Easy open ai下載 - Easy open ai源碼下載

易開-ai

這個儲存庫包含社群維護的 Java 中 OpenAI API 庫，這是在應用程式中使用 GPT 3/4 的最簡單方法。

概述

這個 Java 函式庫（社群維護的函式庫）提供了一種與 OpenAI 的 API 互動的便捷方式，以實現審核和聊天完成。該庫封裝了必要的細節，可以輕鬆地將 OpenAI 的強大模型整合到您的 Java 應用程式中。該項目不是由 OPENAI 維護的，這是一個非官方庫。

貢獻指南

請參考 CONTRIBUTING.md

跑步指南

所有「OPENAI_KEYS」必須透過此處定義的readKeys()函數讀取，該函數允許您同時讀取多個鍵，對於多執行緒任務，建議使用多個鍵以避免速率限制。要運行此功能，您需要在專案根資料夾中擁有keys.txt（可以隨意編輯）。

用法 - 文法

聊天完成API

若要使用聊天完成 API，請依照下列步驟操作：

 Message message = Message . builder ()
       . role ( "user" )
       . content ( "what is the capital of Cambodia?" )
       . build ();

List < Message > messages = new ArrayList <>();
messages . add ( message );

ChatCompletionRequest request = ChatCompletionRequest . builder ()
       . model ( "gpt-3.5-turbo" )
       . messages ( messages )
       . build ();

ChatCompletionResponse response = new EasyopenaiService ( new DAOImpl ()). chatCompletion ( "OPENAI_KEY" , request );

按一下此處跳轉至程式碼範例。

審核API

單一請求，單一回應

若要使用審核 API，請執行下列步驟：

 ModerationAPIRequest request = ModerationAPIRequest . builder ()
        . model ( "text-moderation-latest" )
        . input ( "hello from the other side kill me now" )
        . build ();

ModerationAPIResponse res = new EasyopenaiService ( new DAOImpl ()). getmoderation ( "OPENAI_KEY" , request );

按一下此處跳轉至程式碼範例。

易視API

Vision API 可以這樣使用，隨意在圖像列表中添加 N 個圖像-

 VisionApiResponse responseobj = new EasyVisionService (). VisionAPI ( "OPENAI_KEY" , new EasyVisionRequest ()
       . setModel ( "gpt-4-vision-preview" )
       . setPrompt ( "What are the difference between these images?" )
       . setImageUrls ( new ArrayList < String >() {{
           add ( "https://images.pexels.com/photos/268533/pexels-photo-268533.jpeg?auto=compress&cs=tinysrgb&w=1260&h=750&dpr=1" );
           add ( "https://images.pexels.com/photos/36717/amazing-animal-beautiful-beautifull.jpg?auto=compress&cs=tinysrgb&w=1260&h=750&dpr=1" );
       }}));

. setMaxtokens ( 1000 ) //Optional - for more tokens, defaults to 300

按一下此處跳轉至程式碼範例。

視覺API

若要使用 Vision API，請執行下列步驟：

使用 DAOImpl 實例初始化 EasyopenaiService 物件：

 EasyopenaiService obj = new EasyopenaiService ( new DAOImpl ());

建立 VisionApiRequest 物件：

 VisionApiRequest request = new VisionApiRequest ();

使用圖像 URL 和詳細資訊建立 ImageUrl 物件：

 ImageUrl url = new ImageUrl ();
url . setUrl ( "https://images.pexels.com/photos/18907092/pexels-photo-18907092/free-photo-of-a-photo-of-the-golden-gate-bridge-in-the-sky.jpeg" );
url . setDetail ( "low" );

建立一個 Content 物件清單來表示訊息：

 Content content1 = new Content ();
content1 . setText ( "What’s in this image?" );
content1 . setType ( "text" );

Content content2 = new Content ();
content2 . setImageUrl ( url );
content2 . setType ( "image_url" );

ArrayList < Content > listofContent = new ArrayList <>();
listofContent . add ( content1 );
listofContent . add ( content2 );

使用角色（「使用者」）和內容清單建立 MessageList 物件：

 MessageList messageList = new MessageList ();
messageList . setRole ( "user" );
messageList . setContent ( listofContent );

建立 MessageList 物件的清單：

 ArrayList < MessageList > listofMessage = new ArrayList <>();
listofMessage . add ( messageList );

設定 VisionApiRequest 物件的各種屬性：

 request . setModel ( "gpt-4-vision-preview" );
request . setMaxTokens ( 300 );
request . setMessages ( listofMessage );

進行 API 呼叫並列印回應：

 VisionApiResponse res = obj . visionAPI ( OPENAI_KEY , request );
System . out . println ( "Vision API Response is:" + res );

按一下此處跳轉至程式碼範例。

語音API

語音API可以這樣使用，隨意調整-

 SpeechRequest request = SpeechRequest . builder ()
       . model ( "tts-1" )
       . input ( "Easy OpenAI is best solution." )
       . voice ( "alloy" )
       . build ();

ResponseBody response = new EasyopenaiService ( new DAOImpl ()). createSpeech ( "OPENAI_KEY" , request );

按一下此處跳轉至程式碼範例。

輕鬆轉錄 API

轉錄API可以這樣使用，隨意調整-

 EasyTranscriptionRequest request  =  EasyTranscriptionRequest . builder ()
      . filepath ( "/Users/namankhurpia/Desktop/audio.mp3" )
      . model ( "whisper-1" )
      . build ();

ResponseBody response = new EasyTranscriptionService (). EasyTranscription ( "OPENAI_KEY" , request );

按一下此處跳轉至程式碼範例。

轉錄API

OpenAI 的原始轉錄 API

轉錄API可以這樣使用，隨意調整-

 File audioFile = new File ( "/Users/namankhurpia/Desktop/audio.mp3" );
MultipartBody . Part filePart = MultipartBody . Part . createFormData (
       "file" ,
       audioFile . getName (),
       RequestBody . create ( MediaType . parse ( "audio/*" ), audioFile )
);

RequestBody model = RequestBody . create ( MediaType . parse ( "text/plain" ), "whisper-1" );
RequestBody language = RequestBody . create ( MediaType . parse ( "text/plain" ), "" );
RequestBody prompt = RequestBody . create ( MediaType . parse ( "text/plain" ), "" );
RequestBody response_format = RequestBody . create ( MediaType . parse ( "text/plain" ), "" );
RequestBody temperature = RequestBody . create ( MediaType . parse ( "text/plain" ), "" );

ResponseBody response = new EasyopenaiService ( new DAOImpl ()). createTranscriptions ( "OPENAI_KEY" , filePart , model , language , prompt , response_format , temperature );

按一下此處跳轉至程式碼範例。

映像生成API

影像生成API可以這樣使用，隨意調整-

 ImageRequest imageRequest  = ImageRequest . builder ()
       . prompt ( "Generate a cute dog playing with frizbee" )
       . model ( "dall-e-2" )
       . build ();

ImageResponse response = new EasyopenaiService ( new DAOImpl ()). createImage ( "OPENAI_KEY" , imageRequest );

按一下此處跳轉至程式碼範例。

非同步聊天完成 API

若要非同步使用聊天完成 API，請依照下列步驟操作：

使用AsyncDAOImpl實例初始化EasyopenaiAsyncService物件：

 EasyopenaiAsyncService obj = new EasyopenaiAsyncService ( new AsyncDAOImpl ());

建立一個ChatMessage物件列表來表示對話：

 ChatMessage message = new ChatMessage ();
message . setRole ( "user" );
message . setContent ( "what is the capital of Cambodia?" );

List < ChatMessage > messages = new ArrayList <>();
messages . add ( message );

建立ChatCompletionRequest物件：

 ChatCompletionRequest request = new ChatCompletionRequest ();
request . setModel ( "gpt-3.5-turbo" );
request . setMessages ( messages ); // old conversations as well

進行非同步 API 呼叫：

 ChatCompletionResponse res = obj . getAsyncChatCompletion ( OPENAI_KEY , request );

按一下此處跳轉至程式碼範例。

非同步審核 API

若要非同步使用審核 API，請依照下列步驟操作：

建立一個ModerationAPIRequest物件：

 ModerationAPIRequest request = new ModerationAPIRequest ();
request . setInput ( "kill me now" );
request . setModel ( "text-moderation-latest" );

使用AsyncDAOImpl實例初始化EasyopenaiAsyncService物件：

 EasyopenaiAsyncService obj = new EasyopenaiAsyncService ( new AsyncDAOImpl ());

進行非同步 API 呼叫：

 ModerationAPIResponse res = obj . getASyncModeration ( OPENAI_KEY , request );

按一下此處跳轉至程式碼範例。

多執行緒異步聊天完成 API

對於使用聊天完成 API 進行多執行緒和並發調用，請按照以下步驟操作：

建立ChatCompletionRequestList物件：

 ChatCompletionRequestList list = new ChatCompletionRequestList ( new ArrayList < ChatCompletionRequest >());

將多個ChatCompletionRequest物件加入到清單中：

 // Example request 1
ChatCompletionRequest requestchat = new ChatCompletionRequest ();
requestchat . setModel ( "gpt-3.5-turbo" );
ChatMessage message = new ChatMessage ();
message . setRole ( "user" );
message . setContent ( "what is the capital of India?" );
List < ChatMessage > messages = new ArrayList <>();
messages . add ( message );
requestchat . setMessages ( messages );
list . add ( requestchat );

// Add more requests as needed (requestchat2, requestchat3, requestchat4, etc.)

進行多異步API呼叫：

 EasyopenaiConcurrentService concurrentCalls = new EasyopenaiConcurrentService ();
ChatCompletionResponseList responseList = concurrentCalls . CallMultipleChatCompletionAPI ( OPENAI_KEY , list );
System . out . println ( responseList );

按一下此處跳轉至程式碼範例。

多線程異步審核 API

對於使用審核 API 的多執行緒和並發調用，請按照以下步驟操作：

建立一個ModerationRequestList物件：

 ModerationRequestList requestList = new ModerationRequestList ( new ArrayList < ModerationAPIRequest >());

將多個ModerationAPIRequest物件加入到清單中：

 // Example request 1
ModerationAPIRequest request = new ModerationAPIRequest ();
request . setInput ( "kill me now" );
request . setModel ( "text-moderation-latest" );
requestList . add ( request );

// Add more requests as needed (request2, request3, request4, etc.)

進行多異步API呼叫：

 EasyopenaiConcurrentService concurrentCalls = new EasyopenaiConcurrentService ();
ModerationResponseList responseList = concurrentCalls . CallMultipleModerationAPI ( OPENAI_KEY , requestList );
System . out . println ( responseList );

按一下此處跳轉至程式碼範例。

EMMC EasyOpenAI 多執行緒多鍵 ChatCompletionAPI

這允許您使用多個鍵進行多個聊天完成呼叫。所有 API 呼叫都是並行執行的，但回應會被主動收集並在所有執行緒完成時發回。

CompletableFuture類別已用於實作。請參考 EasyopenaiConcurrentService.java 檔案查看原始程式碼。

用法範例-

 public static void CallMultipleChatCompletionAPI_multikey_Test ()
{
     //this function read multiple keys from keys.txt file 
     ArrayList < String > keyList = readKeys ();
     
     EasyopenaiConcurrentService concurrentCalls = new EasyopenaiConcurrentService ();
   
     ChatCompletionRequestList list = new ChatCompletionRequestList ( new ArrayList < ChatCompletionRequest >());
   
     //First thread for 
     ChatCompletionRequest requestchat = new ChatCompletionRequest ();
     requestchat . setModel ( "gpt-3.5-turbo" );
     Message message = new Message ();
     message . setRole ( "user" );
     message . setContent ( "what is the capital of India?" );
     List < Message > messages = new ArrayList <>();
     messages . add ( message );
     requestchat . setMessages ( messages );
     list . add ( requestchat );
   
   
     ChatCompletionRequest requestchat2 = new ChatCompletionRequest ();
     requestchat2 . setModel ( "gpt-3.5-turbo" );
     Message message2 = new Message ();
     message2 . setRole ( "user" );
     message2 . setContent ( "what is the capital of combodia?" );
     List < Message > messages2 = new ArrayList <>();
     messages2 . add ( message2 );
     requestchat2 . setMessages ( messages2 );
     list . add ( requestchat2 );
   
   
     ChatCompletionRequest requestchat3 = new ChatCompletionRequest ();
     requestchat3 . setModel ( "gpt-3.5-turbo" );
     Message message3 = new Message ();
     message3 . setRole ( "user" );
     message3 . setContent ( "what is the capital of new zealand?" );
     List < Message > messages3 = new ArrayList <>();
     messages3 . add ( message3 );
     requestchat3 . setMessages ( messages3 );
     list . add ( requestchat3 );
   
   
     ChatCompletionRequest requestchat4 = new ChatCompletionRequest ();
     requestchat4 . setModel ( "gpt-3.5-turbo" );
     Message message4 = new Message ();
     message4 . setRole ( "user" );
     message4 . setContent ( "what is the capital of hawaii? and what is 2+2?" );
     List < Message > messages4 = new ArrayList <>();
     messages4 . add ( message4 );
     requestchat4 . setMessages ( messages4 );
     list . add ( requestchat4 );
   
   
   
     ChatCompletionResponseList responseList = concurrentCalls . CallMultipleChatCompletionAPI ( keyList , list );
     System . out . println ( "response is" + responseList );
}

依賴關係

在使用 OpenAI API 包裝器之前，請確保已安裝所需的依賴項。

梅文

< dependency >
    < groupId >io.github.namankhurpia</ groupId >
    < artifactId >easyopenai</ artifactId >
    < version >x.x.x</ version >
</ dependency >

Groovy - Gradle

 implementation group: 'io.github.namankhurpia', name: 'easyopenai', version: 'x.x.x'

搖籃Java -

 implementation 'io.github.namankhurpia:easyopenai:x.x.x'

Gradle Kotlin-

 implementation("io.github.namankhurpia:easyopenai:x.x.x")

SBT-

 libraryDependencies += "io.github.namankhurpia" % "easyopenai" % "x.x.x"

常春藤-

 <dependency org="io.github.namankhurpia" name="easyopenai" rev="x.x.x"/>

葡萄-

 @Grapes(
@Grab(group='io.github.namankhurpia', module='easyopenai', version='x.x.x')
)

萊寧根

 [io.github.namankhurpia/easyopenai "x.x.x"]

建造者

 'io.github.namankhurpia:easyopenai:jar:x.x.x'

展開

Easy open ai

易開-ai

概述

目錄

單一實例

非同步

多重異步

多非同步多密鑰

貢獻指南

跑步指南

用法 - 文法

聊天完成API

審核API

單一請求，單一回應

易視API

視覺API

語音API

輕鬆轉錄 API

轉錄API

OpenAI 的原始轉錄 API

映像生成API

非同步聊天完成 API

非同步審核 API

多執行緒異步聊天完成 API

多線程異步審核 API

EMMC EasyOpenAI 多執行緒多鍵 ChatCompletionAPI

依賴關係

梅文

Groovy - Gradle

搖籃Java -

Gradle Kotlin-

SBT-

常春藤-

葡萄-

萊寧根

建造者