def test_calculate_job_input_hash():
"""Test calculate_job_input_hash."""
job_spec_1 = {"workflow_workspace": "test"}
workflow_json = {}
job_spec_2 = {}
assert calculate_job_input_hash(job_spec_1,
workflow_json) == calculate_job_input_hash(
job_spec_2, workflow_json)
def _update_job_cache(msg):
"""Update caching information for finished job."""
cached_job = Session.query(JobCache).filter_by(
job_id=msg['caching_info'].get('job_id')).first()
input_files = []
if cached_job:
file_access_times = calculate_file_access_time(
msg['caching_info'].get('workflow_workspace'))
for filename in cached_job.access_times:
if filename in file_access_times:
input_files.append(filename)
else:
return
cmd = msg['caching_info']['job_spec']['cmd']
# removes cd to workspace, to be refactored
clean_cmd = ';'.join(cmd.split(';')[1:])
msg['caching_info']['job_spec']['cmd'] = clean_cmd
if 'workflow_workspace' in msg['caching_info']['job_spec']:
del msg['caching_info']['job_spec']['workflow_workspace']
input_hash = calculate_job_input_hash(msg['caching_info']['job_spec'],
msg['caching_info']['workflow_json'])
workspace_hash = calculate_hash_of_dir(
msg['caching_info'].get('workflow_workspace'), input_files)
if workspace_hash == -1:
return
cached_job.parameters = input_hash
cached_job.result_path = msg['caching_info'].get('result_path')
cached_job.workspace_hash = workspace_hash
Session.add(cached_job)
def _update_job_cache(msg):
"""Update caching information for finished job."""
cached_job = (Session.query(JobCache).filter_by(
job_id=msg["caching_info"].get("job_id")).first())
input_files = []
if cached_job:
file_access_times = calculate_file_access_time(
msg["caching_info"].get("workflow_workspace"))
for filename in cached_job.access_times:
if filename in file_access_times:
input_files.append(filename)
else:
return
cmd = msg["caching_info"]["job_spec"]["cmd"]
# removes cd to workspace, to be refactored
clean_cmd = ";".join(cmd.split(";")[1:])
msg["caching_info"]["job_spec"]["cmd"] = clean_cmd
if "workflow_workspace" in msg["caching_info"]["job_spec"]:
del msg["caching_info"]["job_spec"]["workflow_workspace"]
input_hash = calculate_job_input_hash(msg["caching_info"]["job_spec"],
msg["caching_info"]["workflow_json"])
workspace_hash = calculate_hash_of_dir(
msg["caching_info"].get("workflow_workspace"), input_files)
if workspace_hash == -1:
return
cached_job.parameters = input_hash
cached_job.result_path = msg["caching_info"].get("result_path")
cached_job.workspace_hash = workspace_hash
Session.add(cached_job)
def job_is_cached(job_spec, workflow_json, workflow_workspace):
"""Check if job result exists in the cache."""
input_hash = calculate_job_input_hash(job_spec, workflow_json)
workspace_hash = calculate_hash_of_dir(workflow_workspace)
if workspace_hash == -1:
return None
cached_job = Session.query(JobCache).filter_by(
parameters=input_hash,
workspace_hash=workspace_hash).first()
if cached_job:
return {'result_path': cached_job.result_path,
'job_id': cached_job.job_id}
else:
return None
