Add Azure AI Studio as provider

[helioLJ]

Checklist:

Important

Please review the checklist below before submitting your pull request.

• Please open an issue before creating a PR or link to an existing issue
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I ran dev/reformat(backend) and cd web && npx lint-staged(frontend) to appease the lint gods

Description

This PR adds Azure AI Studio as a new model provider in Dify. It implements support for both LLM and rerank models, allowing users to leverage Azure AI Studio's capabilities within the Dify ecosystem. This addition is particularly beneficial for organizations that exclusively use Azure and need access to models like the Cohere Reranker through Azure AI Studio Catalog.

Close #7326

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update, included: Dify Document
• Improvement, including but not limited to code refactoring, performance optimization, and UI/UX improvement
• Dependency upgrade

Testing Instructions

Please test the following to verify the changes:

• Test Azure AI Studio LLM model integration
  • Configure Azure AI Studio credentials in Dify
  • Create a new application using an Azure AI Studio LLM model
  • Test text generation with the configured model
• Test Azure AI Studio Rerank model integration
  • Configure Azure AI Studio credentials for a rerank model
  • Use the rerank model in a relevant Dify feature (e.g., semantic search)
  • Verify that document reranking works as expected
• Verify that existing model providers still function correctly

[crazywoola]

Thanks for your contribution :) Please resolve the deps conflict and add tests to the new provider accordingly.

[helioLJ]

Thanks for your contribution :) Please resolve the deps conflict and add tests to the new provider accordingly.

You are welcome :)

Conflicts solved, added the tests files and update the larger icon to a better one!

[dyohan9]

I left a few points to adjust, congratulations on your contribution, I'm sure it will help many people.

[helioLJ]

I left a few points to adjust, congratulations on your contribution, I'm sure it will help many people.

Good observations, mate! I ran Ruff linter in the files and added proper exception mapping.

[bhctest123]

Hey guys, I was doing some testing for the community and hit an error. I set up the keys correctly and everything in Azure Ai Studio. see images to see the errors! @crazywoola @takatost @laipz8200 @helioLJ

[helioLJ]

Hey guys, I was doing some testing for the community and hit an error. I set up the keys correctly and everything in Azure Ai Studio. see images to see the errors! @crazywoola @takatost @laipz8200 @helioLJ

Analyzing your deployment and mine that I used for test it (ibm-re2g-reranker-nq-3), they have different Compute Types, the one from IBM uses Managed type, while Cohere models use Consumption type, so I suspect the Authentication type are different for both.

Can you send me the sample code for consumption, in Python, in that "Consume" tab, I will fix that later.

Also, I'm almost pretty sure that using this inference endpoint and API Key can work with Cohere Provider Model, but not 100% sure.

[bhctest123]

Hey @helioLJ I see the reranker you used. that is $80+ dollars per month because you are deploying an actual server with that reranker model running. The cohere endpoint is an API hosted by Microsoft and only costs $2 per 1000 queries. It seems more cost efficient and more likely to help the Dify community. Do you think its possible to support the cohere type of reranker?

[helioLJ]

Hey @helioLJ I see the reranker you used. that is $80+ dollars per month because you are deploying an actual server with that reranker model running. The cohere endpoint is an API hosted by Microsoft and only costs $2 per 1000 queries. It seems more cost efficient and more likely to help the Dify community. Do you think its possible to support the cohere type of reranker?

Sure! I will update to support it, would you mind message me at Discord? I'll probably need your help. 'heliolj' at Dify's server.