def goh_behave(feature_dirs=None, step_dirs=None, test_artifact_dir=None,
listeners=None, dry_run=False, webdriver_url=None,
webdriver_processor=None, tags=None, show_skipped=True, config_file=None,
test_config=None):
"""
runs behave
:param feature_dirs: list of paths to feature files to run, or a single path. Defaults to singleton list of "features"
:type feature_dirs: list
:param step_dirs: list of paths to load steps from. the steps paths will be
searched recursssssively
:type step_dirs: list
:param test_artifact_dir: path to where to store test artifacts. if None, no test
artifacts will be written. note that setting this to None will prevent
screenshots from automatically being taken after each step
:param listeners: list of Listener objects to call for diff pts of the test
:type listeners: list
:param dry_run: if True, behave will just check if all steps are defined
:param webdriver_url: optional. webdriver node/grid url to hit to execute
tests
:param webdriver_processor: provides the ability to process things like
capabilities before they're actually used
:param config_file: a configuration file, formatted as JSON data, that contains
all of the other parameters listed here (except for listeners or webdriver_processor)
:param test_config: a configuration that is a dictionary of the parameters. If a config_file is
also passed, the values in the config_file will be written to the test_config as well. Any existing
keys will have the values overwritten by the value in the config_file
:return: True if the tests passed, else False
:rtype: bool
:raise ParserError: when a feature file couldnt be parsed
:raise FileNotFoundError: when a feature file couldnt be found
:raise InvalidFileLocationError: when a feature path was bad
:raise InvalidFilenameError: when a feature file name was bad
:raise UndefinedStepsError: if some steps were undefined
"""
if config_file:
try:
with open(config_file) as config:
try:
json_config = json.load(config)
except ValueError as e:
raise ValueError(u'Could not parse {} config file as JSON. See sample.config for an example config file. {}'.format(config_file, e))
if test_config:
test_config.update(json_config)
else:
test_config = json_config
except EnvironmentError as e:
raise IOError(u'Could not open the {} config file. See sample.config for an example config file. {}'.format(config_file, e))
if test_config:
log.debug(u'Using test_config:')
for key in test_config:
log.debug(u' {}: {}'.format(key, test_config[key]))
if u'feature_dirs' in test_config and not feature_dirs:
feature_dirs = test_config[u'feature_dirs']
if u'step_dirs' in test_config and not step_dirs:
step_dirs = test_config[u'step_dirs']
if u'test_artifact_dir' in test_config and not test_artifact_dir:
test_artifact_dir = test_config[u'test_artifact_dir']
if u'dry_run' in test_config:
dry_run = test_config[u'dry_run']
if u'webdriver_url' in test_config and not webdriver_url:
webdriver_url = test_config[u'webdriver_url']
if u'tags' in test_config and not tags:
tags = test_config[u'tags']
if u'show_skipped' in test_config:
show_skipped = test_config[u'show_skipped']
if not feature_dirs:
feature_dirs = ["features"]
if isinstance(feature_dirs, basestring):
feature_dirs = [feature_dirs]
args = [u'']
# first run in dry run mode to catch any undefined steps
# if the user specifies dry run mode, then don't do this, since it'll be
# done anyway
# but we're running it always anyway. so what's the difference? if user
# specifies dry run mode, they might also give listeners, etc.
# auto dry run mode is meant to happen silently
if not dry_run:
config = Configuration([u'--dry-run'])
config.format = []
_run_behave(feature_dirs, config, step_dirs=step_dirs)
# output test artifacts
if test_artifact_dir:
args.append(u'--junit')
args.append(u'--junit-directory')
args.append(test_artifact_dir)
if dry_run:
args.append(u'--dry-run')
# setup config for behave's runner
config = Configuration(args)
config.format = [
# outputs pretty output in to stdout while tests are running
u'pretty'
]
# Set the tags if there are any
if tags:
log.debug(u'Running Scenarios with Tag(s): {}'.format(tags))
if isinstance(tags, list):
log.debug(u'Running Scenarios with Tag List')
config.tags = TagExpression(tags)
else:
config.tags = TagExpression(tags.split())
config.show_skipped = show_skipped
if not show_skipped:
log.debug(u'Not showing skipped scenarios.')
# only add html reporter if the test artifacts are enabled
if test_artifact_dir:
config.reporters.append(HtmlReporter(config, test_artifact_dir))
try:
return _run_behave(
feature_dirs,
config,
step_dirs=step_dirs,
webdriver_url=webdriver_url,
webdriver_processor=webdriver_processor,
listeners=listeners,
test_config = test_config
)
except ConfigError as e:
# since we control the configuration file, we shouldn't expose this to
# the user, since there's nothing they can do
log.error(e)
return False
