def test_skip_missing_test_files_usage():
command_line_args = [
'--reference-input', 'tests/data/i2b2_2016_track-1_reference',
'--test-input', 'tests/data/i2b2_2016_track-1_test',
'--skip-missing-files'
]
args = args_and_configs.get_arguments(command_line_args)
assert args.skip_missing_files == True
## Performance should be identical to default
command_line_args = [
'--reference-input', 'tests/data/i2b2_2016_track-1_reference',
'--test-input', 'tests/data/i2b2_2016_track-1_test'
]
args = args_and_configs.get_arguments(command_line_args)
assert args.skip_missing_files == True
def test_print_counts_and_metrics_and_confusion_for_test():
command_line_args = [
'--print-counts', '--print-metrics', '--print-confusion-matrix',
'--test-input', 'tests/data/i2b2_2016_track-1_test'
]
with pytest.raises(SystemExit) as e_info:
args = args_and_configs.get_arguments(command_line_args)
def test_default_ignore_whitespace_flag():
command_line_args = [
'--reference-input', 'tests/data/i2b2_2016_track-1_reference',
'--test-input', 'tests/data/i2b2_2016_track-1_test'
]
args = args_and_configs.get_arguments(command_line_args)
assert args.ignore_whitespace == True
def test_score_missing_test_files_usage():
command_line_args = [
'--reference-input', 'tests/data/i2b2_2016_track-1_reference',
'--test-input', 'tests/data/i2b2_2016_track-1_test',
'--score-missing-files'
]
args = args_and_configs.get_arguments(command_line_args)
assert args.skip_missing_files == False
def test_print_counts_ref_only():
command_line_args = [
'--print-counts', '--no-metrics', '--no-confusion-matrix',
'--reference-input', 'tests/data/i2b2_2016_track-1_reference'
]
args = args_and_configs.get_arguments(command_line_args)
assert args.print_counts
assert not args.print_metrics
assert not args.print_confusion_matrix
assert args.reference_input is not None
assert args.test_input is None
def initialize_for_track1():
command_line_args = [
'--reference-input', 'tests/data/n2c2_2018_track-1_reference',
'--test-input', 'tests/data/n2c2_2018_track-1_test', '--empty-value',
'0.0'
]
args = args_and_configs.get_arguments(command_line_args)
## TODO - we are usurping the test for this done within init_args to make
## the test more protable, for now.
args.empty_value = float(args.empty_value)
file_mapping = {
'103.xml': '103.xml',
'203.xml': '203.xml',
'303.xml': '303.xml'
}
return (args, file_mapping)
def init_args():
##
args = args_and_configs.get_arguments(sys.argv[1:])
## Set up logging
if args.verbose:
log.basicConfig(format="%(levelname)s: %(message)s", level=log.DEBUG)
log.info("Verbose output.")
log.debug("{}".format(args))
else:
log.basicConfig(format="%(levelname)s: %(message)s")
## Configure progressbar peformance
if (args.progressbar_output == 'none'):
args.progressbar_disabled = True
args.progressbar_file = None
else:
args.progressbar_disabled = False
if (args.progressbar_output == 'stderr'):
args.progressbar_file = sys.stderr
elif (args.progressbar_output == 'stdout'):
args.progressbar_file = sys.stdout
## F-score beta values are commonly set to 1, 2, and 0.5 but we
## want to support any values. It's easiest to do filtering at
## this point in the pipeline to standardize beta values and how
## they show up in the pipeline
if ('F' in args.metrics_list):
f_position = args.metrics_list.index('F')
args.metrics_list.pop(f_position)
if (len(args.f_beta_values) == 0):
log.warning(
'F was included in the list of metrics to calculate but no beta values were provided (--f-beta-values <betas>)'
)
else:
## Reverse the list so that they get inserted into the metrics_list
## in the proper order
args.f_beta_values.reverse()
for beta in args.f_beta_values:
if ('F{}'.format(beta) not in args.metrics_list):
args.metrics_list.insert(f_position, 'F{}'.format(beta))
else:
if (len(args.f_beta_values) > 0):
log.warning(
'F beta values were provided but "F" was not included in the list of metrics to calculate (--f-beta-values <betas>)'
)
args.f_beta_values = []
for common_beta in ['1', '2', '0.5']:
if ('F{}'.format(common_beta) in args.metrics_list):
if (common_beta not in args.f_beta_values):
args.f_beta_values.append(common_beta)
## The command line parameters are always initially cast as strings.
## That works fine for some empty values. Sometimes we want to use
## 0 (int) or 0.0 (float) or -1 as empty values. In this case,
## it's best to cast the string to the appropriate numerical
## type for formatting later.
if (args.empty_value is not None and args.empty_value != ''):
try:
args.empty_value = int(args.empty_value)
except ValueError:
log.debug('Default empty_value is not an int')
try:
args.empty_value = float(args.empty_value)
except ValueError:
log.debug('Default empty_value is not a float')
## Resolve conflicts between --ignore-whitespace, --heed-whitespace,
## and --ignore-regex flags. Essentially, if we set something in
## skip_chars, use that. Otherwise, if we tripped --ignore_whitespace
## then set skip_chars accordingly
if (args.ignore_whitespace and args.skip_chars == None):
args.skip_chars = r'[\s]'
## lstrip hack added to handle prefixes and suffixes with dashes
## https://stackoverflow.com/questions/16174992/cant-get-argparse-to-read-quoted-string-with-dashes-in-it
args.file_prefix = args.file_prefix.lstrip()
args.file_suffix[0] = args.file_suffix[0].lstrip()
if (len(args.file_suffix) == 2):
args.file_suffix[1] = args.file_suffix[1].lstrip()
## Initialize the list of annotation attributes to score
args.attributes_list = []
args.scorable_attributes = []
if (isinstance(args.attributes_string, str)):
for attribute_key in args.attributes_string.split(','):
## Strip off any extra whitespace before processing
attribute_key = attribute_key.strip()
attribute_kernel = attribute_key.split('/')
last = len(attribute_kernel) - 1
args.attributes_list.append(
[attribute_kernel[0], attribute_kernel[last]])
## Initialize the list of normalization engines to score
args.normalization_list = []
args.scorable_engines = []
args.normalization_synonyms = {}
if (isinstance(args.normalization_string, str)):
for normalization_key in args.normalization_string.split(','):
## Strip off any extra whitespace before processing
normalization_key = normalization_key.strip()
normalization_kernel = normalization_key.split('/')
last = len(normalization_kernel) - 1
args.normalization_list.append(
[normalization_kernel[0], normalization_kernel[last]])
## Only bother to load the normalization_file if the --score-normalization
## flag was used
args.normalization_synonyms = \
args_and_configs.process_normalization_file( args.normalization_file )
## Initialize the corpuse settings, values, and metrics file
## if it was provided at the command line
if (args.corpus_out):
## Clean out any previous corpus dictionary, in case it exists from
## an old run
with open(args.corpus_out, 'w') as fp:
json.dump({}, fp, sort_keys=True, indent=4)
## Add a few important arguments
scoring_metrics.update_output_dictionary(args.corpus_out, ['args'], [
'reference_config', 'reference_input', 'reference_out',
'test_config', 'test_input', 'test_out', 'score_key', 'fuzzy_flags'
], [
args.reference_config, args.reference_input, args.reference_out,
args.test_config, args.test_input, args.test_out, args.score_key,
args.fuzzy_flags
])
return args
def test_print_counts_neither_ref_nor_test():
command_line_args = [
'--print-counts', '--no-metrics', '--no-confusion-matrix'
]
with pytest.raises(SystemExit) as e_info:
args = args_and_configs.get_arguments(command_line_args)
def test_required_input_flags_test_only():
command_line_args = ['--test-input', 'tests/data/i2b2_2016_track-1_test']
with pytest.raises(SystemExit) as e_info:
args = args_and_configs.get_arguments(command_line_args)
