provider: added support for Hyperbolic AI#1108
Conversation
narengogi
left a comment
narengogi
left a comment
There was a problem hiding this comment.
looks good, just need to remove the boilerplate
| } from './imageGenerate'; | ||
|
|
||
| const HyperbolicConfig: ProviderConfigs = { | ||
| chatComplete: HyperbolicChatCompleteConfig, |
There was a problem hiding this comment.
I can see in the documentation that the chat completions endpoint is openai compliant, you can use the transformers for openai-base and avoid the boilerplate
checkout the implementation for any of groq, x-ai, nebius etc.
the open-ai-base package has the implementation
example PR: https://github.com/Portkey-AI/gateway/pull/801/files
There was a problem hiding this comment.
Thanks for your review! I'm working on the requested changes and will update the pull request soon.
|
DescriptionSummary By MatterAI
🔄 What ChangedThis pull request introduces comprehensive support for Hyperbolic AI as a new provider. Key changes include:
🔍 Impact of the ChangeThis change significantly expands the platform's capabilities by integrating a new AI provider, Hyperbolic AI. Users can now leverage Hyperbolic AI for both chat completion and image generation tasks, offering more flexibility and choice in AI models. The structured integration ensures maintainability and consistency with existing provider patterns. 📁 Total Files Changed
🧪 Test Added(explain each test in detail)No explicit tests were added or described in the pull request. Manual testing is presumed to have been performed to verify the functionality of the new Hyperbolic AI integration for chat completion and image generation. 🔒Security VulnerabilitiesNo new security vulnerabilities were detected in the provided code changes. The API key handling follows standard bearer token authentication, which is secure when the key is managed properly. MotivationAdding support for a new AI provider, Hyperbolic AI, to expand the platform's capabilities for chat completion and image generation. Type of Change
How Has This Been Tested?
Screenshots (if applicable)N/A Checklist
Related IssuesN/A Tip Quality Recommendations
Sequence DiagramsequenceDiagram
actor User
participant ClientApplication
participant ProvidersEntrypoint as src/providers/index.ts
participant HyperbolicConfig as src/providers/hyperbolic/index.ts
participant HyperbolicAPIConfig as src/providers/hyperbolic/api.ts
participant HyperbolicAIApi as Hyperbolic AI API
participant ChatCompleteTransform as src/providers/hyperbolic/chatComplete.ts
participant ImageGenerateTransform as src/providers/hyperbolic/imageGenerate.ts
User->>ClientApplication: Initiates AI Request (e.g., chatComplete, imageGenerate)
ClientApplication->>ProvidersEntrypoint: Call Provider (provider: 'hyperbolic', fn: 'chatComplete'/'imageGenerate', params)
ProvidersEntrypoint->>HyperbolicConfig: Route to HyperbolicConfig
HyperbolicConfig->>HyperbolicAPIConfig: Get API Configuration (fn)
HyperbolicAPIConfig-->>HyperbolicConfig: Returns baseURL, headers, endpoint (e.g., '/v1/chat/completions' or '/v1/image/generation')
HyperbolicConfig->>HyperbolicAIApi: Make API Call (POST baseURL + endpoint, headers, body)
HyperbolicAIApi-->>HyperbolicConfig: Returns Response (stream or JSON)
alt If chatComplete and streaming
HyperbolicConfig->>ChatCompleteTransform: Transform Stream Chunk (responseChunk)
ChatCompleteTransform-->>HyperbolicConfig: Returns Transformed Chunk (data: { ... }
\n)
else If imageGenerate
HyperbolicConfig->>ImageGenerateTransform: Transform Image Response (response, responseStatus)
ImageGenerateTransform-->>HyperbolicConfig: Returns Transformed Response
end
HyperbolicConfig-->>ProvidersEntrypoint: Returns Processed Response
ProvidersEntrypoint-->>ClientApplication: Returns Final Response
ClientApplication-->>User: Displays AI Output
|
|
Important PR Review SkippedPR review skipped as per the configuration setting. Run a manually review by commenting /matter review 💡Tips to use Matter AICommand List
|
|
@narengogi, I have made the necessary changes. Just had one question about the documentation, parameters frequency_penalty and presence_penalty have different value ranges from the OpenAI docs. Also, I have added extra params ( top_k, min_p, repetition_penalty ) which were not present in openai listed params. Please take a look and let me know if anything else is needed. |
|
Important PR Review SkippedPR review skipped as per the configuration setting. Run a manually review by commenting /matter review 💡Tips to use Matter AICommand List
|
|
Important PR Review SkippedPR review skipped as per the configuration setting. Run a manually review by commenting /matter review 💡Tips to use Matter AICommand List
|
3 participants
Description
Added Hyperbolic provider support, including API configuration and response transforms for chat completion and image generation.
Motivation
This adds support for the Hyperbolic provider as requested in issue #672
Type of Change
Related Issues
Fixes #672