你当前正在访问 Microsoft Azure Global Edition 技术文档网站。 如果需要访问由世纪互联运营的 Microsoft Azure 中国技术文档网站,请访问 https://docs.azure.cn

OpenAIClient class

用于与 Azure OpenAI 交互的客户端。

客户端需要 OpenAI 资源的终结点和身份验证方法,例如 API 密钥或令牌。 可以在 OpenAI 资源页中找到 API 密钥和终结点。 它们将位于资源的“密钥和终结点”页中。

身份验证示例:

API 密钥

import { OpenAIClient } from "@azure/openai";
import { AzureKeyCredential } from "@azure/core-auth";

const endpoint = "<azure endpoint>";
const credential = new AzureKeyCredential("<api key>");

const client = new OpenAIClient(endpoint, credential);

Azure Active Directory

import { OpenAIClient } from "@azure/openai";
import { DefaultAzureCredential } from "@azure/identity";

const endpoint = "<azure endpoint>";
const credential = new DefaultAzureCredential();

const client = new OpenAIClient(endpoint, credential);

构造函数

OpenAIClient(KeyCredential, OpenAIClientOptions)

初始化 OpenAIClient 实例,以便与非 Azure OpenAI 终结点一起使用。

OpenAIClient(string, KeyCredential, OpenAIClientOptions)

初始化 OpenAIClient 的实例,以便与 Azure OpenAI 资源一起使用。

OpenAIClient(string, TokenCredential, OpenAIClientOptions)

初始化 OpenAIClient 的实例,以便与 Azure OpenAI 资源一起使用。

方法

getAudioTranscription(string, Uint8Array, GetAudioTranscriptionOptions)

以简单的 JSON 格式返回音频文件的听录。

getAudioTranscription<Format>(string, Uint8Array, Format, GetAudioTranscriptionOptions)

返回音频文件的听录。

getAudioTranslation(string, Uint8Array, GetAudioTranslationOptions)

返回音频文件的翻译。

getAudioTranslation<Format>(string, Uint8Array, Format, GetAudioTranslationOptions)

返回音频文件的翻译。

getChatCompletions(string, ChatRequestMessageUnion[], GetChatCompletionsOptions)

获取提供的聊天消息的聊天完成。 完成支持各种任务,并生成继续或“完成”提供的提示数据的文本。

getCompletions(string, string[], GetCompletionsOptions)

获取提供的输入提示的完成。 完成支持各种任务,并生成继续或“完成”提供的提示数据的文本。

getEmbeddings(string, string[], GetEmbeddingsOptions)

返回给定提示的嵌入。

getImages(string, string, GetImagesOptions)

在出现提示的情况下创建映像。

streamChatCompletions(string, ChatRequestMessageUnion[], GetChatCompletionsOptions)

Lists聊天完成令牌可用于聊天上下文。

streamCompletions(string, string[], GetCompletionsOptions)

Lists完成令牌,因为它们在给定提示下可用。

构造函数详细信息

OpenAIClient(KeyCredential, OpenAIClientOptions)

初始化 OpenAIClient 实例,以便与非 Azure OpenAI 终结点一起使用。

new OpenAIClient(openAiApiKey: KeyCredential, options?: OpenAIClientOptions)

参数

openAiApiKey
KeyCredential

连接到非 Azure OpenAI 终结点时要使用的 API 密钥。

options
OpenAIClientOptions

用于配置客户端的选项。

注解

使用此构造函数初始化的 OpenAIClient 对象只能与非 Azure OpenAI 推理终结点一起使用。 若要将 OpenAIClient 与 Azure OpenAI 资源配合使用,请改用接受资源 URI 和 Azure 身份验证凭据的构造函数。

OpenAIClient(string, KeyCredential, OpenAIClientOptions)

初始化 OpenAIClient 的实例,以便与 Azure OpenAI 资源一起使用。

new OpenAIClient(endpoint: string, credential: KeyCredential, options?: OpenAIClientOptions)

参数

endpoint

string

Azure OpenAI 资源的 URI,包括协议和主机名。 例如:https://my-resource.openai.azure.com

credential
KeyCredential

用于对 Azure OpenAI 资源进行身份验证的密钥凭据。

options
OpenAIClientOptions

用于配置客户端的选项。

注解

此构造函数初始化只能与 Azure OpenAI 资源一起使用的 OpenAIClient 对象。 若要对非 Azure OpenAI 推理终结点使用 OpenAIClient,请改用接受非 Azure OpenAI API 密钥的构造函数。

OpenAIClient(string, TokenCredential, OpenAIClientOptions)

初始化 OpenAIClient 的实例,以便与 Azure OpenAI 资源一起使用。

new OpenAIClient(endpoint: string, credential: TokenCredential, options?: OpenAIClientOptions)

参数

endpoint

string

Azure OpenAI 资源的 URI,包括协议和主机名。 例如:https://my-resource.openai.azure.com

credential
TokenCredential

用于对 Azure OpenAI 资源进行身份验证的令牌凭据。

options
OpenAIClientOptions

用于配置客户端的选项。

方法详细信息

getAudioTranscription(string, Uint8Array, GetAudioTranscriptionOptions)

以简单的 JSON 格式返回音频文件的听录。

function getAudioTranscription(deploymentName: string, fileContent: Uint8Array, options?: GetAudioTranscriptionOptions): Promise<AudioResultSimpleJson>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

fileContent

Uint8Array

要听录的音频文件的内容。

options
GetAudioTranscriptionOptions

此音频听录请求的选项。

返回

音频听录结果为简单的 JSON 格式。

getAudioTranscription<Format>(string, Uint8Array, Format, GetAudioTranscriptionOptions)

返回音频文件的听录。

function getAudioTranscription<Format>(deploymentName: string, fileContent: Uint8Array, format: Format, options?: GetAudioTranscriptionOptions): Promise<AudioResult<Format>>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

fileContent

Uint8Array

要听录的音频文件的内容。

format

Format

结果对象的格式。 有关可能的值,请参阅 AudioResultFormat

options
GetAudioTranscriptionOptions

此音频听录请求的选项。

返回

Promise<AudioResult<Format>>

音频听录会生成你选择的格式。

getAudioTranslation(string, Uint8Array, GetAudioTranslationOptions)

返回音频文件的翻译。

function getAudioTranslation(deploymentName: string, fileContent: Uint8Array, options?: GetAudioTranslationOptions): Promise<AudioResultSimpleJson>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

fileContent

Uint8Array

要翻译的音频文件的内容。

options
GetAudioTranslationOptions

此音频翻译请求的选项。

返回

音频翻译结果。

getAudioTranslation<Format>(string, Uint8Array, Format, GetAudioTranslationOptions)

返回音频文件的翻译。

function getAudioTranslation<Format>(deploymentName: string, fileContent: Uint8Array, format: Format, options?: GetAudioTranslationOptions): Promise<AudioResult<Format>>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

fileContent

Uint8Array

要翻译的音频文件的内容。

format

Format

结果对象的格式。 有关可能的值,请参阅 AudioResultFormat

options
GetAudioTranslationOptions

此音频翻译请求的选项。

返回

Promise<AudioResult<Format>>

音频翻译结果。

getChatCompletions(string, ChatRequestMessageUnion[], GetChatCompletionsOptions)

获取提供的聊天消息的聊天完成。 完成支持各种任务,并生成继续或“完成”提供的提示数据的文本。

function getChatCompletions(deploymentName: string, messages: ChatRequestMessageUnion[], options?: GetChatCompletionsOptions): Promise<ChatCompletions>

参数

deploymentName

string

返回

Promise<ChatCompletions>

getCompletions(string, string[], GetCompletionsOptions)

获取提供的输入提示的完成。 完成支持各种任务,并生成继续或“完成”提供的提示数据的文本。

function getCompletions(deploymentName: string, prompt: string[], options?: GetCompletionsOptions): Promise<Completions>

参数

deploymentName

string

prompt

string[]

返回

Promise<Completions>

getEmbeddings(string, string[], GetEmbeddingsOptions)

返回给定提示的嵌入。

function getEmbeddings(deploymentName: string, input: string[], options?: GetEmbeddingsOptions): Promise<Embeddings>

参数

deploymentName

string

input

string[]

返回

Promise<Embeddings>

getImages(string, string, GetImagesOptions)

在出现提示的情况下创建映像。

function getImages(deploymentName: string, prompt: string, options?: GetImagesOptions): Promise<ImageGenerations>

参数

deploymentName

string

prompt

string

返回

Promise<ImageGenerations>

streamChatCompletions(string, ChatRequestMessageUnion[], GetChatCompletionsOptions)

Lists聊天完成令牌可用于聊天上下文。

function streamChatCompletions(deploymentName: string, messages: ChatRequestMessageUnion[], options?: GetChatCompletionsOptions): Promise<EventStream<ChatCompletions>>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

messages

ChatRequestMessageUnion[]

要用于此请求的聊天上下文消息。

options
GetChatCompletionsOptions

此聊天完成请求的聊天完成选项。

返回

聊天完成令牌的异步迭代。

streamCompletions(string, string[], GetCompletionsOptions)

Lists完成令牌,因为它们在给定提示下可用。

function streamCompletions(deploymentName: string, prompt: string[], options?: GetCompletionsOptions): Promise<EventStream<Omit<Completions, "usage">>>

参数

deploymentName

string

使用 Azure OpenAI) 时 (模型部署的名称;使用非 Azure OpenAI) 用于此请求时 (模型名称。

prompt

string[]

用于此请求的提示。

options
GetCompletionsOptions

此完成请求的完成选项。

返回

Promise<EventStream<Omit<Completions, "usage">>>

完成令牌的异步迭代。