feat: http docs

client/src/constants/flow/ModuleTemplate.ts

@@ -278,7 +278,6 @@ export const AnswerModule: FlowModuleTemplateType = {
     Input_Template_TFSwitch,
     {
       key: SpecialInputKeyEnum.answerText,
-      value: '',
       type: FlowInputItemTypeEnum.textarea,
       valueType: FlowValueTypeEnum.string,
       label: '回复的内容',
@@ -331,8 +330,7 @@ export const ClassifyQuestionModule: FlowModuleTemplateType = {
       label: '系统提示词',
       description:
         '你可以添加一些特定内容的介绍，从而更好的识别用户的问题类型。这个内容通常是给模型介绍一个它不知道的内容。',
-      placeholder: '例如: \n1. Laf 是一个云函数开发平台……\n2. Sealos 是一个集群操作系统',
-      value: ''
+      placeholder: '例如: \n1. Laf 是一个云函数开发平台……\n2. Sealos 是一个集群操作系统'
     },
     Input_Template_History,
     Input_Template_UserChatInput,
@@ -393,9 +391,7 @@ export const ContextExtractModule: FlowModuleTemplateType = {
       label: '提取要求描述',
       description: '写一段提取要求，告诉 AI 需要提取哪些内容',
       required: true,
-      placeholder:
-        '例如: \n1. 你是一个实验室预约助手。根据用户问题，提取出姓名、实验室号和预约时间',
-      value: ''
+      placeholder: '例如: \n1. 你是一个实验室预约助手。根据用户问题，提取出姓名、实验室号和预约时间'
     },
     Input_Template_History,
     {

docSite/docs/flow-modules/examples/lab_assistant.md

@@ -0,0 +1 @@
+# 实验室助手

docSite/docs/flow-modules/examples/web_search_google.md

@@ -0,0 +1,70 @@
+# Google Search
+
+![](./imgs/google_search_1.png)
+![](./imgs/google_search_2.png)
+
+As shown in the above images, with the help of the HTTP module, you can easily integrate a search engine. Here, we take calling the Google Search API as an example.
+
+## Register Google Search API
+
+[Refer to this article to register the Google Search API](https://zhuanlan.zhihu.com/p/174666017)
+
+## Create a Google Search interface
+
+[Here, we use laf to quickly implement an interface, which can be written and published without deployment. Click to open laf cloud](https://laf.dev/), make sure to open the POST request method.
+
+```ts
+import cloud from '@lafjs/cloud';
+const googleSearchKey = '';
+const googleCxId = '';
+const baseurl = 'https://www.googleapis.com/customsearch/v1';
+export default async function (ctx: FunctionContext) {
+  const { searchKey } = ctx.body;
+  if (!searchKey) {
+    return {
+      prompt: ''
+    };
+  }
+  try {
+    const { data } = await cloud.fetch.get(baseurl, {
+      params: {
+        q: searchKey,
+        cx: googleCxId,
+        key: googleSearchKey,
+        c2coff: 1,
+        start: 1,
+        num: 5,
+        dateRestrict: 'm[1]'
+      }
+    });
+    const result = data.items.map((item) => item.snippet).join('\n');
+    return {
+      prompt: `Here are the search results from Google: ${result}`,
+      searchKey: `\nSearch term: ${searchKey}`
+    };
+  } catch (err) {
+    console.log(err);
+    return {
+      prompt: ''
+    };
+  }
+}
+```
+
+## Workflow
+
+Drag out a FastGPT workflow as shown in the image, where the request URL of the HTTP module is the interface address, and the input and output parameters are as follows:
+**Input**
+
+```
+searchKey: Search Key Word
+```
+
+**Output**
+
+```
+prompt: Search Result
+```
+
+- The HTTP module will send the searchKey to laf, and laf will perform a Google search based on the received input. It will then return the search results through the prompt parameter.
+- After receiving the response, the HTTP module connects to the prompt of the "AI Dialogue" to guide the model in providing an answer.

docSite/i18n/zh-Hans/docusaurus-plugin-content-docs/current/flow-modules/examples/lab_assistant.md

@@ -0,0 +1 @@
+# 实验室助手

docSite/i18n/zh-Hans/docusaurus-plugin-content-docs/current/flow-modules/examples/web_search_google.md

@@ -0,0 +1,73 @@
+# 谷歌搜索
+
+![](./imgs/google_search_1.png)
+
+![](./imgs/google_search_2.png)
+
+如上图，利用 HTTP 模块，你可以轻松的外接一个搜索引擎。这里以调用 google search api 为例。
+
+## 注册 google search api
+
+[参考这篇文章，注册 google search api](https://zhuanlan.zhihu.com/p/174666017)
+
+## 写一个 google search 接口
+
+[这里用 laf 快速实现一个接口，即写即发布，无需部署。点击打开 laf cloud](https://laf.dev/)，务必打开 POST 请求方式。
+
+```ts
+import cloud from '@lafjs/cloud';
+
+const googleSearchKey = '';
+const googleCxId = '';
+const baseurl = 'https://www.googleapis.com/customsearch/v1';
+
+export default async function (ctx: FunctionContext) {
+  const { searchKey } = ctx.body;
+
+  if (!searchKey) {
+    return {
+      prompt: ''
+    };
+  }
+
+  try {
+    const { data } = await cloud.fetch.get(baseurl, {
+      params: {
+        q: searchKey,
+        cx: googleCxId,
+        key: googleSearchKey,
+        c2coff: 1,
+        start: 1,
+        num: 5,
+        dateRestrict: 'm[1]'
+      }
+    });
+    const result = data.items.map((item) => item.snippet).join('\n');
+    return { prompt: `这是 google 搜索的结果: ${result}`, searchKey: `\n搜索词为: ${searchKey}` };
+  } catch (err) {
+    console.log(err);
+    return {
+      prompt: ''
+    };
+  }
+}
+```
+
+## 编排
+
+按上图拖出一个 FastGPT 编排组合，其中 HTTP 模块的请求地址为接口地址，出入参如下：
+
+**入参**
+
+```
+searchKey: 搜索词
+```
+
+**出参**
+
+```
+prompt: 搜索结果
+```
+
+- HTTP 模块会将 searchKey 发送到 laf，laf 接收后去进行谷歌搜索，并将搜索的结果通过 prompt 参数返回。
+- 返回后，HTTP 模块连接到【AI 对话】的提示词，引导模型进行回答。