-
Notifications
You must be signed in to change notification settings - Fork 66
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert PayloadProcessor ModelId to remain consistent with original value #113
Conversation
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: RobGeada The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
@RobGeada -- can you add some details on how to test this fix? And it would be great to add a unit test as well |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @RobGeada -- could you add a unit test that teases out the bug and shows how your change fixes it? The test should fail before your fix and pass with your fix. Thank you!
@ckadner it isn't actually a bug that this PR is fixing, it's just reverting to an older behavior. Prior to #90, the PayloadProcessor received the following modelId:
After PR90, the payload processor receives:
where String modelOrVModelId = validateModelId(midIt.next(), isVModel);
if (isVModel){
vModelId = modelOrVModelId;
resolvedModelId = ModelMeshApi.resolvedModelId.getIfExists();
} else {
// vModelId remains null in this case
resolvedModelId = modelOrVModelId;
} (see code here) Essentially, the payload process used to receive the output of
therefore restoring the original behavior. |
@njhill any updates here? |
@RobGeada apologies for the delay in getting to this. I think the reason that this was changed in #90 was to ensure consistency. I don't think it's good for the payload processor to get this id without knowing whether it's a model or vModel id. When used as part of kserve/modelmesh-serving, the model id will always be the vModel id with some additional suffix. Do you think that this would be sufficient for your purposes? Otherwise we should probably just add a (possibly-null) vModelId string field to the |
And if we did the latter, we could even also add a |
Also related: #88 |
Yeah, that works for me, I'll close this |
Motivation
PR 90 changed the
modelId
passed to thePayloadProcessor
to theresolvedModelId
, instead of the originalmodelOrVModelId
.Modifications
This PR reverts this change to restore the expected model ID to
modelOrVModelId
, whose value is either found insidevModelId
orresolvedModelId
depending on the flagisvModel
Result
modelId
remains consistent with previous behaviour in payload processor payloads