dataetst is a python library to be used in verifying output from out business processes.

datatest exports a single class datatest.DataTest which expects a pandas.DataFrame in its constructor. datatest.DataTest exposes four methods:

• equals(row_id, column, value) - tests that the value at the given row, col index is of the same type and equal to the input value
• matches(row_id, column, regex) - tests that the value at the given row, col index matches regex
• matches_all(column, regex) - tests that all values in the given column match regex
• finish() - report errors and exit with appropriate status code

DataTest.finish() must be called to recieve the test reporting and error code. The module will complain loudly if a test exits before DataTest.finish() is called.

the script

#!/usr/bin/env python2.7
""" simple demonstration of compass test class """

from datatest import DataTest
import pandas as pd

TDF = pd.DataFrame({'char':['a', 'b', 'c'], 'num':[1, 2, 3]}, index=['j1', 'j2', 'j3'])

DATA_TEST = DataTest(TDF)
DATA_TEST.equals('j1', 'char', 'a')
DATA_TEST.equals('j1', 'num', 0) # this test fails
DATA_TEST.matches('j1', 'char', r'\D')
DATA_TEST.matches('j1', 'num', r'\d')
DATA_TEST.matches_all('char', '[a-z]')
DATA_TEST.finish()

would show output:

ERROR: 1/5 tests failed:

Bad value in num at id: j1
                Expected 0
                Found 1

the script

#!/usr/bin/env python2.7
""" simple demonstration of compass test class """

from datatest import DataTest
import pandas as pd

TDF = pd.DataFrame({'char':['a', 'b', 'c'], 'num':[1, 2, 3]}, index=['j1', 'j2', 'j3'])

DATA_TEST = DataTest(TDF)
DATA_TEST.equals('j1', 'char', 'a')
DATA_TEST.equals('j1', 'num', 1) # this test passes 
DATA_TEST.matches('j1', 'char', r'\D')
DATA_TEST.matches('j1', 'num', r'\d')
DATA_TEST.matches_all('char', '[a-z]')
DATA_TEST.finish()

would show output:

OK: 5 tests passed

datatest is a helper script included with the datatest module which locates and executes tests. Its behavior is as follows:

• If no command line arguments are given, it recursively descends into ./testing and calls python on all files ending with _datatest.py
• If command line arguments are given, it calls python on all arguments ending with _datatest.py
• In either case it prints the name of the file being executed padded with ##### and its exit code is 0 if all tests pass, 1 (with a special message) if no tests are found, and the number of test files with errors in all other cases.

consider the ```testing/`` directory of this repo. its contents are as follows:

testing
├── error_datatest.py
└── ok_datatest.py

running datatest in the base directory of ths repo exits with code 1 and produces

#########################
testing/error_datatest.py
#########################

WARNING: The following IDs were not tested at all: j2 j3
ERROR: 3/6 tests failed:

Bad value in num at id: j1
                Expected 0
                Found 1

Bad format in num at id j1:
                Expected to match \D
                Found 1

Bad format in num at id j3:
                Expected to match [012]
                Found 3


######################
testing/ok_datatest.py
######################

OK: 7 tests passed

running datatest testing/ok_datatest.py exits with code 0 and produces

######################
testing/ok_datatest.py
######################

OK: 7 tests passed

running datatest fooBar.js exits with code 1 and produces

No tests found