I'm using Sorl Thumbnail 12.4a1, Pillow 4.0.0 in a django project
It works fine for most images but for some of them I run into unexpected behaviour.
The original image is portrait jpg 765x1110
After running this code: get_thumbnail('/media/ryzsUIKZVUc.jpg', '464', upscale=False)
I receive a 464x320 image with black background on it's left and right and cropped on top and bottom
this is the original: https://www.dropbox.com/s/ia7906ec19xjhro/ryzsUIKZVUc.jpg?dl=0
and this is the result: https://www.dropbox.com/s/adgut5zkw4xln6e/62b2439654a00312defafddb862eda9b.jpg?dl=0
p.s. tried to upload the original here as an image but it was automatically converted to portrait.. could it mean something?
The frame appears because of the exif tag. The size of the image is not correctly determined.
You can see this using exiftool:
$ exiftool ryzsUIKZVUc.jpg
ExifTool Version Number : 10.10
File Name : ryzsUIKZVUc.jpg
Directory : .
File Size : 210 kB
File Modification Date/Time : 2017:03:30 12:33:51+03:00
File Access Date/Time : 2017:03:30 12:33:55+03:00
File Inode Change Date/Time : 2017:03:30 12:33:51+03:00
File Permissions : rw-rw-r--
File Type : JPEG
File Type Extension : jpg
MIME Type : image/jpeg
JFIF Version : 1.01
Comment : CREATOR: gd-jpeg v1.0 (using IJG JPEG v62), quality = 99.
Exif Byte Order : Big-endian (Motorola, MM)
Processing Software : Windows Photo Editor 10.0.10011.16384
Orientation : Rotate 270 CW
Software : Windows Photo Editor 10.0.10011.16384
Modify Date : 2017:03:07 09:28:15
Date/Time Original : 2017:03:07 09:27:46
Create Date : 2017:03:07 09:27:46
Sub Sec Time Original : 00
Sub Sec Time Digitized : 00
Color Space : sRGB
Padding : (Binary data 2060 bytes, use -b option to extract)
Compression : JPEG (old-style)
X Resolution : 96
Y Resolution : 96
Resolution Unit : inches
Thumbnail Offset : 4608
Thumbnail Length : 8169
About : uuid:faf5bdd5-ba3d-11da-ad31-d33d75182f1b
Creator Tool : Windows Photo Editor 10.0.10011.16384
Image Width : 1110
Image Height : 765
Encoding Process : Baseline DCT, Huffman coding
Bits Per Sample : 8
Color Components : 3
Y Cb Cr Sub Sampling : YCbCr4:2:0 (2 2)
Image Size : 1110x765
Megapixels : 0.849
Create Date : 2017:03:07 09:27:46.00
Date/Time Original : 2017:03:07 09:27:46.00
Thumbnail Image : (Binary data 8169 bytes, use -b option to extract)
Or use PIL
>>> from PIL import Image
>>> from PIL.ExifTags import TAGS
>>>
>>> im = Image.open('ryzsUIKZVUc.jpg')
>>> exif = im._getexif()
>>> data = {}
>>>
>>> for tag, value in exif.items():
... decoded = TAGS.get(tag, tag)
... data[decoded] = value
...
>>>
>>> data['Orientation']
8
To fix this, you can before generating a preview, (rotate image) save the picture without this tag.
For example,
def orient(image_path):
""" Check are this img oriented """
img = Image.open(image_path)
# check is has exif data
if not hasattr(img, '_getexif') or img._getexif() is None:
return
# if has information - try rotate img
orientation = img._getexif().get(0x112)
rotate_values = {3: 180, 6: 270, 8: 90}
if orientation in rotate_values:
img = img.rotate(rotate_values[orientation], expand=True)
img.save(image_path, quality=100)
Or easier, change the settings THUMBNAIL_ORIENTATION = False. In this case, the image simply stops rotating. And no more black lines are created.