Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

grycap/oscar_python

Repository files navigation

Python OSCAR client

Tests PyPI

This package provides a client to interact with OSCAR (https://oscar.grycap.net) clusters and services. It is available on Pypi with the name oscar-python.

Contents

Client

Initialize a client with basic authentication

 options_basic_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'user':'username',
 'password':'password',
 'ssl':'True'}
 client = Client(options = options_basic_auth)

Initialize a client OIDC authentication

If you want to use OIDC tokens to authenticate with EGI Check-In, you can use the OIDC Agent to create an account configuration for the EGI issuer (https://aai.egi.eu/auth/realms/egi/) and then initialize the client specifying the shortname of your account like follows.

 options_oidc_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'shortname':'oidc-agent-shortname',
 'ssl':'True'}
 
 client = Client(options = options_oidc_auth)

If you already have a valid token, you can use the parameter oidc_token instead.

 options_oidc_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'oidc_token':'token',
 'ssl':'True'}
 
 client = Client(options = options_oidc_auth)

An example of using a generated token is if you want to use EGI Notebooks. Since you can't use oidc-agent on the Notebook, you can make use of the generated token that EGI provides on path /var/run/secrets/egi.eu/access_token.

If you have a valid refresh token (long live token), you can use the parameter refresh_token instead.

 options_oidc_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'refresh_token':'token',
 'ssl':'True'}
 
 client = Client(options = options_oidc_auth)

You can get a refresh token from EGI Check-In using the Token Portal.

In case of using other OIDC provider you must provide two additional parameters token_endpoint and scopes:

 options_oidc_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'refresh_token':'token',
 'scopes': ["openid", "profile", "email"],
 'token_endpoint': "http://issuer.com/token",
 'client_id': "your_client_id"
 'ssl':'True'}
 
 client = Client(options = options_oidc_auth)

Sample usage

  • Sample code that creates a client and gets information about the cluster
from oscar_python.client import Client
options_basic_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'user':'username',
 'password':'password',
 'ssl':'True'}
client = Client(options = options)
# 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
options_basic_auth = {'cluster_id':'cluster-id',
 'endpoint':'https://cluster-endpoint',
 'user':'username',
 'password':'password',
 'ssl':'True'}
client = Client(options = options)
try:
 client.create_service("/absolute_path/cowsay.yaml")
 response = client.run_service("cowsay", input = '{"message": "Hi there"}') 
 if response.status_code == 200:
 print(response.text)
except Exception as err:
 print("Failed with: ", err)

Client methods

Cluster methods

get_cluster_info

# get the cluster information
info = client.get_cluster_info() # returns an HTTP response or an HTTPError

get_cluster_config

# get the cluster config
config = client.get_cluster_config() # returns an http response or an HTTPError

Service methods

get_service

# get the definition of a service 
service = client.get_service("service_name") # returns an http response or an HTTPError

list_services

# get a list of all the services deployed 
services = client.list_services() # returns an http response or an HTTPError

Note : Both path_to_fdl and the script path inside the fdl must be absolute.

create_service

# create a service 
err = client.create_service("path_to_fdl" | "JSON_definition") # returns nothing if the service is created or an error if something goes wrong

update_service

# update a service 
err = client.update_service("service_name","path_to_fdl" | "JSON_definition") # returns nothing if the service is created or an error if something goes wrong

remove_service

# remove a service 
response = client.remove_service("service_name") # returns an http response

run_service

input, output and timeout are optional parameters.

# make a synchronous execution 
response = client.run_service("service_name", input="input", output="out.png", timeout=100) # returns an http response
# make an asynchronous execution
response = client.run_service("service_name", input="input", async_call=True) # returns an http response

Logs methods

get_job_logs

# get logs of a job
logs = client.get_job_logs("service_name", "job_id") # returns an http response

list_jobs

# get a list of jobs in a service
log_list = client.list_jobs("service_name") # returns an http response
# to get more jobs use the page parameter
log_list = client.list_jobs("service_name",page="token_to_next_page") # returns an http response

remove_job

# remove a job of a service
response = client.remove_job("service_name", "job_id") # returns an http response

remove_all_jobs

# remove all jobs in a service
response = client.remove_all_jobs("service_name") # returns an http response

Storage usage

You can create a storage object to operate over the different storage providers defined on a service with the method create_storage_client. This constructor returns a storage object with methos to interact with the storage providers.

The default constructor, seen as follows, will create a provider to interact with the default MinIO instance through the user's credentials.

storage_service = client.create_storage_client() # returns a storage object

Additionally, if you need to interact with specific storage providers defined on a service, the constructor accepts a svc parameter where you can state the service name from which to search for additional credentials.

storage_service = client.create_storage_client("service_name") # returns a storage object

Note : The storage_provider parameter on the storage methods follows the format: ["storage_provider_type"].["storage_provider_name"] where storage_provider_type is one of the suported storage providers (minIO, S3, Onedata or webdav) and storage_provider_name is 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", "remote_path") # returns json

upload_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")

About

A Python client to interact with OSCAR clusters and services

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Contributors 7

AltStyle によって変換されたページ (->オリジナル) /