Added support to fetch models for groq

src/modules/llms/server/openai/groq.wiretypes.ts

@@ -0,0 +1,13 @@
+import { z } from 'zod';
+
+
+// [Groq] Models List API - Response
+
+export const wireGroqModelsListOutputSchema = z.object({
+  id: z.string(),
+  object: z.literal('model'),
+  created: z.number(),
+  owned_by: z.string(),
+  active: z.boolean(),
+});
+

src/modules/llms/server/openai/models.data.ts

@@ -4,7 +4,7 @@ import type { ModelDescriptionSchema } from '../llm.server.types';
 import { wireMistralModelsListOutputSchema } from './mistral.wiretypes';
 import { wireOpenrouterModelsListOutputSchema } from './openrouter.wiretypes';
 import { wireTogetherAIListOutputSchema } from '~/modules/llms/server/openai/togetherai.wiretypes';
-
+import { wireGroqModelsListOutputSchema } from './groq.wiretypes';
 
 // [Azure] / [OpenAI]
 const _knownOpenAIChatModels: ManualMappings = [
@@ -778,23 +778,35 @@ export function perplexityAIModelSort(a: ModelDescriptionSchema, b: ModelDescrip
 
 // Groq
 
-const _knownGroqModels: ModelDescriptionSchema[] = [
+const _knownGroqModels: ManualMappings = [
+  // {
+  //   id: 'lama2-70b-4096',
+  //   label: 'Llama 2 70B Chat',
+  //   description: 'Llama 2 is a collection of pretrained and fine-tuned generative text models.',
+  //   contextWindow: 4096,
+  //   interfaces: [LLM_IF_OAI_Chat],
+  // },
   {
-    id: 'lama2-70b-4096',
-    label: 'Llama 2 70B Chat',
-    description: 'Llama 2 is a collection of pretrained and fine-tuned generative text models.',
-    contextWindow: 4096,
-    interfaces: [LLM_IF_OAI_Chat],
-  },
-  {
-    id: 'mixtral-8x7b-32768',
+    idPrefix: 'mixtral-8x7b-32768',
     label: 'Mixtral 8x7B Instruct v0.1',
     description: 'The Mixtral-8x7B Large Language Model (LLM) is a pretrained generative Sparse Mixture of Experts.',
     contextWindow: 32768,
     interfaces: [LLM_IF_OAI_Chat],
   }
 ];
 
+export function groqModelToModelDescription(_model: unknown): ModelDescriptionSchema {
+  const model = wireGroqModelsListOutputSchema.parse(_model);
+  return fromManualMapping(_knownGroqModels, model.id, model.created, undefined, {
+    idPrefix: model.id,
+    label: model.id.replaceAll(/[_-]/g, ' '),
+    description: 'New Model',
+    contextWindow: 32768,
+    interfaces: [LLM_IF_OAI_Chat],
+    hidden: true,
+  });
+}
+
 export function groqModelDescriptions() {
   // change this implementation once upstream implements some form of models listing
   return _knownGroqModels;

src/modules/llms/server/openai/openai.router.ts

@@ -11,11 +11,10 @@ import { Brand } from '~/common/app.config';
 import { fixupHost } from '~/common/util/urlUtils';
 
 import { OpenAIWire, WireOpenAICreateImageOutput, wireOpenAICreateImageOutputSchema, WireOpenAICreateImageRequest } from './openai.wiretypes';
-import { azureModelToModelDescription, groqModelDescriptions, lmStudioModelToModelDescription, localAIModelToModelDescription, mistralModelsSort, mistralModelToModelDescription, oobaboogaModelToModelDescription, openAIModelToModelDescription, openRouterModelFamilySortFn, openRouterModelToModelDescription, perplexityAIModelDescriptions, perplexityAIModelSort, togetherAIModelsToModelDescriptions } from './models.data';
+import { azureModelToModelDescription, groqModelToModelDescription, lmStudioModelToModelDescription, localAIModelToModelDescription, mistralModelsSort, mistralModelToModelDescription, oobaboogaModelToModelDescription, openAIModelToModelDescription, openRouterModelFamilySortFn, openRouterModelToModelDescription, perplexityAIModelDescriptions, perplexityAIModelSort, togetherAIModelsToModelDescriptions } from './models.data';
 import { llmsChatGenerateWithFunctionsOutputSchema, llmsListModelsOutputSchema, ModelDescriptionSchema } from '../llm.server.types';
 import { wilreLocalAIModelsApplyOutputSchema, wireLocalAIModelsAvailableOutputSchema, wireLocalAIModelsListOutputSchema } from './localai.wiretypes';
 
-
 const openAIDialects = z.enum([
   'azure', 'groq', 'lmstudio', 'localai', 'mistral', 'oobabooga', 'openai', 'openrouter', 'perplexity', 'togetherai',
 ]);
@@ -133,10 +132,6 @@ export const llmOpenAIRouter = createTRPCRouter({
         return { models };
       }
 
-      // [Groq]: use a custom list of models 
-      // https://console.groq.com/docs/models
-      if (access.dialect === 'groq')
-        return { models: groqModelDescriptions() };
 
       // [Perplexity]: there's no API for models listing (upstream: https://docs.perplexity.ai/discuss/65cf7fd19ac9a5002e8f1341)
       if (access.dialect === 'perplexity')
@@ -164,6 +159,11 @@ export const llmOpenAIRouter = createTRPCRouter({
       // every dialect has a different way to enumerate models - we execute the mapping on the server side
       switch (access.dialect) {
 
+        case 'groq':
+          models = openAIModels
+            .map(groqModelToModelDescription);
+          break;
+
         case 'lmstudio':
           models = openAIModels
             .map(({ id }) => lmStudioModelToModelDescription(id));