7

I'm working on a photo website where I want the user to be able to upload a portrait or landscape oriented photo. The maximum width should be 1250px, but the maximum height could be 1667px if it's in portrait mode. When I upload photos in portrait orientation, they show up rotated 90 degrees to the left. Is there a way using Pillow to make sure the photo stays in the correct orientation?

This is my code:

class Result(models.Model):
    result01        = models.FileField(upload_to=get_upload_file_name, null=True, blank=True)
    result01thumb   = models.FileField(upload_to=get_upload_file_name, null=True, blank=True)

    def save(self):
        super(Result, self).save()
        if self.result01:
            size = 1667, 1250
            image = Image.open(self.result01)
            image.thumbnail(size, Image.ANTIALIAS)
            fh = storage.open(self.result01.name, "w")
            format = 'png'
            image.save(fh, format)
            fh.close()

It's important that users be able to upload photos from their phones while they're mobile, so the correct orientation is really important. Is there anything I can do here?

Jason Boyce
  • 213
  • 1
  • 2
  • 14
  • Possible duplicate of [Using PIL to auto rotate picture taken with cell phone and accelorometer](http://stackoverflow.com/questions/12133612/using-pil-to-auto-rotate-picture-taken-with-cell-phone-and-accelorometer) – Jerzyk Jun 05 '16 at 03:27

1 Answers1

6

You can try something like this to resize and auto-rotate (based on exif information) an image using Pillow.

def image_resize_and_autorotate(infile, outfile):
    with Image.open(infile) as image:
        file_format = image.format
        exif = image._getexif()

        image.thumbnail((1667, 1250), resample=Image.ANTIALIAS)

        # if image has exif data about orientation, let's rotate it
        orientation_key = 274 # cf ExifTags
        if exif and orientation_key in exif:
            orientation = exif[orientation_key]

            rotate_values = {
                3: Image.ROTATE_180,
                6: Image.ROTATE_270,
                8: Image.ROTATE_90
            }

            if orientation in rotate_values:
                image = image.transpose(rotate_values[orientation])

        image.save(outfile, file_format)
Augusto Destrero
  • 4,095
  • 1
  • 23
  • 25
  • Great, that worked for me! One question, why can't you access `image.format` after performing `image.transpose`? – J. Diaz Sep 11 '22 at 15:09