Use the value of `regresed_by` in crafting the training classes for the RegressionRange model (hashtag3759)

[vikramsubramanian]

Currently, we use the regressionwindow-wanted keyword and the value of the cf_has_regression_range field to determine whether there is a regression range:

We should also use the value of the regresed_by field. If there is a value in the field, we should consider it as "has regression range".

)

[mayil-ai[bot]]

Possible Solution

• Update the get_labels method in the RegressionRangeModel class within bugbug/models/regressionrange.py to include the regresed_by field in the logic that determines if a bug has a regression range.
• Modify the condition that checks for the presence of a regression range to also consider if the regresed_by field is not empty (i.e., it has one or more entries).

class RegressionRangeModel(BugModel):
    # ... [existing code] ...

    def get_labels(self):
        classes = {}

        for bug_data in bugzilla.get_bugs():
            if "regression" not in bug_data["keywords"]:
                continue

            bug_id = int(bug_data["id"])
            # Check if 'regressionwindow-wanted' is in keywords or 'cf_has_regression_range' is 'yes'
            # or 'regresed_by' field is not empty.
            if ("regressionwindow-wanted" in bug_data["keywords"] or
                bug_data.get("cf_has_regression_range") == "yes" or
                bug_data.get("regresed_by")):
                classes[bug_id] = 1
            else:
                classes[bug_id] = 0

        # ... [existing logging code] ...

        return classes, [0, 1]

• Ensure that the regresed_by field is being retrieved correctly from the Bugzilla data source and that it is a list or other iterable that can be truthy when not empty.
• There is no need to modify the hypothetical determine_regression_range method since it is not present in the provided code snippets.

Code snippets to check

bugbug → models → regressionrange.py

This snippet contains the logic for determining whether a bug has a regression range based on keywords and the 'cf_has_regression_range' field, which is directly related to the issue. https://github.com/Mayil-AI/bugbug-21dec23/blob/0acd00da46afbb37bd047c0bce06ce7cfad21568/bugbug/models/regressionrange.py#L22-L102