We Keep Coding

img2pdf AlphaChannelError: what is the best way to remove alphachannel

I have set of images from which I create pdf by the following code
with io.BytesIO() as tmp_io:
tmp_io.write(img2pdf.convert(img_file_paths))
result_bytes = tmp_io.getvalue()
One of files contains alpha channel and I got
raise AlphaChannelError("Refusing to work on images with alpha channel")
What is the simplest way to remove alpha channel and save to pdf rgb channels?

Here is bit ugly solution from myself
def remove_alpha_from_image(image_path):
im = Image.open(image_path)
im.load()
try:
background = Image.new("RGB", im.size, (255, 255, 255))
background.paste(im, mask=im.split()[3]) # 3 is the alpha channel
im = background
except IndexError: # img is not RGBA
pass
name_hash_md5 = md5(bytes(image_path, encoding="utf-8")) # noqa: S303
name = name_hash_md5.hexdigest()
if not os.path.exists(TMP_DIR):
os.makedirs(TMP_DIR)
path = f"{TMP_DIR}{name}.pdf"
im.save(path, "PNG", resolution=100.0)
return path
with io.BytesIO() as tmp_io:
try:
tmp_io.write(img2pdf.convert(file_paths))
except img2pdf.AlphaChannelError:
tmp_io.write(img2pdf.convert([remove_alpha_from_image(path) for path in file_paths]))
result_bytes = tmp_io.getvalue()

Here's a utility I put together - only tested in a single app so not sure how general it is, but should be turnkey. Tested in python 3.9
def image2pdf(image: bytes or str, allow_lossy=True, **rgba_to_kwds) -> bytes:
"""
Converts an image to PDF, optionally allowing for lossy conversion.
:param image: if non RGBA image, this can be any valid input to img2pdf. If RGBA, then must be str (ie. path to image)
or bytes representation of image.
:param allow_lossy: if img2pdf.convert fails with AlphaChannelError, tries to downsample
:param rgba_to_kwds: kwds to _rgba_to
:return: bytes representation of PDF image. To save to disk
pdfBytes=image2pdf(someImage)
with open('converted.pdf', 'w') as f:
f.write(pdfBytes)
"""
try:
pdf_bytes = img2pdf.convert(image)
except img2pdf.AlphaChannelError as alphaError:
if allow_lossy:
rgbBytes = _rgba_to(image)
pdf_bytes = img2pdf.convert(rgbBytes, **rgba_to_kwds)
else:
raise alphaError
return pdf_bytes
def _rgba_to(image: bytes or str, to='RGB', intermediate='PNG') -> bytes:
logging.warning(f"Image has alpha channel... downsampling (newtype={to}, intermediate={intermediate}) and converting")
# Image is a filepath
if isinstance(image, str):
img = Image.open(image)
converted: Image = img.convert(to)
# Image is a bytestream
elif isinstance(image, bytes):
buffered = io.BytesIO(image)
img = Image.open(buffered)
converted: Image = img.convert(to)
else:
raise Exception(f"rgba downsampling only supported for images of type str (ie. filepath) or bytes - got {type(image)}")
buf = io.BytesIO()
converted.save(buf, format=intermediate)
byte_im = buf.getvalue()
return byte_im
def test_convert_png_image_with_alphachannel_to_pdf():
img_path = "some-rgba-image.png"
pdf_bytes = image2pdf(img_path)
# Uncomment if want to view the pdf
with open('converted.pdf', "wb") as f:
f.write(pdf_bytes)

Exception Type: OSError Exception Value: cannot write mode RGBA as JPEG

I have upload more than 10 images with the help of the following code in my django application(using image pillow library).
def save(self):
im = Image.open(self.image)
output = BytesIO()
im = im.resize((500, 500))
im.save(output, format='JPEG', optimize=True, quality=95)
output.seek(0)
self.image = InMemoryUploadedFile(output, 'ImageField', "%s.jpg" % self.image.name.split('.')[0], 'image/jpeg',
sys.getsizeof(output), None)
super(Report_item, self).save()
But now my Django application is giving me following error.
Exception Type: OSError
Exception Value:
cannot write mode RGBA as JPEG
I got the solution from one of the answers at StackOverflow to change the image type to from any other to png.
Now my code looks like this but the process seems a bit slow as compared to it was earlier.
def save(self):
im = Image.open(self.image)
output = BytesIO()
im = im.resize((500, 500))
im.save(output, format='PNG', optimize=True, quality=95)
output.seek(0)
self.image = InMemoryUploadedFile(output, 'ImageField', "%s.png" % self.image.name.split('.')[0], 'image/jpeg',
sys.getsizeof(output), None)
super(Report_item, self).save()
Now the image is uploading without any issue.
Please explain to me why I got that error after uploading more than 10 pics.
Is this the right way to do so. What if I want to save in jpeg format?
My application is in production. So I want to push only the right code for the right job. Please help me to figure out.

Please explain to me why I got that error after uploading more than 10 pics
A: there is some image which is RGBA or P mode, which not support directly convert to JPG
Explanation:
summary 1 and 2:
backgroud
JPG not support alpha = transparency
RGBA, P has alpha = transparency
RGBA= Red Green Blue Alpha
result
cannot write mode RGBA as JPEG
cannot write mode P as JPEG
solution
before save to JPG, discard alpha = transparency
such as: convert Image to RGB
then save to JPG
-> your here:
so you use convert to PNG then save
im.save(output, format='PNG', optimize=True, quality=95)
Is this the right way to do so? What if I want to save in jpeg format?
A: No.
Recommend to use:
rgba_or_p_im = original_im
if rgba_or_p_im.mode in ["RGBA", "P"]:
rgb_im = rgba_or_p_im.convert("RGB")
rgb_im.save("xxx.jpg")

How to change image format without writing it to disk using Python Pillow

I got Pillow image that i got from the Internet:
response= urllib2.urlopen(<url to gif image>)
img = Image.open(cStringIO.StringIO(response.read()))
I want to use it with tesserocr but it wont work with GIF images.
If I save the image as PNG img.save("tmp.png") and load it img = Image.open("tmp.png") everything works.
Is there a way to do this conversion without writing to disk?

import io
from PIL import Image
def convertImageFormat(imgObj, outputFormat=None):
"""Convert image format
Args:
imgObj (Image): the Pillow Image instance
outputFormat (str): Image format, eg: "JPEG"/"PNG"/"BMP"/"TIFF"/...
more refer: https://pillow.readthedocs.io/en/stable/handbook/image-file-formats.html
Returns:
bytes, binary data of Image
Raises:
"""
newImgObj = imgObj
if outputFormat and (imgObj.format != outputFormat):
imageBytesIO = io.BytesIO()
imgObj.save(imageBytesIO, outputFormat)
newImgObj = Image.open(imageBytesIO)
return newImgObj
call example:
pngImgFile = "xxx.png"
pngImgObj = Image.open(pngImgFile)
convertToFormat = "JPEG"
convertedJpgImgBytes = convertImageFormat(pngImgObj, convertToFormat)
advanced version convertImageFormat can refer my lib crifanPillow.py
import io
from PIL import Image
def convertImageFormat(imgObj, outputFormat=None, isOptimize=False, isKeepPrevValues=True):
"""Convert image format
Args:
imgObj (Image): the Pillow Image instance
outputFormat (str): Image format, eg: "JPEG"/"PNG"/"BMP"/"TIFF"/...
more refer: https://pillow.readthedocs.io/en/stable/handbook/image-file-formats.html
isOptimize (bool): do optimize when using save to convert format
isKeepPrevValues (bool): keep previous property values, such as: filename
Returns:
bytes, binary data of Image
Raises:
"""
newImgObj = imgObj
if outputFormat and (imgObj.format != outputFormat):
imageBytesIO = io.BytesIO()
if isOptimize:
imgObj.save(imageBytesIO, outputFormat, optimize=True)
else:
imgObj.save(imageBytesIO, outputFormat)
newImgObj = Image.open(imageBytesIO)
if isKeepPrevValues:
if imgObj.filename:
newImgObj.filename = imgObj.filename
return newImgObj

The solution was very simple:
response= urllib2.urlopen(<url to gif image>)
img = Image.open(cStringIO.StringIO(response.read()))
img = img.convert("RGB")
Note that you need to remove the alpha channel info to make image compatible with tesserocr

Use custom scrapy imagePipeline to download images and overwrite existing images

I am practising using scrapy to crop image with a custom imagePipeline.
I am using this code:
class MyImagesPipeline(ImagesPipeline):
def get_media_requests(self, item, info):
for image_url in item['image_urls']:
yield Request(image_url)
def convert_image(self, image, size=None):
if image.format == 'PNG' and image.mode == 'RGBA':
background = Image.new('RGBA', image.size, (255, 255, 255))
background.paste(image, image)
image = background.convert('RGB')
elif image.mode != 'RGB':
image = image.convert('RGB')
if size:
image = image.copy()
image.thumbnail(size, Image.ANTIALIAS)
else:
# cut water image TODO use defined image replace Not cut
x,y = image.size
if(y>120):
image = image.crop((0,0,x,y-25))
buf = StringIO()
try:
image.save(buf, 'JPEG')
except Exception, ex:
raise ImageException("Cannot process image. Error: %s" % ex)
return image, buf
It works well but have a problem.
If there are original images in the folder,
then run the spider,
the images it download won't replace the original one.
How can I get it to over-write the original images ?

There is an expiration setting, it is by default 90 days.

PIL - Convert GIF Frames to JPG

I tried to convert an gif to single images with Python Image Library,
but it results in weird frames
The Input gif is:
Source Image http://longcat.de/gif_example.gif
In my first try, i tried to convert the image with Image.new to an
RGB image, with 255,255,255 as white background - like in any other
example i've found on the internet:
def processImage( infile ):
try:
im = Image.open( infile )
except IOError:
print "Cant load", infile
sys.exit(1)
i = 0
try:
while 1:
background = Image.new("RGB", im.size, (255, 255, 255))
background.paste(im)
background.save('foo'+str(i)+'.jpg', 'JPEG', quality=80)
i += 1
im.seek( im.tell() + 1 )
except EOFError:
pass # end of sequence
but it results in weird output files:
Example #1 http://longcat.de/gif_example1.jpg
My second try was, to convert the gif in an RGBA first, and then use
its transparency mask, to make the transparent pieces white:
def processImage( infile ):
try:
im = Image.open( infile )
except IOError:
print "Cant load", infile
sys.exit(1)
i = 0
try:
while 1:
im2 = im.convert('RGBA')
im2.load()
background = Image.new("RGB", im2.size, (255, 255, 255))
background.paste(im2, mask = im2.split()[3] )
background.save('foo'+str(i)+'.jpg', 'JPEG', quality=80)
i += 1
im.seek( im.tell() + 1 )
except EOFError:
pass # end of sequence
which results in an output like this:
Example #2 http://longcat.de/gif_example2.jpg
The advantage over the first try was, that the first frame looks pretty good
But as you can see, the rest is broken
What should i try next?
Edit:
I think i came a lot closer to the solution
Example #3 http://longcat.de/gif_example3.png
I had to use the palette of the first image for the other images,
and merge it with the previous frame (for gif animations which use
diff-images)
def processImage( infile ):
try:
im = Image.open( infile )
except IOError:
print "Cant load", infile
sys.exit(1)
i = 0
size = im.size
lastframe = im.convert('RGBA')
mypalette = im.getpalette()
try:
while 1:
im2 = im.copy()
im2.putpalette( mypalette )
background = Image.new("RGB", size, (255,255,255))
background.paste( lastframe )
background.paste( im2 )
background.save('foo'+str(i)+'.png', 'PNG', quality=80)
lastframe = background
i += 1
im.seek( im.tell() + 1 )
except EOFError:
pass # end of sequence
But i actually dont know, why my transparency is black, instead of white
Even if i modify the palette (change the transparency channel to white)
or use the transparency mask, the background is still black

First of all, JPEG doesn't support transparency! But that's not the only problem.. As you move to the next frame of the GIF the palette information is lost (problem witn PIL?) - so PIL is unable to correctly convert to the RGBA framework (Hence the first frame is okish, but all the others are screwy). So the work-around is to add the palette back in for every frame, (which is what you were doing in your last code example, but your trouble was that you were saving as RGB not RGBA so you had no alpha/ transparency channel. Also you were doing a few unnecessary things..). Anyhow, here are the .png's with transparency and the corrected code, hope its of some use :)
import Image
import sys
def processImage(infile):
try:
im = Image.open(infile)
except IOError:
print "Cant load", infile
sys.exit(1)
i = 0
mypalette = im.getpalette()
try:
while 1:
im.putpalette(mypalette)
new_im = Image.new("RGBA", im.size)
new_im.paste(im)
new_im.save('foo'+str(i)+'.png')
i += 1
im.seek(im.tell() + 1)
except EOFError:
pass # end of sequence
processImage('gif_example.gif')

When viewing an image on an image viewer, even when transparency is set to zero, it tends to display the image as black. One way to be sure that your image is truly transparent is to merge it over another. The 'emoticon' should be seen whilst not obstructing the other image.Try:
background = Image.open('someimage.jpg') #an existing image
foreground = Image.open('foo.jpg') #one of the above images
background.paste(foreground, (0,0), foreground)
background.save('trial.jpg') #the composite image
Theoretically, if you open 'trial.jpg' in the image viewer and the content of the initial image is preserved and on top of it lies the foo image then you'll know for sure if it's just the image viewer and your images are fine...

source here
Image.open('image.gif').convert('RGB').save('image.jpg')

This works for me. The following example shows converting image.gif to 8 jpg format images with white background.
from PIL import Image
from PIL import GifImagePlugin
def gif2jpg(file_name: str, num_key_frames: int, trans_color: tuple):
"""
convert gif to `num_key_frames` images with jpg format
:param file_name: gif file name
:param num_key_frames: result images number
:param trans_color: set converted transparent color in jpg image
:return:
"""
with Image.open(file_name) as im:
for i in range(num_key_frames):
im.seek(im.n_frames // num_key_frames * i)
image = im.convert("RGBA")
datas = image.getdata()
newData = []
for item in datas:
if item[3] == 0: # if transparent
newData.append(trans_color) # set transparent color in jpg
else:
newData.append(tuple(item[:3]))
image = Image.new("RGB", im.size)
image.getdata()
image.putdata(newData)
image.save('{}.jpg'.format(i))
gif2jpg("image.gif", 8, (255, 255, 255)) # convert image.gif to 8 jpg images with white background