Update locally referenced images in your blog posts with cloud references in S3

replaceImages.js

// This script will scan all files in a given directory for locally referenced images. It will take the images, rename them with a standard convention, then upload them to s3 with a 'public-read' ACL. The script will then update the reference to the local image with the link to the version in the cloud.
// If there are any issues during operation, the script will output either a 'skipped-posts.json' file that lists the posts that were not processed or a 'failed-image-uploads.json' file that lists images that failed to upload into S3.

// Arguments for operation
//
// [0] blogPostDirectory - relative path from the root where all blog posts live
// [1] imageDirectory - relative path from the root where all images live
// [2] bucketName - name of the S3 bucket to upload the local images to
// [3] awsProfileName - name of the aws profile on your machine that has access to the S3 bucket

//
// Example invocation
//
// node replaceImages.js 'content/blog' static myimagebucket sandbox
//

const blogPostDirectory = process.argv[2]; // content/blog
const imageDirectory = process.argv[3]; // static
const bucketName = process.argv[4];
const awsProfileName = process.argv[5] ?? 'default';


const fs = require('fs');
const frontmatter = require('@github-docs/frontmatter');
const AWS = require('aws-sdk');
const { S3Client, PutObjectCommand } = require('@aws-sdk/client-s3');

async function run() {
  const credentials = new AWS.SharedIniFileCredentials({ profile: awsProfileName });
  const s3 = new S3Client({ credentials: credentials });

  const imagesToRemove = [];
  const skippedPosts = [];
  const failedImageUploads = [];

  const posts = fs.readdirSync(blogPostDirectory);
  for (const post of posts) {
    console.log(`Processing blog post: ${post}`);

    let data = fs.readFileSync(`${blogPostDirectory}/${post}`, 'utf8');
    const metadata = frontmatter(data);
    if (!metadata?.data?.title) {
      console.log(`No title found for post: ${post}. Skipping replacement.`);
      skippedPosts.push(post);
      continue;
    }

    // Strip title of special characters and make lowercase
    const prefix = metadata.data.title.toLowerCase().replace(/ /g, '_').replace(/[^\w\s]/gi, '').trim();

    if (!metadata?.data?.image) {
      console.log(`${post} has no featured image.`);
    }
    else if (metadata?.data?.image?.startsWith('images/')) {
      const pieces = metadata.data.image.split('/');
      const imageFileName = pieces[pieces.length - 1];
      const extension = imageFileName.split('.')[1];

      const originalImage = `./${imageDirectory}/${metadata.data.image}`;
      const response = await s3.send(new PutObjectCommand({
        Bucket: bucketName,
        Key: `${prefix}_feature.${extension}`,
        ACL: 'public-read',
        Body: fs.readFileSync(originalImage)
      }));

      if (response.ETag) { // Indicates a successful PUT
        data = data.replace(metadata.data.image, `https://${bucketName}.s3.amazonaws.com/${prefix}_feature.${extension}`);
        if (!imagesToRemove.includes(originalImage)) {
          imagesToRemove.push(originalImage);
        }
      } else {
        console.error('Error replacing feature image');
        failedImageUploads.push({
          post: post,
          image: metadata.data.image
        });
      }
    }

    // look for markdown links that start with (../images/post). Can have N number of leading dots
    const images = data.matchAll(/(\(\.\..*\/images\/post\/.*\))/g);
    let count = 0;
    let allReplacementsSuccessful = true;
    for (const image of images) {
      const pieces = image[1].split('/');
      const originalFileName = pieces[pieces.length - 1];

      const extension = originalFileName.split('.')[1].replace(')', '');
      const fileName = `${prefix}_${count++}.${extension}`;
      const originalImage = `./${imageDirectory}/${pieces.filter(p => !p.includes('..')).join('/')}`.replace(')', '');

      const response = await s3.send(new PutObjectCommand({
        Bucket: bucketName,
        Key: fileName,
        ACL: 'public-read',
        Body: fs.readFileSync(originalImage)
      }));

      if (response.ETag) {
        data = data.replace(image[1].replace('(', '').replace(')', ''), `https://${bucketName}.s3.amazonaws.com/${fileName}`);
        if (!imagesToRemove.includes(originalImage)) {
          imagesToRemove.push(originalImage);
        }
      } else {
        console.error(`Unable to save image to S3 ${fileName}`);
        failedImageUploads.push({
          post: post,
          image: image[1]
        });
      }
    }

    fs.writeFileSync(`${blogPostDirectory}/${post}`, data);
    console.log(`${post} updated complete.`);
  }

  for (const imageToRemove of imagesToRemove) {
    console.log(`removing ${imageToRemove}`);
    fs.rmSync(imageToRemove);
  }

  if (skippedPosts.length) {
    fs.writeFileSync('skipped-posts.json', JSON.stringify(skippedPosts, null, 2));
  }

  if (failedImageUploads.length) {
    fs.writeFileSync('failed-image-uploads.json', JSON.stringify(failedImageUploads, null, 2));
  }
}

run();