def register_outputs(job):
'''
This adds outputs from the workflow to the list of Resources owned by the client
This way they are able to download files produced by the workflow
'''
config_path = os.path.join(THIS_DIR, 'wdl_job_config.cfg')
config_dict = utils.load_config(config_path)
# pull together the components of the request to the Cromwell server
outputs_endpoint = config_dict['outputs_endpoint']
outputs_url_template = Template(settings.CROMWELL_SERVER_URL +
outputs_endpoint)
outputs_url = outputs_url_template.render({'job_id': job.job_id})
try:
response = requests.get(outputs_url)
response_json = json.loads(response.text)
if (response.status_code == 404) or (response.status_code
== 400) or (response.status_code
== 500):
job.project.status = 'Analysis completed. Error encountered when collecting final outputs.'
job.project.error = True
job.project.save()
handle_exception(
None, 'Query for job failed with message: %s' %
response_json['message'])
else: # the request itself was OK
outputs = response_json['outputs']
output_filepath_list = parse_outputs(outputs)
environment = settings.CONFIG_PARAMS['cloud_environment']
for p in output_filepath_list:
size_in_bytes = get_resource_size(p)
full_destination_with_prefix = move_resource_to_user_bucket(
job, p)
# add the Resource to the database:
r = Resource(source=environment,
path=full_destination_with_prefix,
name=os.path.basename(p),
owner=job.project.owner,
size=size_in_bytes)
r.save()
# add a ProjectResource to the database, so we can tie the Resource created above with the analysis project:
apr = AnalysisProjectResource(analysis_project=job.project,
resource=r)
apr.save()
except Exception as ex:
print(
'An exception was raised when requesting job outputs from cromwell server'
)
print(ex)
message = 'An exception occurred when trying to query outputs from Cromwell. \n'
message += 'Job ID was: %s' % job.job_id
message += 'Project ID was: %s' % job.project.analysis_uuid
message += str(ex)
raise JobOutputsException(message)
def __init__(self, upload_data):
#instantiate the wrapped classes:
self.uploader = self.uploader_cls(upload_data)
self.launcher = self.launcher_cls()
# get the config params for the uploader:
uploader_cfg = self.uploader_cls.get_config(self.config_file)
additional_cfg = utils.load_config(self.config_file, self.config_keys)
uploader_cfg.update(additional_cfg)
self.config_params = uploader_cfg
def main():
models = {"baseline": BaselineModel, "second_order": SecondOrderModel}
config = load_config()
for part in ["nouns", "verbs"]:
for phase in ["public", "private"]:
with open(config[f"{phase}_test_{part}_path"], 'r', encoding='utf-8') as f:
test_data = f.read().split("\n")[:-1]
baseline = models[config["model"]](params=config, part=part, phase=phase)
print("Model loaded")
results = baseline.predict_hypernyms(list(test_data))
save_to_file(results, config[f"{phase}_output_{part}_path"], baseline.ruwordnet)
def post(self, request, *args, **kwargs):
if not request.user.is_staff:
return HttpResponseForbidden()
try:
payload = request.POST
analysis_uuid = payload['cnap_uuid']
analysis_project = AnalysisProject.objects.get(
analysis_uuid=analysis_uuid)
except analysis.models.AnalysisProject.DoesNotExist as ex:
return HttpResponseBadRequest(
'Could not find a project with that UUID')
# now have a project, but to kill the job, we need a SubmittedJob
try:
sj = SubmittedJob.objects.get(project=analysis_project)
cromwell_id = sj.job_id
# send Cromwell a message to abort the job:
# read config to get the names/locations/parameters for job submission
config_path = os.path.join(THIS_DIR, 'wdl_job_config.cfg')
config_dict = utils.load_config(config_path)
# pull together the components of the POST request to the Cromwell server
abort_endpoint_str = config_dict['abort_endpoint']
abort_url_template = Template(settings.CROMWELL_SERVER_URL +
abort_endpoint_str)
abort_url = abort_url_template.render({'job_id': cromwell_id})
r = requests.post(abort_url)
if r.status_code != 200:
return HttpResponseBadRequest(
'Did not return a proper response code from Cromwell. Reason was: %s'
% r.text)
else:
# reset the project attributes
analysis_project.error = False
analysis_project.completed = False
analysis_project.started = False
analysis_project.message = ''
analysis_project.status = ''
analysis_project.save()
# finally, delete the submitted job
sj.delete()
return JsonResponse({'message': 'Job has been aborted.'})
except analysis.models.SubmittedJob.DoesNotExist:
return HttpResponseBadRequest(
'Could not find a running job for project %s' %
analysis_project.analysis_uuid)
def get_config(cls, config_filepath):
return utils.load_config(config_filepath, cls.config_keys)
with open(output_path, 'w', encoding='utf-8') as w:
w.write(f"{vectors.shape[0]} {vectors.shape[1]}\n")
for word, vector in zip(words, vectors):
vector_line = " ".join(map(str, vector))
w.write(f"{word.upper()} {vector_line}\n")
def process_data(input_file, output_file):
with open(input_file, 'r', encoding='utf-8') as f:
dataset = f.read().lower().split("\n")[:-1]
w2v_vec.vectorize_data(dataset, output_file)
if __name__ == '__main__':
from helpers.utils import load_config
config = load_config()
w2v_vec = wiki2vecVectorizer(config["vectorizer_path"])
ruwordnet = RuWordnet(db_path=config["db_path"], ruwordnet_path=config["ruwordnet_path"], with_lemmas=False)
noun_synsets = defaultdict(list)
verb_synsets = defaultdict(list)
for sense_id, synset_id, text in ruwordnet.get_all_senses():
if synset_id.endswith("N"):
noun_synsets[synset_id].append(text.lower())
elif synset_id.endswith("V"):
verb_synsets[synset_id].append(text.lower())
w2v_vec.vectorize_ruwordnet(noun_synsets, "models/vectors/ruwordnet_nouns.txt")
w2v_vec.vectorize_ruwordnet(verb_synsets, "models/vectors/ruwordnet_verbs.txt")
process_data("../data/public_test/verbs_public.tsv", "models/vectors/verbs_public.txt")
process_data("../data/public_test/nouns_public.tsv", "models/vectors/nouns_public.txt")
def execute_wdl(analysis_project, staging_dir, run_precheck=False):
'''
This function performs the actual work of submitting the job
'''
# read config to get the names/locations/parameters for job submission
config_path = os.path.join(THIS_DIR, 'wdl_job_config.cfg')
config_dict = utils.load_config(config_path)
# the path of the input json file:
wdl_input_path = os.path.join(staging_dir, WDL_INPUTS)
# pull together the components of the POST request to the Cromwell server
submission_endpoint = config_dict['submit_endpoint']
submission_url = settings.CROMWELL_SERVER_URL + submission_endpoint
payload = {}
payload = {'workflowType': config_dict['workflow_type'], \
'workflowTypeVersion': config_dict['workflow_type_version']
}
# load the options file so we can fill-in the zones:
options_json = {}
current_zone = get_zone_as_string()
if current_zone:
options_json['default_runtime_attributes'] = {'zones': current_zone}
options_json_str = json.dumps(options_json)
options_io = io.BytesIO(options_json_str.encode('utf-8'))
files = {
'workflowOptions': options_io,
'workflowInputs': open(wdl_input_path, 'rb')
}
if run_precheck:
files['workflowSource'] = open(
os.path.join(staging_dir, settings.PRECHECK_WDL), 'rb')
else:
files['workflowSource'] = open(
os.path.join(staging_dir, settings.MAIN_WDL), 'rb')
zip_archive = os.path.join(staging_dir, ZIPNAME)
if os.path.exists(zip_archive):
files['workflowDependencies'] = open(zip_archive, 'rb')
# start the job:
try:
response = requests.post(submission_url, data=payload, files=files)
except Exception as ex:
print('An exception was raised when requesting cromwell server:')
print(ex)
message = 'An exception occurred when trying to submit a job to Cromwell. \n'
message += 'Project ID was: %s' % str(analysis_project.analysis_uuid)
message += str(ex)
analysis_project.status = '''
Error on job submission. An administrator has been automatically notified of this error.
Thank you for your patience.
'''
analysis_project.error = True
analysis_project.save()
handle_exception(ex, message=message)
raise ex
response_json = json.loads(response.text)
if response.status_code == 201:
if response_json['status'] == 'Submitted':
job_id = response_json['id']
if run_precheck:
job_status = 'Checking input data...'
else:
job_status = 'Job submitted...'
job = SubmittedJob(project=analysis_project,
job_id=job_id,
job_status=job_status,
job_staging_dir=staging_dir,
is_precheck=run_precheck)
job.save()
# update the project also:
analysis_project.started = True # should already be set
analysis_project.start_time = datetime.datetime.now()
analysis_project.status = job_status
analysis_project.save()
else:
# In case we get other types of responses, inform the admins:
message = 'Job was submitted, but received an unexpected response from Cromwell:\n'
message += response.text
handle_exception(None, message=message)
else:
message = 'Did not submit job-- status code was %d, and response text was: %s' % (
response.status_code, response.text)
analysis_project.status = '''
Error on job submission. An administrator has been automatically notified of this error.
Thank you for your patience.
'''
analysis_project.error = True
analysis_project.save()
handle_exception(None, message=message)
def check_job():
'''
Used for pinging the cromwell server to check job status
'''
terminal_actions = {'Succeeded': handle_success, 'Failed': handle_failure}
precheck_terminal_actions = {
'Succeeded': handle_precheck_success,
'Failed': handle_precheck_failure
}
other_states = ['Submitted', 'Running']
config_path = os.path.join(THIS_DIR, 'wdl_job_config.cfg')
config_dict = utils.load_config(config_path)
# pull together the components of the request to the Cromwell server
query_endpoint = config_dict['query_status_endpoint']
query_url_template = Template(settings.CROMWELL_SERVER_URL +
query_endpoint)
# get the job IDs for active jobs:
active_job_set = SubmittedJob.objects.all()
print('%d active jobs found.' % len(active_job_set))
for job in active_job_set:
query_url = query_url_template.render({'job_id': job.job_id})
try:
response = requests.get(query_url)
response_json = json.loads(response.text)
if (response.status_code
== 404) or (response.status_code
== 400) or (response.status_code == 500):
handle_exception(
None, 'Query for job failed with message: %s' %
response_json['message'])
else: # the request itself was OK
status = response_json['status']
# if the job was in one of the finished states, execute some specific logic
if status in terminal_actions.keys():
if job.is_precheck:
precheck_terminal_actions[status](
job
) # call the function to execute the logic for this end-state
else:
terminal_actions[status](
job
) # call the function to execute the logic for this end-state
elif status in other_states:
# any custom behavior for unfinished tasks
# can be handled here if desired
# update the job status in the database
job.job_status = status
job.save()
project = job.project
project.status = status
project.save()
else:
# has some status we do not recognize
message = 'When querying for status of job ID: %s, ' % job.job_id
message += 'received an unrecognized response: %s' % response.text
job.job_status = 'Unknown'
job.save()
try:
warnings_sent = Warning.objects.get(job=job)
print(
'When querying cromwell for job status, received unrecognized status. Notification suppressed'
)
except analysis.models.Warning.DoesNotExist:
handle_exception(None, message=message)
# add a 'Warning' object in the database so that we don't
# overwhelm the admin email boxes.
warn = Warning(message=message, job=job)
warn.save()
except Exception as ex:
print(
'An exception was raised when requesting job status from cromwell server'
)
print(ex)
message = 'An exception occurred when trying to query a job. \n'
message += 'Job ID was: %s' % job.job_id
message += 'Project ID was: %s' % job.project.analysis_uuid
message += str(ex)
try:
warnings_sent = Warning.objects.get(job=job)
print(
'Error when querying cromwell for job status. Notification suppressed'
)
except analysis.models.Warning.DoesNotExist:
handle_exception(ex, message=message)
# add a 'Warning' object in the database so that we don't
# overwhelm the admin email boxes.
warn = Warning(message=message, job=job)
warn.save()
raise ex
def handle_precheck_failure(job):
'''
If a pre-check job failed, something was wrong with the inputs.
We query the cromwell metadata to get the error so the user can correct it
'''
config_path = os.path.join(THIS_DIR, 'wdl_job_config.cfg')
config_dict = utils.load_config(config_path)
# pull together the components of the request to the Cromwell server
metadata_endpoint = config_dict['metadata_endpoint']
metadata_url_template = Template(settings.CROMWELL_SERVER_URL +
metadata_endpoint)
metadata_url = metadata_url_template.render({'job_id': job.job_id})
try:
response = requests.get(metadata_url)
response_json = response.json()
stderr_file_list = walk_response('', response_json, 'stderr')
error_obj_list = log_client_errors(job, stderr_file_list)
# update the AnalysisProject instance:
project = job.project
project.completed = False
project.success = False
project.error = True
project.status = 'Issue encountered with inputs.'
project.message = ''
project.finish_time = datetime.datetime.now()
project.save()
# inform the client of this problem so they can fix it (if allowed):
email_address = project.owner.email
current_site = Site.objects.get_current()
domain = current_site.domain
project_url = reverse('analysis-project-execute',
args=[
project.analysis_uuid,
])
url = 'https://%s%s' % (domain, project_url)
context = {'site': url, 'user_email': email_address}
if project.restart_allowed:
email_template_path = 'email_templates/analysis_fail_with_recovery.html'
email_plaintxt_path = 'email_templates/analysis_fail_with_recovery.txt'
email_subject = 'email_templates/analysis_fail_subject.txt'
else:
email_template_path = 'email_templates/analysis_fail.html'
email_plaintxt_path = 'email_templates/analysis_fail.txt'
email_subject = 'email_templates/analysis_fail_subject.txt'
email_template = get_jinja_template(email_template_path)
email_html = email_template.render(context)
email_plaintxt_template = get_jinja_template(email_plaintxt_path)
email_plaintxt = email_plaintxt_template.render(context)
email_subject = open(email_subject).readline().strip()
send_email(email_plaintxt, email_html, email_address, email_subject)
if not project.restart_allowed:
# a project that had a pre-check failed, but a restart was NOT allowed.
# need to inform admins:
message = 'Job (%s) experienced failure during pre-check. No restart was allowed. Staging dir was %s' % (
job.job_id, job.job_staging_dir)
subject = 'Cromwell job failure on pre-check'
notify_admins(message, subject)
# delete the failed job:
job.delete()
except Exception as ex:
print('An exception was raised when requesting metadata '
'from cromwell server following a pre-check failure')
print(ex)
message = 'An exception occurred when trying to query metadata. \n'
message += 'Job ID was: %s' % job.job_id
message += 'Project ID was: %s' % job.project.analysis_uuid
message += str(ex)
try:
warnings_sent = Warning.objects.get(job=job)
print(
'Error when querying cromwell for metadata. Notification suppressed'
)
except analysis.models.Warning.DoesNotExist:
handle_exception(ex, message=message)
# add a 'Warning' object in the database so that we don't
# overwhelm the admin email boxes.
warn = Warning(message=message, job=job)
warn.save()
raise ex
elif val == 'False' or val == 'false':
CONFIG_PARAMS[key] = False
# using the value of EXPIRATION_PERIOD_DAYS from the config, set a timedelta:
# This logic could be altered as desired:
EXPIRATION_PERIOD = datetime.timedelta(
days=int(CONFIG_PARAMS['expiration_period_days']))
# These are the days on which clients are reminded of pending deletion of Resources:
EXPIRATION_REMINDER_DAYS = [
int(x.strip())
for x in CONFIG_PARAMS['expiration_reminder_days'].split(',')
]
additional_sections = [GOOGLE_DRIVE, DROPBOX, GOOGLE]
LIVE_TEST_CONFIG_PARAMS = utils.load_config(
os.path.join(CONFIG_DIR, 'live_tests.cfg'), additional_sections)
# Configuration for upload providers and compute environments:
UPLOADER_CONFIG = {
'CONFIG_PATH': os.path.join(CONFIG_DIR, 'uploaders.cfg'),
# for each item in the following dictionary, there needs to be a section
# header in the config file located at UPLOADER_CONFIG.CONFIG_PATH
'UPLOAD_SOURCES': [DROPBOX, GOOGLE_DRIVE]
}
DOWNLOADER_CONFIG = {
'CONFIG_PATH': os.path.join(CONFIG_DIR, 'downloaders.cfg'),
# for each item in the following dictionary, there needs to be a section