Add deployment modal for model registry

[DaoDaoNoCode]

JIRA: RHOAIENG-8396

Description

When you deploy a model based on a model registry, the following modal will show:

If anything wrong with the project, the corresponding error message will pop up:

Otherwise, the current KServe modal or InferenceService modal will pop up depending on the project's serving platform. The name will be pre-filled as {modelName} - {versionName}:

There is a helper text if the model registry version contains a prompt model format:

The data connections part will be decided by the model artifact URI.
If there are more than 1 matches, we don't pre-select the data connection:

If there is only one match, we pre-select the recommended data connection:

If there is no match, we pre-select the new data connection field:

The path will always be pre-filled based on the URI parsing result:

How Has This Been Tested?

1. Create project A, B and C
2. For project A, make it an unsupported serving platform (or don't select a serving platform if both platforms are supported)
3. Go to Model registry page, choose one model registry and click on Deploy on the version action
4. Choose project A, check the error message
5. For project B and C, make the serving platform selected (or make your cluster only support one platform)
6. Choose project B, check if the name field is auto-filled
7. Check the data connection field is pre-selected as New data connection and path field is pre-filled
8. Create one matching secret in project B and 2 matching secrets in project C ("matching" means the secret has the same bucket, endpoint and region as the model registry version artifact URI you are going to deploy with)
9. Choose project B again, check the data connection field is pre-selected as Existing data connection and the only matching data connection is labeled and pre-selected
10. Choose project C, check the data connection field is pre-selected as Existing data connection and both data connections in the dropdown menu are labeled as recommended, but neither is pre-selected

Test Impact

Add a unit test to test the util function that shows the error message when choosing the project.

Request review criteria:

Self checklist (all need to be checked):

• The developer has manually tested the changes and verified that the changes work
• Commits have been squashed into descriptive, self-contained units of work (e.g. 'WIP' and 'Implements feedback' style messages have been removed)
• Testing instructions have been added in the PR body (for PRs involving changes that are not immediately obvious).
• The developer has added tests or explained why testing cannot be added (unit or cypress tests for related changes)

If you have UI changes:

• Included any necessary screenshots or gifs if it was a UI change.
• Included tags to the UX team if it was a UI/UX change.

After the PR is posted & before it merges:

• The developer has tested their solution on a cluster by using the image produced by the PR to main

[yih-wang]

Thanks for the work @DaoDaoNoCode! Just a few comments -

1. Can you stretch the modal width in the first image to the same width as the other modals?
2. Regarding the pre-filled model name, I can't see what happens if a version has been deployed multiple times. The ideal behavior is if the name has already been used within the selected project, add a postfix such as ‘-2’, ‘-3’ to differentiate it.
3. Regarding the helper text to the prompt model format - the helper text will only show up after a model server is selected, and the model framework dropdown is available.
4. For data connection, when there is no match and the new data connection is selected by default - a warning is missing, check details in the mockup

[DaoDaoNoCode]

@yih-wang Thanks for the review! I will update it!

[mturley]

This is looking really good! I have one actual change request (the comment in useLabeledDataConnections) and then just some questions and suggestions.

[DaoDaoNoCode]

1. Can you stretch the modal width in the first image to the same width as the other modals?
2. Regarding the pre-filled model name, I can't see what happens if a version has been deployed multiple times. The ideal behavior is if the name has already been used within the selected project, add a postfix such as ‘-2’, ‘-3’ to differentiate it.
3. Regarding the helper text to the prompt model format - the helper text will only show up after a model server is selected, and the model framework dropdown is available.
4. For data connection, when there is no match and the new data connection is selected by default - a warning is missing, check details in the mockup

Based on the comments above, I updated the following:

1. For your first comment, I updated the size of the initial modal

2. For the second comment, I found it's a little bit hard from the technical side to implement what you requested. We will need to fetch all the models in that project and it's hard to decide the postfix number (e.g. There are model-1, model-3 and model-4, what should the auto-generated number be? model-2 or model-5?). I prefer we keep it as it is now, and if the name is duplicated, the server will always prompt an error when the user submits the model. (cc @mturley Might need your input here)
3. For the third comment, I updated the helper text to make it only show when there are items in the dropdown menu.

No items in the menu:

Items in the menu:

4. For the fourth comment, I updated the alert:

@yih-wang Let me know what you think, thanks!

[yih-wang]

@DaoDaoNoCode All the changes look good. Thanks for the rapid work!
Regarding the auto-generated deployment name, my primary concern is to avoid duplicate names and prevent user frustration from seeing duplicated names error continuously. If the original proposal doesn't work, we can implement a workaround by adding a timestamp as postfix to each name. So we'll have a format like [modelname-versionname-yyyymmddhhmmss], which should effectively eliminate duplicates as well.

[DaoDaoNoCode]

Thanks @yih-wang I made the requested change.

[yih-wang]

LGTM, thanks @DaoDaoNoCode !

[mturley]

Pulled and tested on the modelregistry-ui cluster - everything looks and works great here!! Thanks so much for the hard work @DaoDaoNoCode !

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mturley

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• frontend/OWNERS [mturley]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment