Data review

A data review runs a published checklist against a run. Each place a check fails becomes an event for an engineer to review.

For checking live data as it arrives, see Trigger alerts.

Connect the client

This guide assumes you have installed the client and setup your API keys.

from nominal.core import NominalClient

client = NominalClient.from_profile("default")

Run checklists against runs

data_review_builder() returns a DataReviewBuilder. Add one execute_checklist() per run and checklist pair, then start them all with initiate():

builder = client.data_review_builder()
builder.execute_checklist("<RUN_RID>", "<CHECKLIST_RID>", commit="<COMMIT_SHA>")  # pin a version
builder.execute_checklist("<OTHER_RUN_RID>", "<CHECKLIST_RID>")  # latest version
builder.add_integration("<INTEGRATION_RID>")  # optional: notify when checks fail

for review in builder.initiate():
    print(review.nominal_url, len(review.get_events()))

Copy checklist RIDs and commit SHAs from the Checklists page. For a run with several assets, pass asset= to choose the one to review.

initiate waits for the reviews to finish by default, but doesn’t wait for notifications. Pass wait_for_completion=False to return immediately, then call poll_for_completion() on each review.

Notifications support Slack, Opsgenie and webhook integrations.

Read results later

review = client.get_data_review("<DATA_REVIEW_RID>")
events = review.get_events()

search_data_reviews() and search_data_reviews() find reviews by run or asset. To review the results in the UI, see Data review.