Remove pipeline from matcher initialization API #2569
Conversation
|
Is it a temporary solution? We should later unify the design of |
|
Job PR-2569-c54b2cc is done. |
zhiqiangdon
requested review from
sxjscience,
FANGAreNotGnu,
bryanyzhu,
cheungdaven,
suzhoum and
yongxinw
According to our general design discussions, predictor may need the |
zhiqiangdon
changed the title
zhiqiangdon
changed the title
PR updated. Now matcher uses only |
|
Job PR-2569-d423d4b is done. |
| @@ -501,3 +502,7 @@ def convert_data_for_ranking( | |||
| response_data = pd.DataFrame({response_column: data[response_column].unique().tolist()}) | |||
|
|
|||
| return data_with_label, query_data, response_data, label_column | |||
|
|
|||
|
|
|||
| def is_matching(pipeline: str): | |||
There was a problem hiding this comment.
Shall we use
?problem_property_dict[pipeline].is_matchingThere was a problem hiding this comment.
There is circular import issue if importing problem_property_dict from predictor.py. We may refactor this later.
| self._problem_type = problem_type.lower() if problem_type is not None else None | ||
| self._pipeline = pipeline.lower() if pipeline is not None else None | ||
| self._problem_type = None # always infer problem type for matching. | ||
| self._pipeline = problem_type.lower() if problem_type is not None else None |
There was a problem hiding this comment.
In the next PR, we can remove pipeline.
|
We can refactor the implementation further after #2578 |
|
Job PR-2569-f723802 is done. |
Issue #, if available:
Need to align some details between matcher and predictor before fully merging them.
Description of changes:
pipelinefrom the matcher initialization API.pipelinefrominfer_metrics(). We still need to know whether it's a matching task. But no longer require the data modalities in matching.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.