@ai-sdk/openai-compatible 0.0.1

package/CHANGELOG.md

@@ -0,0 +1,7 @@
+# @ai-sdk/openai-compatible
+
+## 0.0.1
+
+### Patch Changes
+
+- 962978b: feat (packages/openai-compatible): Base for OpenAI-compatible providers.

package/LICENSE

@@ -0,0 +1,13 @@
+Copyright 2023 Vercel, 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.

package/README.md

@@ -0,0 +1,82 @@
+# AI SDK - OpenAI Compatible Provider
+
+This package provides a foundation for implementing providers that expose an OpenAI-compatible API.
+
+The primary [OpenAI provider](../openai/README.md) is more feature-rich, including OpenAI-specific experimental and legacy features. This package offers a lighter-weight alternative focused on core OpenAI-compatible functionality.
+
+## Setup
+
+The provider is available in the `@ai-sdk/openai-compatible` module. You can install it with
+
+```bash
+npm i @ai-sdk/openai-compatible
+```
+
+## Provider Instance
+
+You can import the provider creation method `createOpenAICompatible` from `@ai-sdk/openai-compatible`:
+
+```ts
+import { createOpenAICompatible } from '@ai-sdk/openai-compatible';
+```
+
+## Example
+
+```ts
+import { createOpenAICompatible } from '@ai-sdk/openai-compatible';
+import { generateText } from 'ai';
+
+const { text } = await generateText({
+  model: createOpenAICompatible({
+    baseURL: 'https://api.example.com/v1',
+    name: 'example',
+    headers: {
+      Authorization: `Bearer ${process.env.MY_API_KEY}`,
+    },
+  }).chatModel('meta-llama/Llama-3-70b-chat-hf'),
+  prompt: 'Write a vegetarian lasagna recipe for 4 people.',
+});
+```
+
+### Including model ids for auto-completion
+
+```ts
+import { createOpenAICompatible } from '@ai-sdk/openai-compatible';
+import { generateText } from 'ai';
+
+type ExampleChatModelIds =
+  | 'meta-llama/Llama-3-70b-chat-hf'
+  | 'meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo'
+  | (string & {});
+
+type ExampleCompletionModelIds =
+  | 'codellama/CodeLlama-34b-Instruct-hf'
+  | 'Qwen/Qwen2.5-Coder-32B-Instruct'
+  | (string & {});
+
+type ExampleEmbeddingModelIds =
+  | 'BAAI/bge-large-en-v1.5'
+  | 'bert-base-uncased'
+  | (string & {});
+
+const model = createOpenAICompatible<
+  ExampleChatModelIds,
+  ExampleCompletionModelIds,
+  ExampleEmbeddingModelIds
+>({
+  baseURL: 'https://api.example.com/v1',
+  name: 'example',
+  headers: {
+    Authorization: `Bearer ${process.env.MY_API_KEY}`,
+  },
+});
+
+// Subsequent calls to e.g. `model.chatModel` will auto-complete the model id
+// from the list of `ExampleChatModelIds` while still allowing free-form
+// strings as well.
+
+const { text } = await generateText({
+  model: model.chatModel('meta-llama/Llama-3-70b-chat-hf'),
+  prompt: 'Write a vegetarian lasagna recipe for 4 people.',
+});
+```

package/dist/index.d.mts

@@ -0,0 +1,170 @@
+import { ProviderV1, LanguageModelV1, EmbeddingModelV1, LanguageModelV1ObjectGenerationMode } from '@ai-sdk/provider';
+import { FetchFunction } from '@ai-sdk/provider-utils';
+
+type OpenAICompatibleChatModelId = string;
+interface OpenAICompatibleChatSettings {
+    /**
+  A unique identifier representing your end-user, which can help the provider to
+  monitor and detect abuse.
+    */
+    user?: string;
+}
+
+type OpenAICompatibleCompletionModelId = string;
+interface OpenAICompatibleCompletionSettings {
+    /**
+  Echo back the prompt in addition to the completion.
+     */
+    echo?: boolean;
+    /**
+  Modify the likelihood of specified tokens appearing in the completion.
+  
+  Accepts a JSON object that maps tokens (specified by their token ID in
+  the GPT tokenizer) to an associated bias value from -100 to 100. You
+  can use this tokenizer tool to convert text to token IDs. Mathematically,
+  the bias is added to the logits generated by the model prior to sampling.
+  The exact effect will vary per model, but values between -1 and 1 should
+  decrease or increase likelihood of selection; values like -100 or 100
+  should result in a ban or exclusive selection of the relevant token.
+  
+  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
+  token from being generated.
+     */
+    logitBias?: Record<number, number>;
+    /**
+  The suffix that comes after a completion of inserted text.
+     */
+    suffix?: string;
+    /**
+  A unique identifier representing your end-user, which can help OpenAI to
+  monitor and detect abuse. Learn more.
+     */
+    user?: string;
+}
+
+type OpenAICompatibleEmbeddingModelId = string;
+interface OpenAICompatibleEmbeddingSettings {
+    /**
+  The number of dimensions the resulting output embeddings should have.
+  Only supported in text-embedding-3 and later models.
+     */
+    dimensions?: number;
+    /**
+  A unique identifier representing your end-user, which can help OpenAI to
+  monitor and detect abuse. Learn more.
+  */
+    user?: string;
+}
+
+interface OpenAICompatibleProvider<CHAT_MODEL_IDS extends string = string, COMPLETION_MODEL_IDS extends string = string, EMBEDDING_MODEL_IDS extends string = string> extends ProviderV1 {
+    (modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    languageModel(modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    chatModel(modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    completionModel(modelId: COMPLETION_MODEL_IDS, settings?: OpenAICompatibleCompletionSettings): LanguageModelV1;
+    textEmbeddingModel(modelId: EMBEDDING_MODEL_IDS, settings?: OpenAICompatibleEmbeddingSettings): EmbeddingModelV1<string>;
+}
+interface OpenAICompatibleProviderSettings {
+    /**
+     Base URL for the API calls.
+     */
+    baseURL?: string;
+    /**
+     Custom headers to include in the requests.
+     */
+    headers?: Record<string, string>;
+    /**
+     Custom fetch implementation. You can use it as a middleware to intercept requests,
+     or to provide a custom fetch implementation for e.g. testing.
+     */
+    fetch?: FetchFunction;
+    /**
+     Provider name.
+     */
+    name?: string;
+}
+/**
+Create an OpenAICompatible provider instance.
+ */
+declare function createOpenAICompatible<CHAT_MODEL_IDS extends string, COMPLETION_MODEL_IDS extends string, EMBEDDING_MODEL_IDS extends string>(options: OpenAICompatibleProviderSettings): OpenAICompatibleProvider<CHAT_MODEL_IDS, COMPLETION_MODEL_IDS, EMBEDDING_MODEL_IDS>;
+
+type OpenAICompatibleChatConfig = {
+    provider: string;
+    headers: () => Record<string, string | undefined>;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    fetch?: FetchFunction;
+    /**
+  Default object generation mode that should be used with this model when
+  no mode is specified. Should be the mode with the best results for this
+  model. `undefined` can be specified if object generation is not supported.
+    */
+    defaultObjectGenerationMode?: LanguageModelV1ObjectGenerationMode;
+};
+declare class OpenAICompatibleChatLanguageModel implements LanguageModelV1 {
+    readonly specificationVersion = "v1";
+    readonly supportsStructuredOutputs = false;
+    readonly modelId: OpenAICompatibleChatModelId;
+    readonly settings: OpenAICompatibleChatSettings;
+    private readonly config;
+    constructor(modelId: OpenAICompatibleChatModelId, settings: OpenAICompatibleChatSettings, config: OpenAICompatibleChatConfig);
+    get defaultObjectGenerationMode(): 'json' | 'tool' | undefined;
+    get provider(): string;
+    private getArgs;
+    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+}
+
+type OpenAICompatibleCompletionConfig = {
+    provider: string;
+    headers: () => Record<string, string | undefined>;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    fetch?: FetchFunction;
+};
+declare class OpenAICompatibleCompletionLanguageModel implements LanguageModelV1 {
+    readonly specificationVersion = "v1";
+    readonly defaultObjectGenerationMode: undefined;
+    readonly modelId: OpenAICompatibleCompletionModelId;
+    readonly settings: OpenAICompatibleCompletionSettings;
+    private readonly config;
+    constructor(modelId: OpenAICompatibleCompletionModelId, settings: OpenAICompatibleCompletionSettings, config: OpenAICompatibleCompletionConfig);
+    get provider(): string;
+    private getArgs;
+    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+}
+
+type OpenAICompatibleEmbeddingConfig = {
+    /**
+  Override the maximum number of embeddings per call.
+     */
+    maxEmbeddingsPerCall?: number;
+    /**
+  Override the parallelism of embedding calls.
+    */
+    supportsParallelCalls?: boolean;
+    provider: string;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    headers: () => Record<string, string | undefined>;
+    fetch?: FetchFunction;
+};
+declare class OpenAICompatibleEmbeddingModel implements EmbeddingModelV1<string> {
+    readonly specificationVersion = "v1";
+    readonly modelId: OpenAICompatibleEmbeddingModelId;
+    private readonly config;
+    private readonly settings;
+    get provider(): string;
+    get maxEmbeddingsPerCall(): number;
+    get supportsParallelCalls(): boolean;
+    constructor(modelId: OpenAICompatibleEmbeddingModelId, settings: OpenAICompatibleEmbeddingSettings, config: OpenAICompatibleEmbeddingConfig);
+    doEmbed({ values, headers, abortSignal, }: Parameters<EmbeddingModelV1<string>['doEmbed']>[0]): Promise<Awaited<ReturnType<EmbeddingModelV1<string>['doEmbed']>>>;
+}
+
+export { OpenAICompatibleChatLanguageModel, type OpenAICompatibleChatSettings, OpenAICompatibleCompletionLanguageModel, type OpenAICompatibleCompletionSettings, OpenAICompatibleEmbeddingModel, type OpenAICompatibleEmbeddingSettings, type OpenAICompatibleProvider, type OpenAICompatibleProviderSettings, createOpenAICompatible };

package/dist/index.d.ts

@@ -0,0 +1,170 @@
+import { ProviderV1, LanguageModelV1, EmbeddingModelV1, LanguageModelV1ObjectGenerationMode } from '@ai-sdk/provider';
+import { FetchFunction } from '@ai-sdk/provider-utils';
+
+type OpenAICompatibleChatModelId = string;
+interface OpenAICompatibleChatSettings {
+    /**
+  A unique identifier representing your end-user, which can help the provider to
+  monitor and detect abuse.
+    */
+    user?: string;
+}
+
+type OpenAICompatibleCompletionModelId = string;
+interface OpenAICompatibleCompletionSettings {
+    /**
+  Echo back the prompt in addition to the completion.
+     */
+    echo?: boolean;
+    /**
+  Modify the likelihood of specified tokens appearing in the completion.
+  
+  Accepts a JSON object that maps tokens (specified by their token ID in
+  the GPT tokenizer) to an associated bias value from -100 to 100. You
+  can use this tokenizer tool to convert text to token IDs. Mathematically,
+  the bias is added to the logits generated by the model prior to sampling.
+  The exact effect will vary per model, but values between -1 and 1 should
+  decrease or increase likelihood of selection; values like -100 or 100
+  should result in a ban or exclusive selection of the relevant token.
+  
+  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
+  token from being generated.
+     */
+    logitBias?: Record<number, number>;
+    /**
+  The suffix that comes after a completion of inserted text.
+     */
+    suffix?: string;
+    /**
+  A unique identifier representing your end-user, which can help OpenAI to
+  monitor and detect abuse. Learn more.
+     */
+    user?: string;
+}
+
+type OpenAICompatibleEmbeddingModelId = string;
+interface OpenAICompatibleEmbeddingSettings {
+    /**
+  The number of dimensions the resulting output embeddings should have.
+  Only supported in text-embedding-3 and later models.
+     */
+    dimensions?: number;
+    /**
+  A unique identifier representing your end-user, which can help OpenAI to
+  monitor and detect abuse. Learn more.
+  */
+    user?: string;
+}
+
+interface OpenAICompatibleProvider<CHAT_MODEL_IDS extends string = string, COMPLETION_MODEL_IDS extends string = string, EMBEDDING_MODEL_IDS extends string = string> extends ProviderV1 {
+    (modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    languageModel(modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    chatModel(modelId: CHAT_MODEL_IDS, settings?: OpenAICompatibleChatSettings): LanguageModelV1;
+    completionModel(modelId: COMPLETION_MODEL_IDS, settings?: OpenAICompatibleCompletionSettings): LanguageModelV1;
+    textEmbeddingModel(modelId: EMBEDDING_MODEL_IDS, settings?: OpenAICompatibleEmbeddingSettings): EmbeddingModelV1<string>;
+}
+interface OpenAICompatibleProviderSettings {
+    /**
+     Base URL for the API calls.
+     */
+    baseURL?: string;
+    /**
+     Custom headers to include in the requests.
+     */
+    headers?: Record<string, string>;
+    /**
+     Custom fetch implementation. You can use it as a middleware to intercept requests,
+     or to provide a custom fetch implementation for e.g. testing.
+     */
+    fetch?: FetchFunction;
+    /**
+     Provider name.
+     */
+    name?: string;
+}
+/**
+Create an OpenAICompatible provider instance.
+ */
+declare function createOpenAICompatible<CHAT_MODEL_IDS extends string, COMPLETION_MODEL_IDS extends string, EMBEDDING_MODEL_IDS extends string>(options: OpenAICompatibleProviderSettings): OpenAICompatibleProvider<CHAT_MODEL_IDS, COMPLETION_MODEL_IDS, EMBEDDING_MODEL_IDS>;
+
+type OpenAICompatibleChatConfig = {
+    provider: string;
+    headers: () => Record<string, string | undefined>;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    fetch?: FetchFunction;
+    /**
+  Default object generation mode that should be used with this model when
+  no mode is specified. Should be the mode with the best results for this
+  model. `undefined` can be specified if object generation is not supported.
+    */
+    defaultObjectGenerationMode?: LanguageModelV1ObjectGenerationMode;
+};
+declare class OpenAICompatibleChatLanguageModel implements LanguageModelV1 {
+    readonly specificationVersion = "v1";
+    readonly supportsStructuredOutputs = false;
+    readonly modelId: OpenAICompatibleChatModelId;
+    readonly settings: OpenAICompatibleChatSettings;
+    private readonly config;
+    constructor(modelId: OpenAICompatibleChatModelId, settings: OpenAICompatibleChatSettings, config: OpenAICompatibleChatConfig);
+    get defaultObjectGenerationMode(): 'json' | 'tool' | undefined;
+    get provider(): string;
+    private getArgs;
+    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+}
+
+type OpenAICompatibleCompletionConfig = {
+    provider: string;
+    headers: () => Record<string, string | undefined>;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    fetch?: FetchFunction;
+};
+declare class OpenAICompatibleCompletionLanguageModel implements LanguageModelV1 {
+    readonly specificationVersion = "v1";
+    readonly defaultObjectGenerationMode: undefined;
+    readonly modelId: OpenAICompatibleCompletionModelId;
+    readonly settings: OpenAICompatibleCompletionSettings;
+    private readonly config;
+    constructor(modelId: OpenAICompatibleCompletionModelId, settings: OpenAICompatibleCompletionSettings, config: OpenAICompatibleCompletionConfig);
+    get provider(): string;
+    private getArgs;
+    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+}
+
+type OpenAICompatibleEmbeddingConfig = {
+    /**
+  Override the maximum number of embeddings per call.
+     */
+    maxEmbeddingsPerCall?: number;
+    /**
+  Override the parallelism of embedding calls.
+    */
+    supportsParallelCalls?: boolean;
+    provider: string;
+    url: (options: {
+        modelId: string;
+        path: string;
+    }) => string;
+    headers: () => Record<string, string | undefined>;
+    fetch?: FetchFunction;
+};
+declare class OpenAICompatibleEmbeddingModel implements EmbeddingModelV1<string> {
+    readonly specificationVersion = "v1";
+    readonly modelId: OpenAICompatibleEmbeddingModelId;
+    private readonly config;
+    private readonly settings;
+    get provider(): string;
+    get maxEmbeddingsPerCall(): number;
+    get supportsParallelCalls(): boolean;
+    constructor(modelId: OpenAICompatibleEmbeddingModelId, settings: OpenAICompatibleEmbeddingSettings, config: OpenAICompatibleEmbeddingConfig);
+    doEmbed({ values, headers, abortSignal, }: Parameters<EmbeddingModelV1<string>['doEmbed']>[0]): Promise<Awaited<ReturnType<EmbeddingModelV1<string>['doEmbed']>>>;
+}
+
+export { OpenAICompatibleChatLanguageModel, type OpenAICompatibleChatSettings, OpenAICompatibleCompletionLanguageModel, type OpenAICompatibleCompletionSettings, OpenAICompatibleEmbeddingModel, type OpenAICompatibleEmbeddingSettings, type OpenAICompatibleProvider, type OpenAICompatibleProviderSettings, createOpenAICompatible };