arangopipe.arangopipe_storage package¶
Submodules¶
arangopipe.arangopipe_storage.arangopipe_admin_api module¶
Created on Thu Apr 25 09:30:33 2019
@author: Rajiv Sambasivan
-
class
arangopipe.arangopipe_storage.arangopipe_admin_api.ArangoPipeAdmin(reuse_connection=True, config=None, persist_conn=True, client_url=None)¶ Bases:
object-
add_edge_definition_to_arangopipe(edge_col_name, edge_name, from_vertex_name, to_vertex_name)¶
-
add_edges_to_arangopipe(edge_col_name, from_vertex_list, to_vertex_list)¶
-
add_vertex_to_arangopipe(vertex_to_create)¶
-
check_repeated_creation(api_data)¶
-
create_config()¶
-
create_db(db_srv_host, db_srv_port, db_serv_name, db_end_point, db_dbName, db_user_name, db_password, db_conn_protocol)¶
-
create_enterprise_ml_graph(db_replication_factor)¶
-
delete_all_databases(preserve=['arangopipe', 'facebook_db', 'fb_node2vec_db', 'node2vecdb', '_system'])¶
-
delete_arangomldb()¶
-
delete_database(db_to_delete)¶
-
get_config()¶
-
has_edge(edge_name)¶
-
has_vertex(vertex_name)¶
-
register_deployment(dep_tag)¶
-
register_project(p)¶
-
remove_edge_definition_from_arangopipe(edge_name, purge=True)¶
-
remove_vertex_from_arangopipe(vertex_to_remove, purge=True)¶
-
set_connection_params(config)¶
-
arangopipe.arangopipe_storage.arangopipe_api module¶
Created on Sat Apr 13 08:35:58 2019
@author: Rajiv Sambasivan
-
class
arangopipe.arangopipe_storage.arangopipe_api.ArangoPipe(config)¶ Bases:
objectAn instance of ArangoPipe is meant to be used to log a run of a pipeline execution. To use it: (1) Create a ArangoPipe object (2) Register your dataset with ArangoPipe (3) Register your featureset with ArangoPipe (4) Register you model with ArangoPipe
-
find_entity(attrib_name, attrib_value, asset_type)¶
-
get_collection_from_id(id_str)¶
-
get_config()¶
-
heart_beat()¶
-
init_graph()¶ Initialize a graph when an instance of ArangoPipe is provisioned.
-
insert_into_edge_type(edge_name, from_vdoc, to_vdoc, document=None)¶
-
insert_into_vertex_type(vertex_type_name, document)¶
-
is_valid_id_str(id_str)¶
-
link_entities(src_id, dest_id)¶
-
log_run(ri)¶ Log a run. Logging a run requires specifying a dataset, featureset and a model against which this run is recored. A run records model parameters and model performance. The run object is probably most useful for the analysis of model performance with respect to a featureset, model hyper-parameters and a dataset.
-
log_serving_perf(sp, dep_tag, userid='authorized user')¶ Log serving performance against a deployed model. The user making the request needs to be authorized to log this performance update. A serving performance vertex is created and is linked with its deployment vertex
-
lookup_dataset(dataset_name)¶ Return a dataset identifier given a name. This can be used to get the dataset id that is used to log run information associated with execution of the pipeline.
-
lookup_entity(asset_name, asset_type)¶
-
lookup_entity_by_id(entity_id)¶
-
lookup_featureset(feature_set_name)¶ Return a featureset identifier given a name. This can be used to get the featureset id that is used to log run information associated with execution of the pipeline.
-
lookup_model(model_name)¶ Return a model identifier given a name. This can be used to get the model id that is used to log run information associated with execution of the pipeline.
-
lookup_modelparams(tag_value)¶ Return a model parameter result given a tag.
-
lookup_modelperf(tag_value)¶ Return a model dev performance given a tag.
-
register_dataset(ds_info, user_id='authorized_user')¶ Register a dataset. The operation requires specifying a user id. If the user id is permitted to register a dataset, then the registration proceeds, otherwise an unauthorized operation is indicated.
-
register_featureset(fs_info, dataset_id, user_id='authorized_user')¶ Register a featureset. ManagedServiceConnParamThe operation requires specifying a user id. If the user id is permitted to register a featureset, then the registration proceeds, otherwise an unauthorized operation is indicated.
-
register_model(mi, user_id='authorized_user', project='Wine-Quality-Regression-Modelling')¶ Register a model. The operation requires specifying a user id. If the user id is permitted to register a model, then the registration proceeds, otherwise an unauthorized operation is indicated.
-
arangopipe.arangopipe_storage.arangopipe_config module¶
Created on Fri Jun 14 09:17:50 2019
@author: Rajiv Sambasivan
arangopipe.arangopipe_storage.custom_http_client module¶
Spyder Editor @author: Rajiv Sambasivan
-
class
arangopipe.arangopipe_storage.custom_http_client.CustomHTTPClient(username, password)¶ Bases:
arango.http.HTTPClientMy custom HTTP client with cool features.
-
create_session(host)¶ Return a new requests session given the host URL.
This method must be overridden by the user.
- Parameters
host (str | unicode) – ArangoDB host URL.
- Returns
Requests session object.
- Return type
requests.Session
-
send_request(session, method, url, params=None, data=None, headers=None, auth=None)¶ Send an HTTP request.
This method must be overridden by the user.
- Parameters
session (requests.Session) – Requests session object.
method (str | unicode) – HTTP method in lowercase (e.g. “post”).
url (str | unicode) – Request URL.
headers (dict) – Request headers.
params (dict) – URL (query) parameters.
data (str | unicode | bool | int | list | dict) – Request payload.
auth (tuple) – Username and password.
- Returns
HTTP response.
- Return type
arango.response.Response
-
arangopipe.arangopipe_storage.managed_service_conn_parameters module¶
Created on Thu Dec 5 09:05:19 2019
@author: Rajiv Sambasivan
-
class
arangopipe.arangopipe_storage.managed_service_conn_parameters.ManagedServiceConnParam¶ Bases:
object-
property
DB_CONN_PROTOCOL¶
-
property
DB_NAME¶
-
property
DB_NOTIFICATION_EMAIL¶
-
property
DB_PASSWORD¶
-
property
DB_REPLICATION_FACTOR¶
-
property
DB_ROOT_USER¶
-
property
DB_ROOT_USER_PASSWORD¶
-
property
DB_SERVICE_END_POINT¶
-
property
DB_SERVICE_HOST¶
-
property
DB_SERVICE_NAME¶
-
property
DB_SERVICE_PORT¶
-
property
DB_USER_NAME¶
-
property