REST API UDFs¶

Exasol MLflow Plugin contains a list of User Defined Scripts (UDFs) to access MLflow’s REST API and get the results as SQL tables in Exasol.

The UDFs require the plugin’s SLC to be deployed and activated, see Using a Script Language Container.

UDF `ARTIFACTS_LIST`¶

Calls MLflow REST API endpoint artifacts.

SELECT ARTIFACTS_LIST(
    connection_name, path
);

Input Column	SQL Type	Comment
`path`	`VARCHAR(2000000)`	recommended

Output Column	SQL Type
`path`	`VARCHAR(2000000)`
`is_dir`	`BOOLEAN`
`file_size`	`DECIMAL(18,0)`

UDF `EXPERIMENTS_SEARCH`¶

Calls MLflow REST API endpoint experiments/search.

SELECT EXPERIMENTS_SEARCH(
    connection_name, filter, view_type, order_by, max_results
);

Input Column	SQL Type	Comment
`filter`	`VARCHAR(2000000)`
`view_type`	`VARCHAR(2000000)`
`order_by`	`VARCHAR(2000000)`
`max_results`	`DECIMAL(18,0)`

Output Column	SQL Type
`experiment_id`	`VARCHAR(2000000)`
`name`	`VARCHAR(2000000)`
`artifact_location`	`VARCHAR(2000000)`
`lifecycle_stage`	`VARCHAR(2000000)`
`updated`	`TIMESTAMP(3)`
`created`	`TIMESTAMP(3)`
`effective_trace_archival_retention`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `GATEWAY_ENDPOINTS_LIST`¶

Calls MLflow REST API endpoint gateway/endpoints/list.

SELECT GATEWAY_ENDPOINTS_LIST(
    connection_name, provider, secret_id
);

Input Column	SQL Type	Comment
`provider`	`VARCHAR(2000000)`
`secret_id`	`VARCHAR(2000000)`

Output Column	SQL Type
`endpoint_id`	`VARCHAR(2000000)`
`name`	`VARCHAR(2000000)`
`created_at`	`TIMESTAMP(3)`
`last_updated_at`	`TIMESTAMP(3)`
`created_by`	`VARCHAR(2000000)`
`last_updated_by`	`VARCHAR(2000000)`
`routing_strategy`	`VARCHAR(2000000)`
`experiment_id`	`VARCHAR(2000000)`
`usage_tracking`	`BOOLEAN`
`fallback_strategy`	`VARCHAR(2000000)`
`fallback_max_attempts`	`DECIMAL(18,0)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `GATEWAY_MODEL_DEFINITIONS_LIST`¶

Calls MLflow REST API endpoint gateway/model-definitions/list.

SELECT GATEWAY_MODEL_DEFINITIONS_LIST(
    connection_name, provider, secret_id
);

Input Column	SQL Type	Comment
`provider`	`VARCHAR(2000000)`
`secret_id`	`VARCHAR(2000000)`

Output Column	SQL Type
`model_definition_id`	`VARCHAR(2000000)`
`name`	`VARCHAR(2000000)`
`secret_id`	`VARCHAR(2000000)`
`secret_name`	`VARCHAR(2000000)`
`provider`	`VARCHAR(2000000)`
`model_name`	`VARCHAR(2000000)`
`created_at`	`TIMESTAMP(3)`
`last_updated_at`	`TIMESTAMP(3)`
`created_by`	`VARCHAR(2000000)`
`last_updated_by`	`VARCHAR(2000000)`
`fallback_strategy`	`VARCHAR(2000000)`
`fallback_max_attempts`	`DECIMAL(18,0)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `MODEL_VERSIONS_GET`¶

Calls MLflow REST API endpoint model-versions/get.

SELECT MODEL_VERSIONS_GET(
    connection_name, name, version
);

Input Column	SQL Type	Comment
`name`	`VARCHAR(2000000)`	mandatory
`version`	`VARCHAR(2000000)`	mandatory

Output Column	SQL Type
`name`	`VARCHAR(2000000)`
`version`	`VARCHAR(2000000)`
`created`	`TIMESTAMP(3)`
`updated`	`TIMESTAMP(3)`
`user_id`	`VARCHAR(2000000)`
`current_stage`	`VARCHAR(2000000)`
`description`	`VARCHAR(2000000)`
`source`	`VARCHAR(2000000)`
`run_id`	`VARCHAR(2000000)`
`status`	`VARCHAR(2000000)`
`status_message`	`VARCHAR(2000000)`
`run_link`	`VARCHAR(2000000)`
`aliases`	`VARCHAR(2000000)`
`model_id`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `MODEL_VERSIONS_GET_DOWNLOAD_URI`¶

Calls MLflow REST API endpoint model-versions/get-download-uri.

SELECT MODEL_VERSIONS_GET_DOWNLOAD_URI(
    connection_name, name, version
);

Input Column	SQL Type	Comment
`name`	`VARCHAR(2000000)`	mandatory
`version`	`VARCHAR(2000000)`	mandatory

Output Column	SQL Type
`artifact_uri`	`VARCHAR(2000000)`

UDF `MODEL_VERSIONS_SEARCH`¶

Calls MLflow REST API endpoint model-versions/search.

SELECT MODEL_VERSIONS_SEARCH(
    connection_name, filter, order_by, max_results
);

Input Column	SQL Type	Comment
`filter`	`VARCHAR(2000000)`
`order_by`	`VARCHAR(2000000)`
`max_results`	`DECIMAL(18,0)`

Output Column	SQL Type
`name`	`VARCHAR(2000000)`
`version`	`VARCHAR(2000000)`
`created`	`TIMESTAMP(3)`
`updated`	`TIMESTAMP(3)`
`user_id`	`VARCHAR(2000000)`
`current_stage`	`VARCHAR(2000000)`
`description`	`VARCHAR(2000000)`
`source`	`VARCHAR(2000000)`
`run_id`	`VARCHAR(2000000)`
`status`	`VARCHAR(2000000)`
`status_message`	`VARCHAR(2000000)`
`run_link`	`VARCHAR(2000000)`
`aliases`	`VARCHAR(2000000)`
`model_id`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `REGISTERED_MODEL_GET`¶

Calls MLflow REST API endpoint registered-models/get.

SELECT REGISTERED_MODEL_GET(
    connection_name, name
);

Input Column	SQL Type	Comment
`name`	`VARCHAR(2000000)`	mandatory

Output Column	SQL Type
`name`	`VARCHAR(2000000)`
`created`	`TIMESTAMP(3)`
`updated`	`TIMESTAMP(3)`
`user_id`	`VARCHAR(2000000)`
`description`	`VARCHAR(2000000)`
`deployment_job_id`	`VARCHAR(2000000)`
`deployment_job_state`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `REGISTERED_MODELS_SEARCH`¶

Calls MLflow REST API endpoint registered-models/search.

SELECT REGISTERED_MODELS_SEARCH(
    connection_name, filter, order_by, max_results
);

Input Column	SQL Type	Comment
`filter`	`VARCHAR(2000000)`
`order_by`	`VARCHAR(2000000)`
`max_results`	`DECIMAL(18,0)`

Output Column	SQL Type
`name`	`VARCHAR(2000000)`
`created`	`TIMESTAMP(3)`
`updated`	`TIMESTAMP(3)`
`user_id`	`VARCHAR(2000000)`
`description`	`VARCHAR(2000000)`
`deployment_job_id`	`VARCHAR(2000000)`
`deployment_job_state`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `REGISTERED_MODELS_GET_LATEST_VERSIONS`¶

Calls MLflow REST API endpoint registered-models/get-latest-versions.

SELECT REGISTERED_MODELS_GET_LATEST_VERSIONS(
    connection_name, name, stages
);

Input Column	SQL Type	Comment
`name`	`VARCHAR(2000000)`	mandatory
`stages`	`VARCHAR(2000000)`

Output Column	SQL Type
`name`	`VARCHAR(2000000)`
`version`	`VARCHAR(2000000)`
`created`	`TIMESTAMP(3)`
`updated`	`TIMESTAMP(3)`
`user_id`	`VARCHAR(2000000)`
`current_stage`	`VARCHAR(2000000)`
`description`	`VARCHAR(2000000)`
`source`	`VARCHAR(2000000)`
`run_id`	`VARCHAR(2000000)`
`status`	`VARCHAR(2000000)`
`status_message`	`VARCHAR(2000000)`
`run_link`	`VARCHAR(2000000)`
`aliases`	`VARCHAR(2000000)`
`model_id`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

UDF `RUNS_SEARCH`¶

Calls MLflow REST API endpoint runs/search.

SELECT RUNS_SEARCH(
    connection_name, experiment_ids, filter, run_view_type, order_by, max_results
);

Input Column	SQL Type	Comment
`experiment_ids`	`VARCHAR(2000000)`	recommended
`filter`	`VARCHAR(2000000)`
`run_view_type`	`VARCHAR(2000000)`
`order_by`	`VARCHAR(2000000)`
`max_results`	`DECIMAL(18,0)`

Output Column	SQL Type
`run_id`	`VARCHAR(2000000)`
`run_uuid`	`VARCHAR(2000000)`
`run_name`	`VARCHAR(2000000)`
`experiment_id`	`VARCHAR(2000000)`
`user_id`	`VARCHAR(2000000)`
`status`	`VARCHAR(2000000)`
`start_time`	`TIMESTAMP(3)`
`end_time`	`TIMESTAMP(3)`
`artifact_uri`	`VARCHAR(2000000)`
`lifecycle_stage`	`VARCHAR(2000000)`
`tag_key`	`VARCHAR(2000000)`
`tag_value`	`VARCHAR(2000000)`

REST API UDFs¶

UDF ARTIFACTS_LIST¶

UDF EXPERIMENTS_SEARCH¶

UDF GATEWAY_ENDPOINTS_LIST¶

UDF GATEWAY_MODEL_DEFINITIONS_LIST¶

UDF MODEL_VERSIONS_GET¶

UDF MODEL_VERSIONS_GET_DOWNLOAD_URI¶

UDF MODEL_VERSIONS_SEARCH¶

UDF REGISTERED_MODEL_GET¶

UDF REGISTERED_MODELS_SEARCH¶

UDF REGISTERED_MODELS_GET_LATEST_VERSIONS¶

UDF RUNS_SEARCH¶