AutoML Vision을 사용하여 고막 이미지 판별 Bot 만들기

npm install --save fetch-base64

'use strict';
require('dotenv').config();
const fetch = require('fetch-base64');
const express = require('express');
const line = require('@line/bot-sdk');
const PORT = process.env.PORT || 3000;

const config = {
  channelSecret: "自分のチャンネルシークレット",
  channelAccessToken: "自分のアクセストークン"
};

const automl = require('@google-cloud/automl');

// Create client for prediction service.
const mlclient = new automl.PredictionServiceClient();

const projectId = "自分のプロジェクト名";
const computeRegion = "us-central1";
const modelId = "自分のモデルID";
const scoreThreshold = "0.5";

// Get the full path of the model.
const modelFullId = mlclient.modelPath(projectId, computeRegion, modelId);

const aomLabels = {
  "aom": "急性中耳炎",
  "ome": "滲出性中耳炎",
  "normal": "正常鼓膜",
};

const app = express();
app.use(express.static('public')); //追加

app.get('/', (req, res) => res.send('Hello LINE BOT!(GET)')); //ブラウザ確認用(無くても問題ない)


app.post('/webhook', line.middleware(config), (req, res) => {
  console.log(req.body.events);

  //ここのif分はdeveloper consoleの"接続確認"用なので削除して問題ないです。
  if (req.body.events[0].replyToken === '00000000000000000000000000000000' && req.body.events[1].replyToken === 'ffffffffffffffffffffffffffffffff') {
    res.send('Hello LINE BOT!(POST)');
    console.log('疎通確認用');
    return;
  }

  Promise
    .all(req.body.events.map(event => handleEvent(event, req)))
    .then((result) => res.json(result));
});


const client = new line.Client(config);

async function handleEvent(event, req) {
  console.log(req);
  if (event.type !== 'message' || (event.message.type !== 'text' && event.message.type !== 'image')) {
    return Promise.resolve(null);
  }
  const params = {};
  params.score_threshold = scoreThreshold;

  // Read the file content for prediction.
  const data = await fetch.remote({
    url: `https://api-data.line.me/v2/bot/message/${event.message.id}/content`,
    headers: {
      'Authorization': `Bearer ${process.env.CHANNEL_ACCESS_TOKEN}`
    }
  });

  const payload = {};
  payload.image = { imageBytes: data[0] };

  // params is additional domain-specific parameters.
  // currently there is no additional parameters supported.
  const [response] = await mlclient.predict({
    name: modelFullId,
    payload: payload,
    params: params,
  });



  if (response.payload.length === 0) {
    return client.replyMessage(event.replyToken, {
      type: 'text',
      text: `ちょっとわかりませんね・・・`
    });
  }
  return client.replyMessage(event.replyToken, {
    type: 'text',
    text: `出ました！占いではあなたの耳は${aomLabels[response.payload[0].displayName]}と出ています。`
  });
}

app.listen(PORT);
console.log(`Server running at ${PORT}`);

GOOGLE_APPLICATION_CREDENTIALS=./ダウンロードしたJsonファイル名 
CHANNEL_ACCESS_TOKEN=自分のチャンネルアクセストークン