Botkit with QnA Integration keeps repeating itself in Slack

[mattw114862]

Integrating QnA Maker with Slack through botkit seems to cause an issue where the response from the bot keeps repeating itself. I am unsure of how to resolve this. I thought maybe it was reading the messages of itself and responding back but I am not so sure. I also have the same integration for teams and twilio both which work fine.

// `require('dotenv').config();

const express = require("express");
const bodyParser = require("body-parser");
// Import Botkit's core features
const { Botkit } = require("botkit");
// Import a platform-specific adapter for slack.
const { SlackAdapter, SlackEventMiddleware, SlackMessageTypeMiddleware } = require('botbuilder-adapter-slack');
const { BotFrameworkAdapter } = require('botbuilder');
const { TwilioAdapter } = require('botbuilder-adapter-twilio-sms');
const { QnAMaker } = require('botbuilder-ai');

// Create an express server with json and urlencoded body parsing
const webserver = express();
webserver.use((req, res, next) => {
    req.rawBody = '';
    req.on('data', function (chunk) {
        req.rawBody += chunk;
    });
    next();
});
webserver.use(express.json());
webserver.use(express.urlencoded({ extended: true }));

 //Create Bot Framework Adapter.
// See https://aka.ms/about-bot-adapter to learn more about adapters and how bots work.
const botAdapter = new BotFrameworkAdapter({
    appId: process.env.MicrosoftAppId,
   appPassword: process.env.MicrosoftAppPassword
});

// Adaptor for Slack
const slackAdaptor = new SlackAdapter({
    // REMOVE THIS OPTION AFTER YOU HAVE CONFIGURED YOUR APP!
    //enable_incomplete: true,
    // The following options are sufficient for a single team application
    // parameters used to secure webhook endpoint
    verificationToken: process.env.SLACK_VERIFICATION_TOKEN,
    clientSigningSecret: process.env.SLACK_CLIENT_SIGNING_SECRET,
    // auth token for a single-team app
    botToken: process.env.SLACK_BOT_TOKEN
});

// QnA Integration
const qnaMaker = new QnAMaker({
    knowledgeBaseId: process.env.QnAKnowledgebaseId,
    endpointKey: process.env.QnAAuthKey,
    host: process.env.QnAEndpointHostName
});

// Use Slack Middleware to emit events that match their original Slack event types.
slackAdaptor.use(new SlackMessageTypeMiddleware());
slackAdaptor.use(new SlackEventMiddleware());

// Controller for Slack
const slackController = new Botkit({
    webhook_uri: process.env.SLACK_WEBHOOK_URI,
    adapter: slackAdaptor,
    webserver: webserver
});

// Controller for Teams
const teamsController = new Botkit({
    webhook_uri: process.env.TEAMS_WEBHOOK_URI,
    adapterConfig: {
        appId: process.env.TEAMS_APP_ID,
        appPassword: process.env.TEAMS_APP_PASSWORD,
    },
    webserver: webserver
});

// QnA Integration for Teams
teamsController.middleware.ingest.use(async (bot, message, next) => {
    if (message.incoming_message.type === 'message') {
        const qnaResults = await qnaMaker.getAnswers(message.context);
        if (qnaResults[0]) {
            // If we got an answer, send it directly
            //await message.context.sendActivity(qnaResults[0].answer);
            await bot.reply(message, qnaResults[0].answer); // also works
        } else {
            // If we have no other features, we could just say we didn't find any answers:
            //await message.context.sendActivity('No QnA Maker answers were found.');
            // Otherwise, just forward to the next BotHandler to see if there are any other matches
            next();
        }
    } else {
        next();
    }
});

// QnA Integration for Slack
slackController.middleware.ingest.use(async (bot, message, next) => {
    if (message.incoming_message.type === 'message') {
        const qnaResults = await qnaMaker.getAnswers(message.context);
        if (qnaResults[0]) {
            // If we got an answer, send it directly
            //await message.context.sendActivity(qnaResults[0].answer);
            await bot.reply(message, qnaResults[0].answer); // also works
        } else {
            // If we have no other features, we could just say we didn't find any answers:
            //await message.context.sendActivity('No QnA Maker answers were found.');
            // Otherwise, just forward to the next BotHandler to see if there are any other matches
           next();
        }
    } else {
      next();
    }
});

//Adapter and Controller for Twilio
const smsadapter = new TwilioAdapter({
    twilio_number: process.env.TWILIO_NUMBER,
    account_sid: process.env.TWILIO_ACCOUNT_SID,
    auth_token: process.env.TWILIO_AUTH_TOKEN,
});

const smsController = new Botkit({
    adapter: smsadapter,
    webserver: webserver
});

// QnA Integration for Twilio
smsController.middleware.ingest.use(async (bot, message, next) => {
    if (message.incoming_message.type === 'message') {
        const qnaResults = await qnaMaker.getAnswers(message.context);
        if (qnaResults[0]) {
            // If we got an answer, send it directly
            //await message.context.sendActivity(qnaResults[0].answer);
            await bot.reply(message, qnaResults[0].answer); // also works
        } else {
            // If we have no other features, we could just say we didn't find any answers:
            //await message.context.sendActivity('No QnA Maker answers were found.');
            // Otherwise, just forward to the next BotHandler to see if there are any other matches
            next();
        }
    } else {
        next();
    }
});

// Catch-all for errors.
botAdapter.onTurnError = async (context, error) => {
    // This check writes out errors to console log
    // NOTE: In production environment, you should consider logging this to Azure
    //       application insights.
    console.error(`\n [onTurnError]: ${ error }`);
    // Send a message to the user
    await context.sendActivity(`Oops. Something went wrong!`);
    // Clear out state
    await conversationState.delete(context);
};

// load developer-created local custom feature modules.
// Load common features for slack and teams
slackController.loadModules(__dirname + '/features/common');
teamsController.loadModules(__dirname + '/features/common');
// Load platform specific features seperately
slackController.loadModules(__dirname + '/features/slack');
teamsController.loadModules(__dirname + '/features/teams');
smsController.loadModules(__dirname + '/features/twilio');

// For running server locally for development purposes
if (require.main === module) {
    const http = require("http");
    const port = process.env.PORT || 3000;
    const httpserver = http.createServer(webserver);
    httpserver.listen(port, function () {
        console.log("Slack Webhook endpoint online:  http://127.0.0.1:" + port + process.env.SLACK_WEBHOOK_URI);
        console.log("Teams Webhook endpoint online:  http://127.0.0.1:" + port + process.env.TEAMS_WEBHOOK_URI);
    });
};`

• Botkit version:^4.10.0
• Messaging Platform: Slack

[benbrown]

This is most likely due to the fact that your bot is "hearing" messages from itself. Make sure you are not processing messages FROM the bot.

[mattw114862]

After playing around with it I was able to get it fixed. It seems to be an issue with the slack API and how it handles retries in POST. I noticed on Ngrok that it was posting multiple 200 OK responses. I can't really say for sure if it is because it was the bot trying to respond to it's on query, however, I did notice that it only seemed to do this when I had Microsoft's QnA Maker integrated into the bot's responses.

Doing some googling on other people reporting the same issue I was finally able to stop the repeats by adding a few snippets:

// Create an express server with json and urlencoded body parsing
const webserver = express();
webserver.use((req, res, next) => {
    res.set("X-Slack-No-Retry", "1"); // disable retries
    res.status(200).json();
    req.rawBody = '';
    req.on('data', function (chunk) {
        req.rawBody += chunk;
    });
    next();
    return;
});
webserver.use(express.json());
webserver.use(express.urlencoded({ extended: true }));

Hope this helps any one else who experiences the same issue.

[benbrown]

Ah! This happens when the 200 response takes too long to be sent, which can happen when using external services like QNA Maker.

Thanks for the follow-up!