{"id":23201,"date":"2024-11-16T10:20:18","date_gmt":"2024-11-16T02:20:18","guid":{"rendered":"https:\/\/www.1ai.net\/?p=23201"},"modified":"2024-11-16T10:23:55","modified_gmt":"2024-11-16T02:23:55","slug":"ai%e5%b0%8f%e7%a8%8b%e5%ba%8f%e6%90%ad%e5%bb%ba%ef%bc%8c%e5%a6%82%e4%bd%95%e5%bf%ab%e9%80%9f%e5%bc%80%e5%8f%91%e4%b8%80%e6%ac%beai%e5%be%ae%e4%bf%a1%e5%b0%8f%e7%a8%8b%e5%ba%8f","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/23201.html","title":{"rendered":"ai small program build, how to quickly develop an AI weibo small program"},"content":{"rendered":"

In this AI era, combined with AI to make new products or based on your first business to join the AI capabilities, this is where the trend is, so how to access the AI capabilities faster? Next to share with you the integration method of cloud development AI+.<\/p>\n

move<\/strong><\/p>\n

Log in to the WeChat Cloud development backend
\nhttps:\/\/tcb.cloud.tencent.com\/cloud-admin?_tcbProviderId=mp<\/p>\n

\"de87306dj00sn0tz70023d000qm012cm\"<\/p>\n

Then find the AI+ menu<\/p>\n

\"2be293aej00sn0tz7000dd000u00058m\"<\/p>\n

There are two pieces in total:<\/p>\n

1.AI Big Model: basic AI dialog capability suitable for quick access to big models<\/p>\n

2. Customized AI capabilities: AI intelligences applicable to specific businesses are built to support the knowledge base<\/p>\n

Access to AI Big Models<\/strong><\/p>\n

Let's start our experience with the AI Big Model by going to the AI Big Model module<\/p>\n

\"a77f1713j00sn0tz7000xd000u000gxm\"<\/p>\n

Support access to a number of large domestic models, just fill in the Key can be, each large model service background access to the way you can go to see the documentation of each.<\/p>\n

The big model I use here is Wisdom Spectrum, after inputting the key and clicking save in the upper right corner, then the relevant model on the left will have an \"enabled\" status, and it is configured at this time.<\/p>\n

I'll take it from here.Mini Programs<\/a>SDK as an example:<\/p>\n

Step 1: Install the npm package first<\/strong><\/p>\n

Execute the command install command in the directory where the applet package.json is located (usually the miniprogram directory):<\/p>\n

npm i @cloudbase\/js-sdk@2.8.3-beta.0<\/code><\/p>\n

Step 2: Initialize the ai object<\/strong><\/p>\n

const env = \"Fill in your environment ID\";\r\nregisterAuth(cloudbase);\r\nregisterAi(cloudbase);\r\nconst app = cloudbase.init({ env });\r\nconst auth = app.auth({ persistence: \"local\" }); await auth.signInWith(cloudbase); registerAi(cloudbase); registerAi(cloudbase)\r\nawait auth.signInWithOpenId(); \/\/ or use another login method\r\nconst ai = await app.ai(); \/\/ Next you can call ai.\r\n\/\/ Next, you can call the methods provided by the ai module to create the model and so on.\r\n<\/code><\/pre>\n
Then you'll see the console report an error<\/p>\n
\"dd1f2812j00sn0tzt001ed000u0009qm\"<\/p>\n
Because the SDK has a request for the domain name need to be configured in the background of the small program can be or can be in the local settings to check the non-check.<\/p>\n
\"421162cbj00sn0tzt002id000m00152m\"<\/p>\n
Note: Check the box does not check the purpose is to be able to experience the use, if you want to publish online or to configure the domain name configuration in the background<\/p>\n
Step 3: Use AI Big Models to Recommend Streaming Calls<\/strong><\/p>\n
 const aiModel = ai.createModel(\"zhipu\"); \/\/ create the model\r\n    const res = await aiModel.streamText({\r\n      model: \"glm-4-plus\",\r\n      messages: [{\r\n        role: \"user\",\r\n        content: \"Hello, please introduce Li Bai\"\r\n      }, ]\r\n    }); for await (let str of res.textStream)\r\n    for await (let str of res.textStream) {\r\n      console.log(str);\r\n    }\r\n<\/code><\/pre>\n
The call succeeded.<\/p>\n
\"39a60027j00sn0tzt001fd000u000dkm\"<\/p>\n
Access to AI Agent<\/strong><\/p>\n
The first step is to create an Agent Intelligence<\/p>\n
\"5ee00e0fj00sn0tzt000fd000u000dgm\"<\/p>\n
There are some Agent templates built-in by default, you can directly use the templates or innovate a new Agent, enter the name and introduction on the right side and click \"New\" on the bottom right corner.<\/p>\n
\"f6384344j00sn0tzt002kd000u000n6m\"<\/p>\n
You will then be taken to the edit details page, which is divided into three areas, from left to right<\/p>\n
    \n
  1. Agent list area, showing all Agents and the ability to switch between them<\/li>\n
  2. Orchestrate Agent area, support setting up persona and response logic as well as adding knowledge base, openers, questions, question suggestions<\/li>\n
  3. Agent preview debugging area, you can debug the effect in this area after setting up the Agent.<\/li>\n<\/ol>\n
    \"a3a070fej00sn0tzt000td000u000frm\"<\/p>\n
    Let's take a look at the Choreography Agent area, \"Persona and Response Constraints\", where you can click the \"Generate\" button on the right to automatically generate prompts.<\/p>\n
    \"598a1fefj00sn0tzt0029d000u000ukm\"<\/p>\n
    Click on the use of the application can be applied, after the completion of the setup can be in the right area of the dialog test effect<\/p>\n
    \"a1df7882j00sn0tzt001gd000u000ftm\"<\/p>\n
    You can also add a knowledge base for this intelligence, click [Manage Knowledge Base].<\/p>\n
    \"9cc145aaj00sn0tzt000gd000u000d5m\"<\/p>\n
    Go to New to enter a title and description<\/p>\n
    \"68d001c6j00sn0tzt000od000u000i0m\"<\/p>\n
    A folder will appear after success<\/p>\n
    \"aaaf6b1dj00sn0tzt000qd000u000i1m\"<\/p>\n
    Then add the file<\/p>\n
    \"012af30fj00sn0tzt000ed000u0007tm\"<\/p>\n
    Support for single\/multiple add files<\/p>\n
    \"d5a8261dj00sn0tzt0008d000aa0086m\"<\/p>\n
    Let's say I add a document on common car trouble problems<\/p>\n
    \"8ca99fc3j00sn0tzt004ed000u0010rm\"<\/p>\n
    After adding the file, the status will change from \"File Parsing in Progress\" to \"File Parsing and Writing Completed\".<\/p>\n
    \"239b1845j00sn0tzt000fd000u0006um\"<\/p>\n
    Then just go back to the edit page to add the knowledge base folder, and all that's left is some basic configuration of openers and preset questions.<\/p>\n
    \"e0281396j00sn0tzt000ud000u000gem\"<\/p>\n
    Next, let's test the effect of the knowledge base after the completion of the setup, when I sent in the document just mentioned the \"open the door creaking noise\" problem.<\/p>\n
    \"efee6e69j00sn0tzt001qd000ke00m8m\"<\/p>\n
    It will give the source of the cited knowledge base at the end of the answer, which will go a long way towards avoiding the problem of the big model's hallucination on specific content.<\/p>\n
    \"ea6aabb5j00sn0tzt001id000kc00l4m\"<\/p>\n
    Once we're all configured then we can go ahead and use it, clicking on the Access Guidelines in the upper right hand corner will call out the release document.<\/p>\n
    release<\/strong><\/p>\n
    There are three ways to do this:
    \n1. Docking WeChat platform
    \n2. Small program block components
    \n3.SDK call<\/p>\n
    Recommended to use these two ways to docking WeChat platform and SDK calls, docking WeChat platform completely no code is more simple, SDK calls pure code way to call more flexible, if your application scenario is not a dialog mode then you can use the code of the way to implant your business flow.<\/p>\n
    Docking WeChat platform<\/strong>
    \nSupport small program customer service, WeChat customer service, WeChat service number, WeChat subscription number<\/p>\n
    \"9c4b0bc1j00sn0tzt0022d000po014mm\"<\/p>\n
    The configuration process is very simple, just follow the steps to set the AppID and then scan the code for authorization.<\/p>\n
    Applet Block Components<\/strong>
    \nNeed to switch to My Apps from the main menu to create a blank app<\/p>\n
    \"36f2653dj00sn0tzt000xd000k000kkm\"<\/p>\n
    \"4ab97486j00sn0tzt000zd000u000ejm\"<\/p>\n
    Then you'll be taken to the microtransmitter edit page, select the right-hand block and search for \"Agent\".<\/p>\n
    \"7861e790j00sn0tzt000rd000u000eqm\"<\/p>\n
    Configure ID after choosing to play AgentUI<\/p>\n
    \"28e36dbaj00sn0tzt0013d000u000fam\"<\/p>\n
    ID has a copy ID function in the Agent detail name.<\/p>\n
    \"1dfc37c8j00sn0tzt000rd000u000a3m\"<\/p>\n
    Once the ID is configured the data will be synchronized to this UI component and you can modify its properties, events, styles, etc.<\/p>\n
    \"97b94f50j00sn0tzt0013d000u000eim\"<\/p>\n
    After all the modifications are completed we can select the Agent-UI component and switch to the configuration where we can directly download the code package<\/p>\n
    \"fea2001dj00sn0tzt000wd000u000b8m\"<\/p>\n
    Then you can integrate it into your applet according to the follow up documentation<\/p>\n
    \"5e24ac9dj00sn0tzt002md000u000zbm\"<\/p>\n
    SDK calls:<\/p>\n
      \n
    1. Start by establishing a connection.<\/li>\n<\/ol>\n
      \/\/ Before you start, make sure you have configured your miniprogram request with a legitimate domain name, see https:\/\/docs.cloudbase.net\/quick-start\/baas\/wx-mini for details.\r\n\/\/ Install the npm package by running the command in the directory where the applet package.json is located (usually the miniprogram directory):\r\n\/\/ npm i @cloudbase\/js-sdk@2.8.3-beta.0\r\n\/\/ After installation, click \"Tools -> Build npm\" in the menu bar of WeChat Developer Tools.\r\n\r\n\/\/ Introduce the SDK. Here, we've introduced cloudbase-js-sdk in modules as needed, or in its entirety\r\nimport cloudbase from \"@cloudbase\/js-sdk\/app\";\r\nimport { registerAuth } from \"@cloudbase\/js-sdk\/auth\";\r\nimport { registerAi } from \"@cloudbase\/js-sdk\/ai\";\r\n\r\nregisterAuth(cloudbase).\r\nregisterAi(cloudbase); const app = cloudbase.init\r\nconst app = cloudbase.init({\r\n  env: \"your-env\", \/\/ need to be replaced by the actual environment id\r\n}); const auth = app.\r\nconst auth = app.auth({ persistence: \"local\" }); await auth.signInWith\r\nawait auth.signInWithOpenId(); \/\/ Or use another login method\r\nconst ai = await app.ai(); \/\/ Next you can call ai.\r\n\/\/ You can then call the methods provided by the ai module<\/code><\/pre>\n
        \n
      1. Specify a specific Agent dialog<\/li>\n<\/ol>\n
        const res = await ai.bot.sendMessage({\r\n  botId: \"botId-xxx\",\r\n  msg: \"Give me an idiom.\" ,\r\n});\r\n\r\nfor await (let str of res.textStream) {\r\n  console.log(str); }\r\n}\r\n\/\/ Of course\r\n\/\/ Yes, you can!\r\n\/\/ !\r\n\/\/ Here\r\n\/\/ Here's one\r\n\/\/ Idioms\r\n<\/code><\/pre>\n
          \n
        1. View Chat Logs<\/li>\n<\/ol>\n
          const res = await ai.bot.getChatRecords({\r\n  botId: \"botId-xxx\", pageNumber: 1, pageNumber.getChatRecords({\r\n  botId: \"botId-xxx\", pageNumber: 1,\r\n  pageNumber: 1, pageSize: 10,\r\n  sort: \"asc\", }); { botId: \"botId-xxx\r\n});\r\n\/\/ {\r\n\/\/ \"recordList\": [\r\n\/\/ {\r\n\/\/ \"botId\": \"bot-809d4ad1\", \/\/ \"recordId\": \"record-96617446\", \/\/ \"recordList\": [ \/\/ {\r\n\r\n\/\/ \"role\": \"user\", \/\/ \"content\": \"user\", \/\/ \"content\": \"user\", \"user\", \"user\", \"user\", \"user\".\r\n\r\n\/\/ \"conversation\": \"r3Hjz3qgms3UG7z9lwSYbA\", \/\/ \"type\": \"text\".\r\n\r\n\/\/ \"triggerSrc\": \"TCB\"\r\n\/\/ }, \/\/ { \"triggerSrc\": \"TCB\", \/\/ \"triggerSrc\".\r\n\/\/ {\r\n\/\/ \"botId\": \"bot-809d4ad1\", \/\/ \"recordId\": \"bot-809d4ad1\", {\r\n\r\n\r\n\/\/ \"content\": \"Hello! It looks like you may be experiencing some technical issues, or just want to say hello. In either case, I'm here to help. If you have any specific questions or need assistance, please feel free to let me know!\" ,\r\n\/\/ \"conversation\": \"r3Hjz3qgms3UG7z9lwSYbA\",\r\n\/\/ \"type\": \"text\", \/\/ \"triggerSrc\".\r\n\/\/ \"triggerSrc\": \"TCB\"\r\n\/\/ }\r\n\/\/ ], \/\/ \"total\": 2\r\n\/\/ \"total\": 2\r\n\/\/ }\r\n<\/code><\/pre>\n
           <\/p>\n
          Here the whole process of accessing AI is all over.<\/p>\n
          at last<\/strong><\/p>\n
          I hope you can keep up with the AI era and make your own AI applets.<\/p>","protected":false},"excerpt":{"rendered":"
          In this AI era, combined with AI to make new products or based on your first business to join the AI capabilities, this is where the trend is, so how to access the AI capabilities faster? Next to share with you the integration method of cloud development AI+. Steps Login to the WeChat cloud development background https:\/\/tcb.cloud.tencent.com\/cloud-admin?_tcbProviderId=mp and then find the AI+ menu A total of two pieces: 1. AI big model: applies to rapid access to the basic AI dialog capabilities of the big model 2. custom AI capabilities: applies to specific business Custom AI capabilities: for business-specific AI intelligent body building, support for knowledge base Access to the AI big model Let's start with the AI big model experience, enter the AI big model module support<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[4949,1690,3905],"collection":[],"class_list":["post-23201","post","type-post","status-publish","format-standard","hentry","category-jiaocheng","category-baike","tag-ai","tag-1690","tag-3905"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/23201","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=23201"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/23201\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=23201"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=23201"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=23201"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=23201"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}