AI接口大全

中文文档

PIKA接口

环境为python3.x，依赖版本适配python3即可

项目目录

│  api.py             //对外接口，项目启动器
│  cookie.txt				     //你的账号cookie
│  generate.py				    //发送视频生成请求
│  authorization.txt  //账号token
└─static					         //下载后的视频存储

部署

先将本项目pika下载到本地
在cookie.txt填写你的账号的cookie，在authorization.txt填写token
到api.py下发送请求

参生后的数据为一个falsk本地请求连接用于访问，自己部署falsk项目，端口为8081

本地存储在static目录下

ChatGPT接口

环境为python3.7以上，依赖版本适配python3.7以上即可

部署

先将本项目chatgpt下载到本地
在cookie.txt填写你的账号的cookie，在authorization.txt填写token
到api.py下发送请求，此项目为flask项目

@app.route('/createId',methods=['GET'])
def createId():
    conversation_id = make_new_conversation_id()
    return jsonify({'id': conversation_id})

这个是新建一个会话，返回这个唯一标识id，后面的通话就是靠这个id标识，可以创建多个

@app.route('/send', methods=['POST'])
def send():
    conversation_id = request.json.get('conversation_id')
    msg = request.json.get('msg')
    ws = sends(msg, conversation_id)
    return jsonify({'ws': ws})

这是发送消息，用前面返回的id获取，然后返回一个websocket的url通道用于接收消息，由于这个不是很稳定，所以每次前端接收一个url就要创建一个连接，接收完后断开连接，这个websocket通道没有被墙国内可接收，但是发送需要魔法

以下是前端代码，vue3为例

创建新的会话id

function frist_(){
     fetch('http://localhost:8081/createId').then(response => {
         return response.json();
     }).then(data => {
         console.log(data.id);    //返回的就是id
     })
 }

发送消息

import base64 from 'base-64'; 
const fullMessage = ref([]);

//处理websocket所接收的信息，这个逻辑还得优化一下
function processMessage(message) {
    const bodyStartIndex = message.indexOf('"body": "') + '"body": "'.length;
    const bodyEndIndex = message.indexOf('", "more_body":');
    const bodyContent = message.substring(bodyStartIndex, bodyEndIndex);
    const decodedBody = base64.decode(bodyContent);
    fullMessage.value.push(decodedBody);
    if (decodedBody.includes('[DONE]')) {
        try {
            const parsedData = JSON.parse(fullMessage.value[fullMessage.value.length - 2].substring(6));
            const partsValue = parsedData.message.content.parts[0];
            console.log(partsValue);            //解析的message也就是chatgpt所返回的消息
            connectWebSocket.close()
        } catch (e) {
            console.log(e);
        }

    }
}


//连接websocket接收消息
const connectWebSocket = (ws) => {
    const websocket = new WebSocket(ws);

    websocket.onopen = () => {
        console.log('Connected to WebSocket server');
    };

    websocket.onmessage = (event) => {
        processMessage(event.data);
    };

    websocket.onerror = (error) => {
        console.error('WebSocket encountered error: ', error);
    };

    websocket.onclose = () => {
        console.log('WebSocket connection closed');
    };

    return websocket;
};
//发送消息
function sendMessage() {
     fetch('http://localhost:8081/send', {
         method: 'POST',
         headers: {
             'Content-Type': 'application/json'
         },
         body: JSON.stringify({ 'msg': ’发送的消息‘, 'conversation_id': ’frist_()所产生id' })
     }).then(response => {
         return response.json();
     }).then(data => {
         connectWebSocket(data.ws)         //接收wss_url然后创建websocket通道接收消息
     })
 }

国内用户需要魔法代理

有什么bug欢迎issues，或者加入本群930770787

English Documentation

PIKA interface

The environment is python3.x, and the dependency version can be adapted to python3

Project directory

│  api.py           // External interface, project launcher
│  cookie.txt				  // Your account cookie
│  generate.py				 //Send a video generation request
│  authorization.txt  //Your account authorization
└─static					 //Video storage after download

Deployment

Download the pika of the project to your local computer
Fill in the cookie of your account in the cookie.txt and the token in the authorization.txt
Send the request under the test to the api.py

the data after the birth is a FALSK local request connection for access, and you can deploy the FALSK project yourself , and port is 8081

stored locally in the static directory

ChatGPT interface

The environment is Python 3.7 or later, and the version can be adapted to Python 3.7 or later

Deployment

Download the ChatGPT of this project to your local computer first
Fill in the cookie of your account in the cookie.txt and the token in the authorization.txt
Send a request to the api.py, which is a flask project

@app.route('/createId',methods=['GET'])
def createId():
    conversation_id = make_new_conversation_id()
    return jsonify({'id': conversation_id})

This is to create a new session and return this unique identifier ID, and the subsequent calls can be identified by this ID, and multiple can be created

@app.route('/send', methods=['POST'])
def send():
    conversation_id = request.json.get('conversation_id')
    msg = request.json.get('msg')
    ws = sends(msg, conversation_id)
    return jsonify({'ws': ws})

This is to send a message, get it with the id returned earlier, and then return a websocket URL channel for receiving messages, because this is not very stable, so every time the front-end receives a URL, it has to create a connection, and disconnect after receiving it

Here's the front-end code, Vue3 as an example

Create a new session ID

function frist_(){
     fetch('http://localhost:8081/createId').then(response => {
         return response.json();
     }).then(data => {
         console.log(data.id);    //The returned is the ID
     })
 }

Send messages

import base64 from 'base-64'; 
const fullMessage = ref([]);

//To handle the information received by the websocket, this logic needs to be //optimized
function processMessage(message) {
    const bodyStartIndex = message.indexOf('"body": "') + '"body": "'.length;
    const bodyEndIndex = message.indexOf('", "more_body":');
    const bodyContent = message.substring(bodyStartIndex, bodyEndIndex);
    const decodedBody = base64.decode(bodyContent);
    fullMessage.value.push(decodedBody);
    if (decodedBody.includes('[DONE]')) {
        try {
            const parsedData = JSON.parse(fullMessage.value[fullMessage.value.length - 2].substring(6));
            const partsValue = parsedData.message.content.parts[0];
            console.log(partsValue);            //The parsed message is the message returned by ChatGPT
            connectWebSocket.close()
        } catch (e) {
            console.log(e);
        }

    }
}


//Connect to a websocket to receive messages
const connectWebSocket = (ws) => {
    const websocket = new WebSocket(ws);

    websocket.onopen = () => {
        console.log('Connected to WebSocket server');
    };

    websocket.onmessage = (event) => {
        processMessage(event.data);
    };

    websocket.onerror = (error) => {
        console.error('WebSocket encountered error: ', error);
    };

    websocket.onclose = () => {
        console.log('WebSocket connection closed');
    };

    return websocket;
};
//Send a message
function sendMessage() {
     fetch('http://localhost:8081/send', {
         method: 'POST',
         headers: {
             'Content-Type': 'application/json'
         },
         body: JSON.stringify({ 'msg': 'Send a message', 'conversation_id': 'the id' })
     }).then(response => {
         return response.json();
     }).then(data => {
         connectWebSocket(data.ws)         //Receive wss_url then create a websocket channel to receive messages
     })
 }

yang520-bye / AI-interfaces-for-each-pandemic