MLflow prediction requests API Specs¶
Methods
verify_prediction (response_json, expected_keywords_response,) |
verifies if a prediction output is as expected |
|
get_requests (model_name, databricks_cluster_hostname, databricks_workspace_token, settings, keywords, stage_env, request_time_out) |
makes post requests for model inference and verify that inference is within expectations |
verify_prediction¶
- verify_prediction(response_json: dict, expected_keywords_response: str)¶
- load an ML model from MLflow run, raises an exception if type_of_model is not in dictionary
- Parameters:
response_json (dict) – json response of the post request from model endpoint
expected_keywords_response (str) – expected top parent seo name for keywords
- Returns:
non exit response if response matches
- Return type:
bool
get_requests¶
- get_requests(model_name: str, databricks_cluster_hostname: str, databricks_workspace_token: str, settings: dict, keywords: str, stage_env: str = 'Production', request_time_out: int = 60)¶
- load an artifact from MLflow run, accepts `joblib, pkl, dict and yaml` file types
- Parameters:
model_name (str) – name of the registered model
databricks_cluster_hostname (str) – hostname of the databricks cluster
databricks_workspace_token (str) – token of the databricks workspace
settings (dict) – repo settings and configuration
keywords (str) – keywords to be used for prediction
stage_env (str) – stage of the registered model (e.g. Staging or Production)
request_time_out (int) – time out for the request
- Returns:
returns a callable python object; dictionary, pandas dataframe, list
- Return type:
int
