实现 AI 回答 v2ex 帖子中的问题，可使用 deepseek 或其他与openai api对齐的LLM，会保留缓存记录到本地避免大量消耗 token。从 https://github.com/dlzmoe/scripts的v2ex AI 总结帖子脚本 修改而来 https://greasyfork.org/zh-CN/users/741071-falconchen
当前为
// ==UserScript==
// @name v2ex AI 回答问题
// @namespace https://github.com/falconchen/scripts
// @version 0.1.4
// @description 实现 AI 回答 v2ex 帖子中的问题，可使用 deepseek 或其他与openai api对齐的LLM，会保留缓存记录到本地避免大量消耗 token。从 https://github.com/dlzmoe/scripts的v2ex AI 总结帖子脚本 修改而来 https://greasyfork.org/zh-CN/users/741071-falconchen
// @author falconchen
// @match *://v2ex.com/*
// @match *://*.v2ex.com/*
// @match *://www.v2ex.com/*
// @grant GM_xmlhttpRequest
// @grant GM_registerMenuCommand
// @grant GM_unregisterMenuCommand
// @grant GM_openInTab
// @grant GM_getValue
// @grant GM_setValue
// @grant GM_notification
// @grant GM_info
// @icon data:image/png;base64,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
// @license Apache-2.0 license
// ==/UserScript==
(function () {
'use strict';
var menu_ALL = [
['menu_ManualAnswer', '是否开启手动回答 / 自动', '是否开启手动回答 / 自动', true],
];
var menu_ID = [];
for (let i = 0; i < menu_ALL.length; i++) { // 如果读取到的值为 null 就写入默认值
if (GM_getValue(menu_ALL[i][0]) == null) {
GM_setValue(menu_ALL[i][0], menu_ALL[i][3])
};
}
registerMenuCommand();
// 注册脚本菜单
function registerMenuCommand() {
if (menu_ID.length > menu_ALL.length) { // 如果菜单ID数组多于菜单数组，说明不是首次添加菜单，需要卸载所有脚本菜单
for (let i = 0; i < menu_ID.length; i++) {
GM_unregisterMenuCommand(menu_ID[i]);
}
}
for (let i = 0; i < menu_ALL.length; i++) { // 循环注册脚本菜单
menu_ALL[i][3] = GM_getValue(menu_ALL[i][0]);
menu_ID[i] = GM_registerMenuCommand(`${menu_ALL[i][3]?'✅':'❌'} ${menu_ALL[i][1]}`, function () {
menu_switch(`${menu_ALL[i][3]}`, `${menu_ALL[i][0]}`, `${menu_ALL[i][2]}`)
});
}
menu_ID[menu_ID.length] = GM_registerMenuCommand('⚙️ 设置API key配置', function () {
setApiConfig();
});
menu_ID[menu_ID.length] = GM_registerMenuCommand('💬 建议与反馈！', function () {
window.GM_openInTab("https://github.com/falconchen/scripts", {
active: true,
insert: true,
setParent: true
});
});
}
function setApiConfig(callback) {
$('body').append(`
<div class="v2exaianswer">
<input type="text" id="v2exaianswer-apikey" placeholder="sk-xxxxxxx">
<input type="text" id="v2exaianswer-baseurl" placeholder="https://api.openai.com" value="https://api.openai.com">
<input type="text" id="v2exaianswer-model" placeholder="gpt-4o-mini" value="gpt-4o-mini">
<button id="v2exaianswer-save">保存</button>
</div>
`)
$('.v2exaianswer').show();
var v2exaianswerAPI = JSON.parse(localStorage.getItem('v2exaianswerAPI')) || {
apikey: "",
baseurl: "",
model: "",
};
$('#v2exaianswer-apikey').val(v2exaianswerAPI.apikey);
$('#v2exaianswer-baseurl').val(v2exaianswerAPI.baseurl);
$('#v2exaianswer-model').val(v2exaianswerAPI.model);
// 保存
$('#v2exaianswer-save').click(function () {
v2exaianswerAPI = {
apikey: $('#v2exaianswer-apikey').val(),
baseurl: $('#v2exaianswer-baseurl').val(),
model: $('#v2exaianswer-model').val(),
}
localStorage.setItem('v2exaianswerAPI', JSON.stringify(v2exaianswerAPI));
$('.v2exaianswer').remove();
if (callback) callback();
})
}
// 菜单开关
function menu_switch(menu_status, Name, Tips) {
if (menu_status == 'true') {
GM_setValue(`${Name}`, false);
GM_notification({
text: `已关闭 [${Tips}] 功能\n（点击刷新网页后生效）`,
timeout: 3500,
onclick: function () {
location.reload();
}
});
} else {
GM_setValue(`${Name}`, true);
GM_notification({
text: `已开启 [${Tips}] 功能\n（点击刷新网页后生效）`,
timeout: 3500,
onclick: function () {
location.reload();
}
});
}
registerMenuCommand(); // 重新注册脚本菜单
};
// 返回菜单值
function menu_value(menuName) {
for (let menu of menu_ALL) {
if (menu[0] == menuName) {
return menu[3]
}
}
}
$(function () {
})
// 手动回答
function menu_ManualAnswer() {
isCache();
if (menu_value('menu_ManualAnswer')) {
// 手动回答
$('.aianswer').click(function () {
$('.aianswer').hide();
getPostContent();
})
} else {
// 自动回答
$('.aianswer').hide();
getPostContent();
}
}
if (window.location.pathname.indexOf('/t/') > -1) {
menu_ManualAnswer();
}
// 获取帖子内容
function getPostContent() {
var v2exaianswerAPI = JSON.parse(localStorage.getItem('v2exaianswerAPI'));
// 检查是否已配置API信息
if (!v2exaianswerAPI || !v2exaianswerAPI.apikey || !v2exaianswerAPI.baseurl || !v2exaianswerAPI.model) {
// 如果未配置,弹出设置窗口
setApiConfig(getPostContent);
return;
}
$('.gpt-answer-wrap').show();
return new Promise((resolve, reject) => {
const topic_title = $('h1').html();
const topic_content = $('div.topic_content').html();
const allReplies = getAllReplies();
const repliesText = allReplies.map(reply => `${reply.username} (赞赏: ${reply.likes}): ${reply.content}`).join('\n\n');
const v2exprompt = `你是一位智能AI助手。请仔细阅读以下由三重引号分隔的文本,其中包含一个问题或讨论主题以及相关回复。你的任务是:
1. 识别文本中的主要问题或讨论点
2. 考虑所有的回复,特别关注那些获得高赞赏的回复
3. 提供一个全面、有见地且有帮助的回答
4. 如果有多个问题,请逐一回答
5. 如果问题不明确,请尝试理解潜在的意图并给出最佳回答
6. 回答应该简洁、准确,并尽可能提供有用的信息或建议
7. 在回答中特别提及那些获得高赞赏的回复,解释为什么这些回复受到欢迎
8. 提及的网友ID 以 "@网友ID" 后接一个空格表示
请用简体中文回答。不要重复问题,直接给出回答。
"""
标题: ${topic_title}
内容: ${topic_content}
回复 (按赞赏数降序排列):
${repliesText}
"""
`;
console.log(v2exprompt);
fetch(`${v2exaianswerAPI.baseurl}/v1/chat/completions`, {
method: "POST",
headers: {
"Content-Type": "application/json",
Authorization: `Bearer ${v2exaianswerAPI.apikey}`,
},
body: JSON.stringify({
model: v2exaianswerAPI.model,
messages: [{
role: "user",
content: v2exprompt,
}],
temperature: 0.7,
}),
})
.then(response => {
if (!response.ok) {
reject(`HTTP error! status: ${response.status}`);
}
return response.json();
})
.then(gptData => {
$(".gpt-answer").html(`AI 回答：<br>${gptData.choices[0].message.content.replace(/\n/g, '<br>')}`);
$('.airegenerate').show();
let v2exaianswerdata =
JSON.parse(localStorage.getItem("v2exaianswerdata")) || [];
const match = window.location.pathname;
let existingObject = v2exaianswerdata.find((item) => item.name == match);
let newObject = {
name: match,
value: gptData.choices[0].message.content,
};
if (existingObject) {
existingObject.value = newObject.value;
} else {
v2exaianswerdata.push(newObject);
}
// 将帖子回答的数据缓存
localStorage.setItem("v2exaianswerdata", JSON.stringify(v2exaianswerdata));
resolve();
})
.catch(error => {
$(".gpt-answer").html(`抱歉生成失败，请检查配置或者反馈给开发者！`);
$('.airegenerate').show();
});
});
}
// 先判断是否有缓存
function isCache() {
$("#Main .box>.header").after(`<button type="button" class="aianswer">AI回答</button>`);
$("#Main .box>.header").after(
`<div class="gpt-answer-wrap">
<div class="gpt-answer">AI 回答：正在使用 AI 生成回答中，请稍后...</div>
<button type="button" class="airegenerate" style="display:none">重新生成</button>
</div>`
);
let v2exaianswerdata = JSON.parse(localStorage.getItem("v2exaianswerdata")) || [];
const match = window.location.pathname;
let existingObject = v2exaianswerdata.find((item) => item.name === match);
if (existingObject) {
// 存在缓存，拿旧数据
$('.gpt-answer-wrap').show();
$(".gpt-answer").html(`AI 回答：<br>${existingObject.value.replace(/\n/g, '<br>')}`);
$('.airegenerate').show();
$('.aianswer').hide();
} else {
$('.gpt-answer-wrap').hide();
if (!menu_value('menu_ManualAnswer')) {
getPostContent();
}
}
$('.airegenerate').click(() => {
$('.gpt-answer').html(`AI 回答：正在使用 AI 生成回答中，请稍后...`)
$('.airegenerate').hide();
getPostContent();
})
}
function getAllReplies() {
const replies = [];
$('div[id^="r_"]').each(function() {
const $reply = $(this);
const username = $reply.find('a[href^="/member"]').text();
const content = $reply.find('div.reply_content').text();
// 获取赞赏数
let likes = 0;
const $likeSpan = $reply.find('span.small.fade img[src*="heart_neue_red.png"]').parent();
if ($likeSpan.length > 0) {
likes = parseInt($likeSpan.text().trim(), 10) || 0;
}
replies.push({
username: username,
content: content,
likes: likes
});
});
// 按赞赏数降序排序
replies.sort((a, b) => b.likes - a.likes);
return replies;
}
// 使用示例
const allReplies = getAllReplies();
console.log(allReplies);
$('body').append(`<style>.gpt-answer-wrap{background:#fffbd9;border-radius:5px;padding:10px;font-size:14px;color:#303030;margin:0;line-height:1.6;text-align:left}.aianswer{display:flex;outline:0;border:1px solid #eee;background:#ffe27d;color:#626262;padding:4px 10px;cursor:pointer;border-radius:3px}.gpt-answer-wrap .airegenerate{margin-top:6px;outline:0;border:1px solid #eee;background:#ffe27d;color:#626262;padding:4px 10px;cursor:pointer;border-radius:3px}.v2exaianswer{position:fixed;bottom:20px;right:20px;z-index:99999;max-width:400px;padding:20px;border:1px solid #ddd;border-radius:8px;box-shadow:0 2px 10px rgba(0,0,0,.1);background-color:#f9f9f9;display:none}.v2exaianswer input[type=text]{width:100%;padding:10px;margin:10px 0;border:1px solid #ccc;border-radius:4px;font-size:16px;transition:border-color .3s}.v2exaianswer input[type=text]:focus{border-color:#007bff;outline:0}.v2exaianswer button{width:100%;padding:10px;background-color:#007bff;color:#fff;border:none;border-radius:4px;font-size:16px;cursor:pointer;transition:background-color .3s}.v2exaianswer button:hover{background-color:#0056b3}.gpt-answer {
white-space: pre-line;
}</style>`)
})();