def main_test(self, option_function):
test_dir = os.path.dirname(os.path.abspath(__file__))
eval_dir = os.path.join(test_dir, '..', 'eval')
result_dir = os.path.join(test_dir, 'result')
refer_dir = os.path.join(test_dir, 'refer')
bleualign.log = lambda a, b:None
compare_files = []
for test_set, test_argument in [('eval1957', '-d'), ('eval1989', '-e')]:
fr_text = []
de_text = []
for filename in os.listdir(eval_dir):
if filename.startswith(test_set):
attr = filename.split('.')
if len(attr) == 3:
filepath = os.path.join(eval_dir, filename)
if attr[1] != 'clean':
if attr[2] == 'fr':
fr_text.append(filepath)
elif attr[2] == 'de':
de_text.append(filepath)
fr_text.sort()
de_text.sort()
test_files = []
test_files.append((fr_text[0:1], de_text[-3:-2], 'articles'))
test_files.append((fr_text, [], 'sentences'))
test_files.append((fr_text, de_text, 'sentences'))
for fr_file, de_file, filter_type in test_files:
srctotarget_file = fr_file
targettosrc_file = de_file
output_file = self.output_file_path(srctotarget_file, targettosrc_file)
output_path = os.path.join(result_dir , output_file)
options = getattr(self, option_function)(test_argument, filter_type,
srctotarget_file, targettosrc_file, output_path)
a = Aligner(options)
a.mainloop()
output_src, output_target = a.results()
output_src_bad, output_target_bad = a.results_bad()
if option_function == 'fileObjectOptions':
output_src.close()
output_target.close()
output_src_bad.close()
output_target_bad.close()
refer_path = os.path.join(refer_dir , output_file)
compare_files.append((output_path + '-good-s', refer_path + '-good-s', output_src))
compare_files.append((output_path + '-good-t', refer_path + '-good-t', output_target))
compare_files.append((output_path + '-bad-s', refer_path + '-bad-s', output_src_bad))
compare_files.append((output_path + '-bad-t', refer_path + '-bad-t', output_target_bad))
for result_path, refer_path, output_object in compare_files:
self.cmp_files(result_path, refer_path, output_object)
if option_function.startswith('file'):
os.remove(result_path)
def main_test(self, option_function):
test_dir = os.path.dirname(os.path.abspath(__file__))
eval_dir = os.path.join(test_dir, '..', 'eval')
result_dir = os.path.join(test_dir, 'result')
refer_dir = os.path.join(test_dir, 'refer')
bleualign.log = lambda a, b: None
compare_files = []
for test_set, test_argument in [('eval1957', '-d'),
('eval1989', '-e')]:
fr_text = []
de_text = []
for filename in os.listdir(eval_dir):
if filename.startswith(test_set):
attr = filename.split('.')
if len(attr) == 3:
filepath = os.path.join(eval_dir, filename)
if attr[1] != 'clean':
if attr[2] == 'fr':
fr_text.append(filepath)
elif attr[2] == 'de':
de_text.append(filepath)
fr_text.sort()
de_text.sort()
test_files = []
test_files.append((fr_text[0:1], de_text[-3:-2], 'articles'))
test_files.append((fr_text, [], 'sentences'))
test_files.append((fr_text, de_text, 'sentences'))
for fr_file, de_file, filter_type in test_files:
srctotarget_file = fr_file
targettosrc_file = de_file
output_file = self.output_file_path(srctotarget_file,
targettosrc_file)
output_path = os.path.join(result_dir, output_file)
options = getattr(self,
option_function)(test_argument, filter_type,
srctotarget_file,
targettosrc_file,
output_path)
a = Aligner(options)
a.mainloop()
output_src, output_target = a.results()
output_src_bad, output_target_bad = a.results_bad()
if option_function == 'fileObjectOptions':
output_src.close()
output_target.close()
output_src_bad.close()
output_target_bad.close()
refer_path = os.path.join(refer_dir, output_file)
compare_files.append((output_path + '-good-s',
refer_path + '-good-s', output_src))
compare_files.append((output_path + '-good-t',
refer_path + '-good-t', output_target))
compare_files.append((output_path + '-bad-s',
refer_path + '-bad-s', output_src_bad))
compare_files.append(
(output_path + '-bad-t', refer_path + '-bad-t',
output_target_bad))
for result_path, refer_path, output_object in compare_files:
self.cmp_files(result_path, refer_path, output_object)
if option_function.startswith('file'):
os.remove(result_path)
'output-src':
None,
'output-target':
None,
'output-src-bad':
None,
'output-target-bad':
None,
# other options ...
}
options['filter'] = 'sentences'
options['filterthreshold'] = 66
a = Aligner(options)
a.mainloop()
output_src, output_target = a.results()
output_src_bad, output_target_bad = a.results_bad(
) # if you set options['filter']
# output_src is StringIO because options['output-src'] is None
src = output_src.getvalue() # StringIO member function
trg = output_target.getvalue().splitlines() # array of string
print('output_src.getvalue()')
print(src[:30])
print()
print('output_target.getvalue().splitlines()')
print(trg[:3])
print()
print('filterthreshold for choice good part of alignment: {0}%'.format(
options['filterthreshold']))
print('number of good/bad alignmemts: {0}'.format(
len(output_src.getvalue().splitlines()),
len(output_src_bad.getvalue().splitlines())))
# they can be filenames, arrays of strings or io objects.
'srcfile':os.path.join(current_path, '..', 'eval', 'eval1989.de'),
'targetfile': os.path.join(current_path, '..', 'eval', 'eval1989.fr'),
# translations of srcfile and targetfile, not influenced by 'factored'
# they can be filenames, arrays of strings or io objects, too.
'srctotarget': [os.path.join(current_path, '..', 'eval', 'eval1957.europarlfull.fr')],
'targettosrc': [],
# passing filenames or io object for them in respectly.
# if not passing anything or assigning None, they will use StringIO to save results.
'output-src': None, 'output-target': None,
'output-src-bad': None, 'output-target-bad': None,
# other options ...
}
options['filter'] = 'sentences'
options['filterthreshold'] = 66
a = Aligner(options)
a.mainloop()
output_src, output_target = a.results()
output_src_bad, output_target_bad = a.results_bad() # if you set options['filter']
# output_src is StringIO because options['output-src'] is None
src = output_src.getvalue() # StringIO member function
trg = output_target.getvalue().splitlines() # array of string
print('output_src.getvalue()')
print(src[:30])
print()
print('output_target.getvalue().splitlines()')
print(trg[:3])
print()
print('filterthreshold for choice good part of alignment: {0}%'.format(options['filterthreshold']))
print('number of good/bad alignmemts: {0}'.format(len(output_src.getvalue().splitlines()), len(output_src_bad.getvalue().splitlines())))
