[Serverless Framework] S3 に保存した画像から Amazon Textract でテキストデータを取得する

serverless create --template aws-nodejs --name s3-textract --path s3-textract

npm init
npm install aws-sdk

import AWS from 'aws-sdk';

const textract = new AWS.Textract();
const s3 = new AWS.S3();

export const detectText = async (event) => {
  // S3 のイベントからオブジェクトの情報を取得
  const bucket = event.Records[0].s3.bucket.name;
  const objectKey = event.Records[0].s3.object.key;

  const params = {
    DocumentLocation: {
      S3Object: {
        Bucket: bucket,
        Name: objectKey
      }
    }
  };

  try {
    // Textractを呼び出し、テキストを抽出
    const response = await textract.startDocumentTextDetection(params).promise();
    
    // Textractのジョブが完了するのを待つ
    const jobId = response.JobId;
    await waitForJobCompletion(jobId);

    // テキストを取得
    const text = await getTextractText(jobId);
    console.log(text);
    
    // テキストを使用して必要な処理を行う
    const putparams = {
      Bucket: '20231113-text-bucket',
      Key: 'sample.txt',
      Body: text,
      ContentType: 'text/plain',
    };

    await s3.putObject(putparams).promise();

    return {
      statusCode: 200,
      body: 'Text file uploaded successfully.'
    };
  } catch (error) {
    return {
      statusCode: 500,
      body: JSON.stringify(error.message)
    };
  }
};

// Textractのジョブが完了するのを待つ
async function waitForJobCompletion(jobId) {
  const delay = (ms) => new Promise(resolve => setTimeout(resolve, ms));

  while (true) {
    const response = await textract.getDocumentTextDetection({ JobId: jobId }).promise();

    if (response.JobStatus === 'SUCCEEDED') {
      break;
    }

    if (response.JobStatus === 'IN_PROGRESS') {
      await delay(5000);
    }
  }
}

// Textractからテキストを取得
async function getTextractText(jobId) {
  const response = await textract.getDocumentTextDetection({ JobId: jobId }).promise();
  const text = response.Blocks
    .filter(block => block.BlockType === 'LINE')
    .map(block => block.Text)
    .join('\n');
  
  return text;
}

service: s3-textract

frameworkVersion: "3"

provider:
  name: aws
  runtime: nodejs18.x
  region: us-west-1
  iam:
    role:
      statements:
        - Effect: Allow
          Action:
            - s3:*
          Resource:
            - "arn:aws:s3:::20231113-image-bucket/*"
            - "arn:aws:s3:::20231113-text-bucket/*"
        - Effect: Allow
          Action:
            - textract:StartDocumentTextDetection
            - textract:GetDocumentTextDetection
          Resource:
            - "*"

functions:
  detectText:
    handler: handler.detectText
    timeout: 60
    events:
      - s3:
          bucket: 20231113-image-bucket
          event: s3:ObjectCreated:*
          existing: true
    layers:
      - {Ref: AwsLayerLambdaLayer}

resources:
  Resources:
    20231113ImageBucket:
      Type: AWS::S3::Bucket
      Properties:
        BucketName: 20231113-image-bucket
    20231113TextBucket:
      Type: AWS::S3::Bucket
      Properties:
        BucketName: 20231113-text-bucket

layers:
  awsLayer:
    path: aws-layer

.
├── handler.mjs
├── aws-layer
│   └── nodejs
│       └── node_modules
│
└── serverless.yml

sls deploy

aws s3api put-object --bucket 20231113-image-bucket --key sample.png --body ./sample.png

aws s3api list-objects-v2 --bucket 20231113-image-bucket

{
    "Contents": [
        {
            "Key": "sample.png",
            "LastModified": "2023-11-13T13:12:14+00:00",
            "ETag": "\"64e806e55d7fbb9ba011cf19599f97a4\"",
            "Size": 443275,
            "StorageClass": "STANDARD"
        }
    ]
}

aws s3api list-objects-v2 --bucket 20231113-text-bucket

{
    "Contents": [
        {
            "Key": "sample.txt",
            "LastModified": "2023-11-13T13:13:08+00:00",
            "ETag": "\"616ca57c925f619bbd6647609be65a19\"",
            "Size": 800,
            "StorageClass": "STANDARD"
        }
    ]
}

aws s3api get-object --bucket 20231113-text-bucket --key sample.txt sample.txt