refactor: split disabled models for endpoints mapping into chat completions endpoint-disabled and completions endpoint-enabled mappings

[PChaparro]

Describe the change

This PR refactors model mapping logic by splitting the original disabledModelsForEndpoints map into two distinct maps: disabledModelsForChatCompletionsEndpoint and enabledModelsForDeprecatedCompletionsEndpoint.

This improves clarity and simplifies future maintenance of model support across endpoints.

Provide OpenAI documentation link

• https://platform.openai.com/docs/models
• https://platform.openai.com/docs/api-reference/completions
• https://platform.openai.com/docs/api-reference/chat/create

Describe your solution

• Created disabledModelsForChatCompletionsEndpoint to list models not supported by the chat completions endpoint. Maintaining a deny list is simpler since most new models are supported by chat completions by default.

• Created enabledModelsForDeprecatedCompletionsEndpoint to list models that are supported by the legacy completions endpoint. Maintaining an allow list is more appropriate here, since the completions endpoint is deprecated and no longer widely supported.

Tests

Executed the verification command provided in the CONTRIBUTING.md file. All tests passed successfully and the refactor did not introduce any regressions.

Issue: #972.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.46%. Comparing base (774fc9d) to head (cdb4504).
Report is 97 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master     #974       +/-   ##
===========================================
- Coverage   98.46%   85.46%   -13.00%     
===========================================
  Files          24       43       +19     
  Lines        1364     2271      +907     
===========================================
+ Hits         1343     1941      +598     
- Misses         15      308      +293     
- Partials        6       22       +16     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull Request Overview

This PR refactors the model mapping logic for endpoint support by splitting the original mapping into two distinct maps—one for models not supported by the chat completions endpoint and one for models supported by the deprecated completions endpoint.

• Defined a new constant (completionsEndpoint) to standardize the endpoint URL.
• Replaced the single map with two mappings and updated the logic in checkEndpointSupportsModel accordingly.
• Updated the client tests to include an explicit model field in CompletionRequest calls.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
completion.go	Refactored endpoint model mapping logic with new mappings
client_test.go	Updated tests to include explicit model requests for clarity

Comments suppressed due to low confidence (2)

completion.go:149

• [nitpick] Consider clarifying the behavior for models not found in enabledModelsForDeprecatedCompletionsEndpoint. An explicit default or additional comment may improve readability.

if endpoint == completionsEndpoint {

completion.go:153

• [nitpick] Consider adding a comment explaining the default 'true' behavior when a model is not present in disabledModelsForChatCompletionsEndpoint to aid future maintainers.

return !disabledModelsForChatCompletionsEndpoint[model]