""" """ # TODO: Put a comment above. import lines_of_code import main_for_testing import abc import importlib.util import unittest class Tester(): """ Base class for testing student code. """ # TODO: Augment the above comment. def __init__(self, what_to_grade, who_to_grade, where_to_grade): """ Tests the given what_to_grade for the students specified by the given who_to_grade, given that checked-out projects are in the given where_to_grade folder. type: what_to_grade: ThingToGrade type: who_to_grade: WhoToGrade type: where_to_grade: str """ self.what_to_grade = what_to_grade self.who_to_grade = who_to_grade self.where_to_grade = where_to_grade def do_tests_on_students(self): """ Does this Tester's tests on this Tester's what_to_grade for ALL of the students specified by its who_to_grade, using its repo_helper. Also prints appropriate messages in doing so. Returns a dictionary of results whose keys are the students. :rtype dict(str) """ self.initialize_tests() results = {} for student in self.who_to_grade.students: results[student] = self.do_tests_on_student(student) return results def do_tests_on_student(self, student): """ Does this Tester's tests on this Tester's what_to_grade for the given student, using this Tester's repo_helper. Prints appropriate messages, runs the test, and extends the returned result to include the student and what was tested. :rtype TesterResult """ print() print('!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!') print('TESTING:', student) print('!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!') result = self.run_tests(student) result.student = student result.what_was_tested = self.what_to_grade return result def initialize_tests(self): """ Called (once) by do_tests prior to doing tests on all the students. """ # Override in subclass if needed. pass @abc.abstractmethod def run_tests(self, student): """ Runs all of this Tester's tests on the given student. Returns the result of the test as a TesterResult. :rtype TesterResult """ class TesterResult(object): def __init__(self, student=None, what_was_tested=None): """ The result from running a Tester's tests on its WhatToGrade for a single student. :type student: str :type what_was_tested: WhatToGrade """ # TODO: Clarify the above comment. # TODO: This is a base class, describe and organize it as such. self.student = student self.what_was_tested = what_was_tested class ChangesTester(Tester): """ Tests by examining how much (if at all) each module has changed since the student received it. """ def __init__(self, what_to_grade, who_to_grade, where_to_grade): """ Tests the given what_to_grade for the students specified by the given who_to_grade, given that checked-out projects are in the given where_to_grade folder. Tests by examining how much (if at all) each module has changed since the student received it. type: what_to_grade: ThingToGrade type: who_to_grade: WhoToGrade type: where_to_grade: str """ super().__init__(what_to_grade, who_to_grade, where_to_grade) # The following are set by initialize_tests # (which is called by do_tests). self.stats_for_original = None self.stats_for_solution = None def initialize_tests(self): """ For each module to be tested, get and store the statistics for the original and solution versions of the module. """ original = self.what_to_grade.course.username_for_original self.stats_for_original = self.get_stats(original) solution = self.what_to_grade.course.username_for_solution self.stats_for_solution = self.get_stats(solution) def get_stats(self, student): """ Returns ... """ # TODO: Complete the above comment. # FIXME: The folder structure in the following statement # should come from elsewhere. Right now it is both here # and in RepoHelper, and it is specific to projects # as we set them up for Eclipse in 120. folder = self.where_to_grade + student + '/src/' result = {} for module in self.what_to_grade.modules: filename = folder + module # CONSIDER: Move lines_of_code stuff to this class? stats = lines_of_code.evaluate_module(filename) result[module] = stats return result def run_tests(self, student): """ """ # TODO: Augment the above comment. stats = self.get_stats(student) # stats for ALL modules result = ChangesTesterResult() vs_original = {} vs_solution = {} for module in self.what_to_grade.modules: original = self.stats_for_original[module] solution = self.stats_for_solution[module] vs_original[module] = stats[module].minus(original) vs_solution[module] = stats[module].minus(solution) result.stats = stats result.vs_original = vs_original result.vs_solution = vs_solution return result class ChangesTesterResult(TesterResult): """ The result of running a test on a student. Includes: -- student tested -- what was tested (a ThingToGrade) -- for each module that was tested: -- The StatisticsForModule for that module """ def __init__(self, student=None, what_was_tested=None, stats=None, vs_original=None, vs_solution=None): """ :type student: str :type what_was_tested: ThingToGrade :type stats: dict(str, StatisticsForModule) :type vs_original: dict(str, StatisticsForModule) :type vs_solution: dict(str, StatisticsForModule) """ # TODO: Augment the above comment. # CONSIDER: The names for the fields of this class # are poorly chosen, maybe. super().__init__(student, what_was_tested) self.stats = stats self.vs_original = vs_original self.vs_solution = vs_solution # CONSIDER: repr and str below are provided to display results. # But what we really need is to STORE and/or LOG results. # The following will eventually need to change, I suspect. def __repr__(self): format_string = 'ChangesTesterResult({}, {}, {!r}, {!r}, {!r}' return format_string.format(self.student, self.what_was_tested, self.stats, self.vs_original, self.vs_solution) def __str__(self): line1 = 'Student: {}\n'.format(self.student) line2 = 'What was tested: {}\n'.format(self.what_was_tested) module_limit = 11 # Chop module names at this many characters. format_string = '{:' + str(module_limit) + '}' format_string += ' {:>14} {:>15} {:>15} {:>15}\n' header = ('Module', 'Removed', 'vs. Original', 'vs. Solution', 'Module itself') line3 = format_string.format(*header) result = line1 + line2 + line3 modules = sorted(self.stats.keys()) for module in modules: result += self.result_for_module(module, format_string, module_limit) return result def result_for_module(self, module, format_string, module_limit): """ Returns a string representation for the given module in this ChangesTesterResult. """ # TODO: Augment the above comment # CONSIDER: The following is based on things in # StatisticsForModule that could later change # (and be missed here). Encapsulate it all in a single place. transformations = ('nothing_removed', 'wo_blank_lines', 'wo_docstrings', 'wo_comments') labels = ('Nothing', 'Blank lines', ' + docstrings', ' + comments') result = '' for k in range(len(transformations)): transformation = transformations[k] label = labels[k] vs_original = getattr(self.vs_original[module], transformation) vs_solution = getattr(self.vs_solution[module], transformation) stats = getattr(self.stats[module], transformation) trio = (vs_original, vs_solution, stats) stat = [] for k in range(len(trio)): stat.append(' {:3} {:4} {:5}'.format(trio[k].lines, trio[k].words, trio[k].characters)) if transformation == 'nothing_removed': module_name = module[:module_limit] else: module_name = '' result += format_string.format(module_name, label, *stat) return result # ---------------------------------------------------------------------- # Stuff below here is good stuff (mostly) but needs to be reworked. # ---------------------------------------------------------------------- class ReturnedValueTester(Tester): """ Tests functions in modules by, for each function to be tested, determining whether the function returns the correct value and has (only) the correct side effects. More precisely, a ReturnedValueTester does the following for each module that it tests: 1. Reads (from a file) a collection of Tests, where a Test is: -- The name of the function to call. -- The arguments to send to the function in that Test. -- The correct returned value, and the correct side effects (if any), from calling the function with those arguments. 2. For each student to be tested: -- Loads the student's module. -- Applies each Test to that student's module. -- Records whether or not the student's code returned the correct value and had the correct side effects. -- Also records the nature of the failure for failed tests. """ def __init__(self, what_to_grade, who_to_grade, where_to_grade): """ Tests the given what_to_grade for the students specified by the given who_to_grade, given that checked-out projects are in the given where_to_grade folder. type: what_to_grade: FunctionToGrade type: who_to_grade: WhoToGrade type: where_to_grade: str """ super().__init__(what_to_grade, who_to_grade, where_to_grade) # The following are set by initialize_tests # (which is called by do_tests). self.tests = None def initialize_tests(self): """ """ self.tests = {} for module in self.what_to_grade.modules: filename = self.get_filename_with_tests(module) with open(filename, 'r') as f: test_text = f.read() self.tests[module] = self.parse_test_text(test_text) def get_filename_with_tests(self, module_name): suffix = self.what_to_grade.course.suffix_for_test_files tail = module_name.split('.')[0] + suffix + '.py' head = self.where_to_grade head += self.what_to_grade.course.username_for_solution head += '/src/' return head + tail def parse_test_text(self, test_text): """ For each module to be tested, reads the tests to be applied to that module from a text file like this example: @ m2 problem2a [[4, 66, 9, -2, 55, 0], [7, 22, 5, 10, -5, 9]] [11, 88, 14, 8, 50, 9] [[], []] [] [[-1, 0, 1], [1, 0, -1]] [0, 0, 0] @ m2 problem2b [[4, 66, 9, -2, 55, 0], [7, 22, 5, 10, -5, 9]] None [[11, 88, 14, 8, 50, 9], [7, 22, 5, 10, -5, 9]] @ m2 test_problem2a # This tests a TEST function [] # No parameters None # Nothing returned In particular: -- Lines that begin with an @ indicate a function to test, with the first word after the @ being the module name and the second word being the function name. The module name can be abbreviated as indicated above. -- A Test has 2 or 3 lines: -- Line 1: a list of the arguments (So a function with one argument has [BLAH].) -- Line 2: the correct returned value -- Line 3 (if present): the correct value for the arguments AFTER the function call. (If absent, the function should not mutate the argument.) -- Tests for a module must be separated by one or more empty lines (i.e., lines with only whitespace). -- A # character and all characters on the rest of its line are ignored. Returns a list of ReturnedValueTest objects. :rtype list(ReturnedValueTest) """ returned_value_tests = [] functions_to_test = test_text.split('@') n = functions_to_test.count('') for _ in range(n): functions_to_test.remove('') for function_tests in functions_to_test: tests = function_tests.split('\n') n = tests.count('') for _ in range(n): tests.remove('') module_and_function = tests[0].strip().split(' ') module = module_and_function[0] function = module_and_function[1] # TODO: Implement allowing module abbreviations like m2 arguments = eval(tests[1]) # FIXME: ugh! Better way? returned_value = eval(tests[2]) if len(tests) > 3: arguments_after_test = eval(tests[3]) else: arguments_after_test = arguments rv_test = ReturnedValueTest(module, function, arguments, returned_value, arguments_after_test) returned_value_tests.append(rv_test) return returned_value_tests def run_tests(self, student): """ """ # TODO: Augment the above comment. folder = self.where_to_grade + student + '/src/' rv_result = ReturnedValueTestResult() result = {} for module in self.what_to_grade.modules: tests = self.tests[module] pathname = folder + module result[module] = self.run_returned_value_tests(tests, module, pathname, student) rv_result.results_by_module = result return rv_result def run_returned_value_tests(self, tests, module, pathname, student): """ :type tests: list(ReturnedValueTest) :type module: module [A real module, NOT a filename] :type pathname: str [Pathname of the file to be tested] :type student: str """ # TODO: Augment the above comment. # CRITICAL NOTE: The following is ugly. # The challenge is for this function to load a module # WITHOUT RETAINING the definitions loaded when this function # exits. We need that because if student 1 has a correct # function and student 2 does not define that function at all, # the definition from student 1 is used. # There is probably a better way to solve this problem that # what is done below, which is to load the student's module # into a UNIQUELY-NAMED module and use that uniquely-named # module when running the tests. The uniquely-named module # is obtained by appending the student's username to the # module name. That may fail if [something unforeseen]. # A better approach would be to "wipe" the namespace when # this function exits, but I don't know how to do that. module_name = module + '_' + student spec = importlib.util.spec_from_file_location(module_name, pathname) # TODO: The load_module method is deprecated, # but I haven't figured out how to do it that method. # TODO: Need to deal with inability to load, can't read tests, etc. module = spec.loader.load_module(module_name) test = ReturnedValueTestCase(tests, module) result = test.run() result.subtest_results = test.subtest_results # Extend on the fly return result class ReturnedValueTestResult(): """ """ # TODO: Augment the above comment. def __repr__(self): format_string = 'ReturnedValueTestResult({}' return format_string.format(self.results_by_module) # def __str__(self): # for result in self.result: # if result.wasSuccessful(): # print('{:8}: OK'.format(result.student)) # else: # print('{:8} ERRORS: {}'.format(result.student, # result.errors)) # print('{:8} FAILURES: {}'.format(result.student, # result.failures)) # for subtest in result.subtest_results: # if subtest[0] == 'PASSED_TEST': # continue # print('{:8} FAILED {:9}: {}'.format(result.student, # subtest[2][0], # subtest)) class ReturnedValueTest(): def __init__(self, module, function, arguments, returned_value, arguments_after_test): self.module = module self.function = function self.arguments = arguments self.returned_value = returned_value self.arguments_after_test = arguments_after_test def __repr__(self): format_string = 'ReturnedValueTest({}, {}, {!r}, {!r}, {!r}' return format_string.format(self.module, self.function, self.arguments, self.returned_value, self.arguments_after_test) class ReturnedValueTestCase(unittest.TestCase): """ A TestCase that tests a function in a module by running this TestCase's list of ReturnedValueTest instances on this TestCase's module. As such, it determines whether the function returns the correct value and has (only) the correct side effects. """ def __init__(self, tests, module): """ :type tests: list(ReturnedValueTest) :type module: module """ # Augment the above comment. self.subtests = tests self.module = module super().__init__('runSubTestsOnModule') self.subtest_results = [] def runSubTestsOnModule(self): """ """ # TODO: Augment the above comment. print('***** OUTPUT from student\'s run: *****') for test in self.subtests: f = test.function args = test.arguments rv = test.returned_value args_after_call = test.arguments_after_test with self.subTest(): error = '' # so far try: function = getattr(self.module, f) except Exception as exception: # FIXME: These should be ENUMSs. error = 'FUNCTION_NOT_IMPLEMENTED' message = 'Function {} is not implemented' triple = (error, message.format(f), test) self.subtest_results.append(triple) raise exception try: result = function(*args) except Exception as exception: error = 'THROWS_EXCEPTION' message = 'Function {} throws an exception: {}' triple = (error, message.format(f, exception), test) self.subtest_results.append(triple) raise exception try: self.assertEqual(result, rv, 'Wrong returned value') except Exception as exception: error = error + 'WRONG_RETURNED_VALUE' message = 'Expected: {}. Got: {}.' message = message.format(rv, result) try: self.assertEqual(args, args_after_call, 'Wrong mutation') except Exception as exception: if error: error = error + ' and BAD_MUTATION' message = message + ' ' else: error = error + 'BAD_MUTATION' message = '' message2 = 'Expected arguments to be: {}\nGot: {}' message2 = message2.format(args_after_call, args) message = message + message2 if not error: error = 'PASSED_TEST' message = 'OK' triple = (error, message, test) self.subtest_results.append(triple) if error != 'PASSED_TEST': raise exception print() print('***** RESULTS of tests: *****') for result in self.subtest_results: print(result) return self.subtest_results StandardTester = ReturnedValueTester def main(): main_for_testing.main() if __name__ == '__main__': main() # # class TesterTest(unittest.TestCase): # """ # A subclass of unittest.TestCase for testing whether a TESTING # function calls the function to be tested enough times. # For example, it might test whether test_blah() calls blah(...) # at least 4 times (for 4 tests). # A TesterTest: # -- Has a module (NOT name of moudule -- the module itself) # -- Has the name X of a function in that module # -- Has a positive integer N # and # -- Runs test_X() [where X is the name of the function] # (and catches and ignores any exception) # -- Counts how many times X is called # The test passes if that count >= N. Else the test fails. # """ # # def __init__(self, module, name_of_test_function, # name_of_function_it_tests, min_number_of_tests): # self.module = module # self.name_of_test_function = name_of_test_function # self.name_of_function_it_tests = name_of_function_it_tests # self.min_number_of_tests = min_number_of_tests # super().__init__('runTestsOnModule') # self.number_of_calls = 0 # test_function_name = 'test_' + self.function_name # try: # test_function = getattr(self.module, test_function_name) # # except: # message = 'Function {} is not implemented' # self.fail(message.format(test_function_name)) # return # def runTestsOnModule(self): # number_of_calls = 0 # # # If the test_function or the function it tests does not exist, # # fail immediately. # try: # test_function = getattr(self.module, # self.name_of_test_function) # except: # message = 'Function {} is not implemented' # self.fail(message.format(test[0])) # # function_it_tests = getattr(self.module, # self.name_of_function_it_tests) # # Redefine the function the test_function tests # # to include a counter. # # # Call the test_function, catching and ignoring any exceptions. # # # Test passes if number_of_calls >= min_number_of_tests # # def count_calls(function_name, ): # self.number_of_calls = self.number_of_calls + 1 # try: # # # message = 'Function {} is not implemented' # self.fail(message.format(test_function_name)) # return # # # # with self.subTest(i=test): # try: # function = getattr(self.module, test[0]) # except: # message = 'Function {} is not implemented' # self.fail(message.format(test[0])) # continue # try: # result = function(*test[1]) # except Exception as e: # message = 'Function {} throws an exception: {}' # self.fail(message.format(test[0], e)) # continue # self.assertEqual(result, test[2], 'Wrong returned value') # self.assertEqual(test[1], test[3], 'Wrong mutation') # Test more carefully: # 1. Catches inadvertant mutations? # 2. Are the catches for not-implemented and throws-exception correct? # def test_decorator(): # import m1 # old = m1.foo # def bar(): # print('bar') # old() # m1.foo = bar # m1.foo() # project = 'Session16_Test2_201430_mutchler' # src = 'src/' # module = 'm1' # suffix = '.py' # t = UnitTester(project, module) # print(t.tests) # result = t.run_tests() # print(result) # path = folder + session + src + module + suffix # t = StandardTest('m1', 'm1_tests.py') # print(t.read_tests()) # result = t.run() # print(result) # print(result.failures[0]) # t = StandardTest('m1a', 'foo') # result = t.run() # print(result) # print(result.errors) # m5 = importlib.import_module('m5') # m5.foo() # FAILS: # m = importlib.import_module('C:\\EclipseWorkspaces\\csse120\\Session16_Test2_201430_SOLUTION\\src\\m5.py') # m.foo() # folder = 'C:/EclipseWorkspaces/csse120/' # session = 'Session16_Test2_201430_mutchler/' # src = 'src/' # module = 'm6' # suffix = '.py' # path = folder + session + src + module + suffix # spec = importlib.util.spec_from_file_location(module, path) # print(spec.loader) # m = spec.loader.load_module(module) # m.foo() # # module = 'm7' # suffix = '.py' # path = folder + session + src + module + suffix # spec = importlib.util.spec_from_file_location('m6', path) # print(spec.loader) # m = spec.loader.load_module('m6') # m.foo() # FIXME: The next two change from term to term. Unify with Grader. # self.root_folder = 'C:/EclipseWorkspaces/csse120-grading' # self.term = '201430' # self.grading_folder = self.root_folder + '/' + self.term + '/' # self.folder_to_grade = self.grading_folder + self.project + '/' # # print(self.folder_to_grade) # Where the tests are: # project = 'Session16_Test2_201430_SOLUTION/' # FIXME # prefix = folder1 + project + src # suffix_for_tests = '_tests.py' # self.file_with_tests = prefix + self.module_name + suffix_for_tests # self.tests = self.read_tests() # Where the student modules are: