Please enable JavaScript.
Coggle requires JavaScript to display documents.
mlops/authentication_to_azureml.py, mlops/register_model_functions -…
mlops/authentication_to_azureml.py
get_key_vault_name
purpose
Get Key Vault name for databricks secrets from jobs parameter
returns
string
, with the key vault name
authentication_to_azureml
calls (up)
purpose
Set up a connection to azure-ml using service principal, creates a workspace and writes the workspace ARM properties to a config file.
Workspace ARM properties can be uploaded later using the
azureml.core.workspace.Workspace.from_config
method.
needs
key-vault-secret
azure-ml-tenant-id
azure-ml-service-principal-id
azure-ml-service-principal-app-secret
azure-ml-subscription-id
azure-ml-resource-group
azure-ml-workspace-name
returns
azureml.core.Workspace
- A workspace is the top-level resource for Azure Machine Learning, providing a centralized place to work with all the artifacts you create when you use Azure Machine Learning
mlops/register_model_functions
get_local_model
purpose
Creates a model URI from parameters and load a model stored in Python function format.
parameters
model_name
model_version
Must be either a string representation of a integer or a integer
returns
mlflow.pyfunc.PyFuncModel
, a MLflow python function model
get_local_model_dbfs_path
parameters
model_name
model_version
Must be either a string representation of a integer or a integer
returns
string
, with the download location in Model Registry for this model version.
purpose
Creates a unified MLflow client from MLflow Tracking Server and MLflow Registry Server objects with the purpose of creating and managing experiments and runs, registering models and versions.
The purpose is to get the download location in Model Registry for this model version.
register_local_model_into_AzureML
parameters
local_model_path
model_name
ws
stage_tag