Read image format from S3 stream and write to disk with extension

Question

I'm wondering if it's possible to read an incoming image stream from an S3 bucket, extract the mime type and then write it the disk. Here's my current function:

async function retrieveAsset(
  bucket: string,
  asset: string,
  destination: string
) {
  const client = new S3Client({
    region: "auto",
    endpoint: `https://${ACCOUNT_ID}.r2.cloudflarestorage.com`,
    credentials: {
      accessKeyId: ACCESS_KEY_ID,
      secretAccessKey: SECRET_ACCESS_KEY,
    },
  });
  const response = await client.send(
    new GetObjectCommand({
      Bucket: bucket,
      Key: asset
    })
  );
  return new Promise(async (resolve, reject) => {
    const body = response.Body;
    if (body instanceof Readable) {
      const writeStream = createWriteStream(destination);
      body
        .pipe(sharp().metadata((err, meta) => {
           console.log(meta.format)
           // I'm not sure how to change the writable stream name here since it was defined above
        }))
        .pipe(writeStream)
        .on("error", reject)
        .on("close", resolve);
    }
  });
}

I'm having trouble figuring out a PassThrough stream so that I can extract it. Ultimately I want to rename the image file based on the mime type so that I might call:

retrieveAsset("test", "test.jpg", "newtest");

Which would result in an output file of newtest.jpg

Answer 1

Ultimately I want to rename the image file based on the mime type

Then, you don't actually need to do what you're asking for...

I'm wondering if it's possible to read an incoming image stream from an S3 bucket, extract the mime type and then write it the disk.

The type is in the Content-Type header, but its reliability depends on whatever did the upload in the first place. To read this header, you can make a simple HEAD request. Or, you could make a GET request and read the headers before the body data flows in, and disconnect if it's not what you wanted.

There's another issue though... the title of your question states...

Read mime type and dimensions from S3 stream before writing to disk

If you want the dimensions, you need to parse the image format. Most are going to have that data right up front so you could attempt to parse just the front of the file. In fact, you can do a range request and get just the first x bytes of the file for this purpose. You'd have to modify Sharp for this purpose though.

According to your question though, you don't actually need to know any of this information before you write the file. Simply download the file to a temporary file, have Sharp figure out the metadata you need, and rename the file accordingly.