generic_unittest

Module: tests.generic_unittest

Provides a class to implement a generic unittest

The unittests will only instantiate the given class with either a default input set (see test_default_data) or will interpret the data given by the user. In the case that there already is a specialized unittest available, this class will not be called for that specific node.

Author:Andrei Ignat, Mario Michael Krell Created:2014/05/02

Inheritance diagram for pySPACE.tests.generic_unittest:

Class Summary

ParametrizedTestCase([methodName, ...])	This class acts as a wrapper such that different parameters can be passed on to the test case.
OutputTestCase([methodName, current_node, ...])	Serves as a wrapper to test the output given the default input set
InputOutputTestCase([methodName, ...])	Serves as a wrapper for an easy way to test the input-output relation
GenericTestCase([methodName, current_node, ...])	Contains the methods and submethods needed to run the tests

Function Summary

single_node_testing(node_name)	This function facilitates the testing of a single node referred to by using its representation in the DEFAULT_NODE_MAPPING variable.
multiple_node_testing([verbose, report])	This function ensures the testing of all available nodes.

Classes

ParametrizedTestCase

class pySPACE.tests.generic_unittest.ParametrizedTestCase(methodName='runTest', current_node=None, specific_input=None, desired_output=None)

Bases: unittest.case.TestCase

This class acts as a wrapper such that different parameters can be passed on to the test case. It is adapted from the example found here:

http://eli.thegreenplace.net/2011/08/02/python-unit-testing-parametrized-test-cases/

Parameters

methodName:The name of the test to be run current_node:A reference to the node. Should reference the node object and NOT A STRING with the same name specific_input:If a specific input is to be used, it should be defined here

The reason behind the existence of this class is that passing parameters to a class that extends GenericTestCase is close to impossible. If the user does not want to implement an external unittesting package such as nose-parametrized, one must ‘fool’ Python and use a wrapper class such that the unit testing is done in a different class. The end result is that the unittest implementation is done within a class that has an initialization method which obviously accepts external parameters (such as different nodes) and which takes the actual tests from a class that inherits from GenericTestCase.

While this is not the most elegant solution, it was preferred over importing a new module which would have just enlarged the list of the framework dependencies.

Class Components Summary

_get_the_call()	wrapper to get the YAML call
_initialize_node()	initializes the node with the parameters from YAML
_look_for_yaml(the_docstring)	takes a string as an argument and looks for a YAML
_non_trainable_execute()	Execution call for non-trainable nodes
_nspaces(line)	returns the indentation
_trainable_execute()	Execution call for trainable nodes
_which_input(node)	queries the node to find out what type of input it takes
parametrize(current_testcase[, node, input, ...])	Instantiate a new testcase for current_node

__init__(methodName='runTest', current_node=None, specific_input=None, desired_output=None)

static parametrize(current_testcase, node=None, input=None, output=None)

Instantiate a new testcase for current_node

_nspaces(line)

returns the indentation

Parameters

line:string representation of the current line

_look_for_yaml(the_docstring)

takes a string as an argument and looks for a YAML code block inside the string

Parameters

the_docstring:the string in which the method should look for YAML code

_get_the_call()

wrapper to get the YAML call

_initialize_node()

initializes the node with the parameters from YAML

If no YAML specification exists at all, use default initialization.

_which_input(node)

queries the node to find out what type of input it takes

Parameters

node:Reference to the node that is currently tested

_trainable_execute()

Execution call for trainable nodes

_non_trainable_execute()

Execution call for non-trainable nodes

OutputTestCase

class pySPACE.tests.generic_unittest.OutputTestCase(methodName='runTest', current_node=None, specific_input=None, desired_output=None)

Bases: pySPACE.tests.generic_unittest.ParametrizedTestCase

Serves as a wrapper to test the output given the default input set

Principle

The OutputTestCase serves as a framework for quick checks of a node whereby, for the default input set, which can be found in , test_default_data , a user defined output is expected

Class Components Summary

test_output()

compare the output with the expected output

test_output()

compare the output with the expected output

InputOutputTestCase

class pySPACE.tests.generic_unittest.InputOutputTestCase(methodName='runTest', current_node=None, specific_input=None, desired_output=None)

Bases: pySPACE.tests.generic_unittest.ParametrizedTestCase

Serves as a wrapper for an easy way to test the input-output relation

Principle

The InputOutputTestCase serves as a framework for quick checks of a node whereby, for a user defined input, a user defined output is desired.

Class Components Summary

test_input_output()

obtain output for given input and compare it with desired output

test_input_output()

obtain output for given input and compare it with desired output

GenericTestCase

class pySPACE.tests.generic_unittest.GenericTestCase(methodName='runTest', current_node=None, specific_input=None, desired_output=None)

Bases: pySPACE.tests.generic_unittest.ParametrizedTestCase

Contains the methods and submethods needed to run the tests

• whether the node has some sort of documentation
• whether an exemplary call is present in the documentation
• whether the node can be initialized with the default data set
• whether the node can execute on the default data set

An example of how to implement the generic tests in a node specific manner can be found under test_time_domain_features

This example is also explained in detail in the tutorial file

Class Components Summary

shortDescription()	overwritten method that will display the node name as output
test_execution()	execute the node using the default data
test_has_documentation()	check if the node has some sort of documentation
test_has_exemplary_call()	check if there is an exemplary call in the documentation
test_initialize()	check if the node can be initialized using the default data

shortDescription()

overwritten method that will display the node name as output

test_has_documentation()

check if the node has some sort of documentation

test_has_exemplary_call()

check if there is an exemplary call in the documentation

test_initialize()

check if the node can be initialized using the default data

test_execution()

execute the node using the default data

Principle

The test_execution method runs the node on the default data set and then checks whether the type of the output corresponds to the according output type from the theoretical point of view

Functions

single_node_testing()

pySPACE.tests.generic_unittest.single_node_testing(node_name)

This function facilitates the testing of a single node referred to by using its representation in the DEFAULT_NODE_MAPPING variable.

multiple_node_testing()

pySPACE.tests.generic_unittest.multiple_node_testing(verbose=False, report=False)

This function ensures the testing of all available nodes. The results of the test are packed into an HTML file which is saved in the current working directory.