Non-image files are still read

[SansPapyrus683]

Hi!

I was using the compress utility of this library when I found that this function tries to read any uploaded file, not just image files.
This results in error messages being produced like the following:

A simple fix would be as follows in serverless-galleria-util/index.js, but I'm not sure if it would mesh with the other utilities:

      const rawKey = record.s3.object.key;
      const ext = rawKey.split(".").pop();
      const goodTypes = ["jpeg", "jpg", "png"];
      if (!goodTypes.includes(ext)) {
        console.log(`Skipping ${rawKey} as its file extension ${ext} is unsupported`)
        return;
      }

[evanchiu]

Hello @SansPapyrus683!

It looks like jimp has a clear listing of the file types that it supports, so I could probably check the mime type of the file before reading it to confirm that it’s supported.

The thing is, that this set of libraries are focused on building an image processing pipeline. So my expectation is that users are putting images in. If they put in something that’s not an image, the first lambda to pick it up will crash and the file won’t be propagated into the next bucket. Adding something like this gives a possibly clearer error message, but the processing result is the same.

What problem are you hoping for me to solve here?

[SansPapyrus683]

Hi, and thanks for the reply!

Adding something like this gives a possibly clearer error message

This is the main thing I'm looking for, but I completely understand if such code is deemed unnecessary. My issue was that AWS put some auxiliary stuff regarding cost metrics in the S3 bucket along with images, and the pipeline still tried to transform these .csv files and whatnot.