously generated set of results.

This can help you see how the validation results would be affected by using different combinations of custom tests.

You can use log_dir to specify a custom log location.

prev_result = last_comparison('ff.all')

# See how validation would be affected by a set of new/different tests
test_run(prev_result, custom_tests=[<some new tests specified here>])

# See how validation would be affected by *no* custom tests.
test_run(prev_result, custom_tests=[])

test_run(
    prev_results,
    custom_tests=[<some custom tests>,],
    log_dir='/home/wrds/josepd/tmp_log_dir', # Here I'm altering the location the log will be written to
)

test_last_comparison(product_code, run_default_tests=True, custom_tests=None, log_dir=None)

This function is a convenience wrapper around test_run() + last_comparison().

This will:

• find the last comparison (automatically)
• run the specified tests against those results
• produce a log of the results