Created
May 31, 2021 08:21
-
-
Save sroecker/3ffabf589c9de9a528b388254af69979 to your computer and use it in GitHub Desktop.
DataRobot Python API Get Prediction Explanations
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
def upload_dataset_from_catalog(self, dataset): | |
import json | |
import time | |
# client = dr.Client() | |
res = client.post(client.endpoint + "/projects/{}/predictionDatasets/datasetUploads/".format(self.id), json={"datasetId": dataset.id}) | |
dataset_id = json.loads(res.text)['datasetId'] | |
while dataset_id not in [ds.id for ds in self.get_datasets()]: | |
time.sleep(1) | |
return dataset_id | |
def get_prediction_explanations(self, dataset_id): | |
project = dr.Project.get(self.project_id) | |
if project.advanced_options['shap_only_mode'] == False: | |
try: | |
dr.PredictionExplanationsInitialization.get(project.id, self.id) | |
except dr.errors.ClientError as e: | |
if e.json['message'] == 'No prediction explanations initialization found for model': | |
print('Initialising Explanations...') | |
while True: | |
try: | |
pei_job = dr.PredictionExplanationsInitialization.create(project.id, self.id) | |
pei_job.wait_for_completion() | |
except dr.errors.ClientError as e: | |
if e.json['message'] == 'A prerequisite was not satisfied: Feature impact has not been computed for the model. Run feature impact for this model first.': | |
print('Calculating Feature Impact...') | |
self.get_or_request_feature_impact() | |
else: | |
raise(e) | |
else: | |
print('Explanations Initialised') | |
break | |
else: | |
raise(e) | |
print('Calculation Explanations...') | |
while True: | |
try: | |
pe_job = dr.PredictionExplanations.create(project.id, self.id, dataset_id) | |
pe = pe_job.get_result_when_complete() | |
except dr.errors.ClientError as e: | |
if e.json['message'] == 'Predictions must first be generated for this dataset': | |
print('Calculating Predictions...') | |
p_job = self.request_predictions(dataset_id) | |
p_job.get_result_when_complete() | |
else: | |
raise(e) | |
else: | |
print('Explanations Done') | |
break | |
return pe.get_all_as_dataframe() | |
else: | |
shap_matrix_job = dr.models.ShapMatrix.create(project.id, self.id, dataset_id) | |
shap_matrix = shap_matrix_job.get_result_when_complete() | |
return shap_matrix.get_as_dataframe() | |
dr.Project.upload_dataset_from_catalog = upload_dataset_from_catalog | |
dr.Model.get_prediction_explanations = get_prediction_explanations |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment