feat: get the inference provider from a registered store

Signed-off-by: Brian <[email protected]>

Author: bmahabirbu

packages/backend/src/studio-api-impl.spec.ts

@@ -43,6 +43,7 @@ import type { RecipeManager } from './managers/recipes/RecipeManager';
 import type { PodmanConnection } from './managers/podmanConnection';
 import type { NavigationRegistry } from './registries/NavigationRegistry';
 import type { RpcExtension } from '@shared/messages/MessageProxy';
+import type { InferenceProviderRegistry } from './registries/InferenceProviderRegistry';
 
 vi.mock('./ai.json', () => {
   return {
@@ -148,6 +149,7 @@ beforeEach(async () => {
     localRepositoryRegistry,
     {} as unknown as TaskRegistry,
     {} as unknown as InferenceManager,
+    {} as unknown as InferenceProviderRegistry,
     {} as unknown as PlaygroundV2Manager,
     {} as unknown as SnippetManager,
     {} as unknown as CancellationTokenRegistry,

packages/backend/src/studio-api-impl.ts

@@ -33,6 +33,7 @@ import path from 'node:path';
 import type { InferenceServer } from '@shared/models/IInference';
 import type { CreationInferenceServerOptions } from '@shared/models/InferenceServerConfig';
 import type { InferenceManager } from './managers/inference/inferenceManager';
+import type { InferenceProviderRegistry } from './registries/InferenceProviderRegistry';
 import type { Conversation } from '@shared/models/IPlaygroundMessage';
 import type { PlaygroundV2Manager } from './managers/playgroundV2Manager';
 import { getFreeRandomPort } from './utils/ports';
@@ -72,6 +73,7 @@ export class StudioApiImpl implements StudioAPI {
     private localRepositories: LocalRepositoryRegistry,
     private taskRegistry: TaskRegistry,
     private inferenceManager: InferenceManager,
+    private InferenceProviderRegistry: InferenceProviderRegistry,
     private playgroundV2: PlaygroundV2Manager,
     private snippetManager: SnippetManager,
     private cancellationTokenRegistry: CancellationTokenRegistry,
@@ -144,6 +146,10 @@ export class StudioApiImpl implements StudioAPI {
     return this.inferenceManager.getServers();
   }
 
+  async getProviders(): Promise<string[]> {
+    return this.InferenceProviderRegistry.getAll().map(provider => provider.type);
+  }
+
   async requestDeleteInferenceServer(...containerIds: string[]): Promise<void> {
     // Do not wait on the promise as the api would probably timeout before the user answer.
     if (containerIds.length === 0) throw new Error('At least one container id should be provided.');

packages/backend/src/studio.ts

@@ -398,6 +398,7 @@ export class Studio {
       this.#localRepositoryRegistry,
       this.#taskRegistry,
       this.#inferenceManager,
+      this.#inferenceProviderRegistry,
       this.#playgroundManager,
       this.#snippetManager,
       this.#cancellationTokenRegistry,

packages/frontend/src/lib/select/InferenceRuntimeSelect.svelte

@@ -1,34 +1,38 @@
 <script lang="ts">
 import Select from '/@/lib/select/Select.svelte';
-import { InferenceType } from '@shared/models/IInference';
+import { inferenceProviders } from '/@/stores/inferenceProviders';
 
 interface Props {
   disabled?: boolean;
-  value: InferenceType | undefined;
-  exclude?: InferenceType[];
+  value: string | undefined;
+  exclude?: string[];
 }
 let { value = $bindable(), disabled, exclude = [] }: Props = $props();
 
-// Filter options based on optional exclude list
-const options = Object.values(InferenceType).filter(type => type !== InferenceType.NONE && !exclude.includes(type));
+// Create a derived store for providerOptions filtering by exclude and enabled
+let providerOptions = $derived(() => {
+  return $inferenceProviders
+    .filter(name => !exclude.includes(name))
+    .map(name => ({
+      value: name,
+      label: name,
+    }));
+});
 
 function handleOnChange(nValue: { value: string } | undefined): void {
   if (nValue) {
-    value = nValue.value as InferenceType;
+    value = nValue.value;
   } else {
     value = undefined;
   }
 }
 </script>
 
 <Select
-  label="Select Inference Runtime"
-  name="select-inference-runtime"
+  label="Select Inference Provider"
+  name="select-inference-provider"
   disabled={disabled}
   value={value ? { label: value, value: value } : undefined}
   onchange={handleOnChange}
-  placeholder="Select Inference Runtime to use"
-  items={options.map(type => ({
-    value: type,
-    label: type,
-  }))} />
+  placeholder="Select Inference Provider to use"
+  items={providerOptions()} />

packages/frontend/src/pages/PlaygroundCreate.svelte

@@ -36,7 +36,7 @@ let trackingId: string | undefined = undefined;
 // The trackedTasks are the tasks linked to the trackingId
 let trackedTasks: Task[] = [];
 
-// Preset model selection
+// Preset model selection depending on runtime
 $: if (localModels.length > 0) {
   model = localModels[0];
 } else {

packages/frontend/src/stores/inferenceProviders.ts

@@ -0,0 +1,22 @@
+/**********************************************************************
+ * Copyright (C) 2024 Red Hat, Inc.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ ***********************************************************************/
+import { RPCReadable } from '/@/stores/rpcReadable';
+import { MSG_INFERENCE_PROVIDER_UPDATE } from '@shared/Messages';
+import { studioClient } from '/@/utils/client';
+
+export const inferenceProviders = RPCReadable<string[]>([], MSG_INFERENCE_PROVIDER_UPDATE, studioClient.getProviders);

packages/shared/src/StudioAPI.ts

@@ -121,6 +121,11 @@ export interface StudioAPI {
    */
   getInferenceServers(): Promise<InferenceServer[]>;
 
+  /**
+   * Get inference providers
+   */
+  getProviders(): Promise<string[]>;
+
   /**
    * Request to start an inference server
    * @param options The options to use