This package provides an API to interact with OSCAR (https://oscar.grycap.net) clusters and services. It is available on Pypi with the name oscar-python.
- Sample code that creates a client and gets information about the cluster
from oscar_python.client import Client
client = Client("cluster-id","https://cluster-endpoint", "username", "password", True)
# get the cluster information
try:
info = client.get_cluster_info()
print(info.text)
except Exception as err:
print("Failed with: ", err)- Sample code to create a simple service with the cowsay example and make a synchronous invocation.
from oscar_python.client import Client
client = Client("cluster-id","https://cluster-endpoint", "username", "password", True)
try:
client.create_service("/absolute_path/cowsay.yaml")
res = client.run_service("cowsay", '{"message": "Hi there"}')
if res.status_code == 200:
print(res.text)
except Exception as err:
print("Failed with: ", err)get_cluster_info
# get the cluster information
info = client.get_cluster_info() # returns an HTTP response or an HTTPErrorget_cluster_config
# get the cluster config
config = client.get_cluster_config() # returns an http response or an HTTPErrorget_service
# get the definition of a service
service = client.get_service("service_name") # returns an http response or an HTTPErrorlist_services
# get a list of all the services deployed
services = client.list_services() # returns an http response or an HTTPErrorNote : Both
path_to_fdland the script path inside the fdl must be absolute.
create_service
# create a service
err = client.create_service("path_to_fdl") # returns nothing if the service is created or an error if something goes wrongupdate_service
# update a service
err = client.update_service("service_name","path_to_fdl") # returns nothing if the service is created or an error if something goes wrongremove_service
# remove a service
response = client.remove_service("service_name") # returns an http responserun_service
The input parameter may not be passed if the function doesn't require input.
# make a synchronous execution
response = client.run_service("service_name", input="input") # returns an http responseget_job_logs
# get logs of a job
logs = client.get_job_logs("service_name", "job_id") # returns an http responselist_jobs
# get a list of jobs in a service
log_list = client.list_jobs("service_name") # returns an http responseremove_job
# remove a job of a service
response = client.remove_job("service_name", "job_id") # returns an http responseremove_all_jobs
# remove all jobs in a service
response = client.remove_all_jobs("service_name") # returns an http responseYou can create a storage object to operate over the different storage providers defined on a service with the method create_storage_client as follows:
storage_service = client.create_storage_client("service_name") # returns a storage objectNote : The
storage_providerparameter on the storage methods follows the format:["storage_provider_type"].["storage_provider_name"]wherestorage_provider_typeis one of the suported storage providers (minIO, S3, Onedata or webdav) andstorage_provider_nameis the identifier (ex: minio.default)
list_files_from_path
This method returns a JSON with the info except for OneData, which returns an HTTP response.
# get a list of the files of one of the service storage provider
files = storage_service.list_files_from_path("storage_provider") # returns jsonupload_file
# upload a file from a local path to a remote path
response = storage_service.upload_file("storage_provider", "local_path", "remote_path")download_file
# download a file from a remote path to a local path
response = storage_service.download_file("storage_provider", "local_path", "remote_path")