@aahnik I presume you're talking about an animated GIF? If so, see what you think of this
from PIL import Image, ImageSequence
def addOverlay(frame):
im = Image.open('Tests/images/hopper.png')
position = 'topLeft'
coords = {
'topLeft': (0, 0),
'bottomRight': (frame.width - im.width, frame.height - im.height)
}
frame.paste(im, coords[position])
return frame
im = Image.open('Tests/images/iss634.gif')
# Run 'addOverlay' on each frame in the image
frames = ImageSequence.all_frames(im, addOverlay)
# Save the frames as a new image
frames[0].save('out.gif', save_all=True, append_images=frames[1:])
the watermark, had a transparent background
but after putting it on the gif, its background became white
how can i fix this ?
background transparency solved
If I'm reading correctly that is, and you're still having a problem with the colors
Hi there, a user of an open-source program I maintain is getting an error I've never seen before and don't know what the cause is:
stitch_text and above is my own functions, everything below is PIL.
2021-03-03 22:34:52,004:ERROR:An unknown error occurred while creating a notification image: Traceback (most recent call last):
File "trade_worker.py", line 92, in send_trade
File "notification_builder.py", line 81, in build_image
File "notification_builder.py", line 132, in stitch_text
File "PIL\ImageDraw.py", line 373, in text
File "PIL\ImageDraw.py", line 324, in draw_text
File "PIL\ImageFont.py", line 667, in getmask2
OSError: invalid face handlestitch_text and above is my own functions, everything below is PIL.
Any help is greatly appreciated.
And why does the issue only occur for him? Out of 50+ active users, he's the only one to encounter the issue.
ok i'm an idiot, I think Image.Open() does what i need
thanks anyway
like get the maximum height/width for a given fiel size
Interesting question.
I presume you have a large image, and are looking to scale it down? If so, here's a rough solution.
I presume you have a large image, and are looking to scale it down? If so, here's a rough solution.
import io, os
from PIL import Image
def resize(im, factor):
return im.resize((int(im.width * factor), int(im.height * factor)), Image.ANTIALIAS)
def getsize(im, factor):
im_resized = resize(im, factor)
# Save to memory, rather than disk, and check size
with io.BytesIO() as fp:
im_resized.save(fp, "PNG")
return fp.getbuffer().nbytes
def scale_down_to_size(im, destination_size):
size = getsize(im, 1)
if size < destination_size:
print("Already small enough: "+str(size))
return im
# Binary search
factor = 0.5
gap = 0.5
while gap > 0.0005:
gap *= 0.5
size = getsize(im, factor)
if size > destination_size:
factor -= gap
else:
factor += gap
if size > destination_size:
factor -= gap
size = getsize(im, factor)
print("Final size: "+str(size))
return resize(im, factor)
with Image.open("Tests/images/hopper.png") as im:
im = scale_down_to_size(im, 500 * 1024) # 500 kb
im.save("out.png")
Hi! I've been waiting for feedback on #5201 for a while. I'm not complaining, I know how it is maintaining an open source project, and it's a pretty sizable change. I was wondering if there is anything I can do to make it easier to review, or to help out in any other way. Thanks.
Hi there, when I'm trying to save image as webp (larger than 9mb) I getting an error, anyone can help me?
Error: https://hastebin.com/ecokilaray.sql
Code: https://hastebin.com/oteteyexus.lua
Error: https://hastebin.com/ecokilaray.sql
Code: https://hastebin.com/oteteyexus.lua
You might running into https://github.com/python-pillow/Pillow/issues/5461#issuecomment-831114416 - error code 6 from webp, PARTITION0_OVERFLOW - "partition is bigger than 512k"
Without a copy of your image, I can't say for sure. The next version of Pillow due out in a month will help this a bit, by adding in an error code alongside 'encoding error'.
Is there any other way to fix error?
Specifically downloading this file will tigger defender, as well as the release tarball.
I filled a false positive report with micro$oft but in the meantime it's a real nuisance
anybody got any ideas about what's in that file thats setting it off??
The file is a reproduction of a previous CVE - we add it to the repository so that we test it and make sure it stays fixed. So I wouldn't say it's exactly a false positive, it's just only a security problem for Pillow < 8.2.0.
If you'd like, we can move the file out of the main repository and into a separate one. That's how we've dealt with similar reports in the past.
If you'd like, we can move the file out of the main repository and into a separate one. That's how we've dealt with similar reports in the past.
But if getting it out of the release tarball fixes the problem then :thumbsup:
recent updates to how GIFs were handled compounded the issue, as disposal methods 0/1 now leave behind a portion of the image8 array where the current frame is transparent, but if the current frame has a different color table (palette), the indices left in image8 will display the wrong colors
now, generally, GIF frames are meant to be loaded individually and the disposal method signals the app displaying the frames regarding how to handle them as they're loaded... this would mean not stacking frames as a decoder, and returning frames independently. This would be the cleanest fix, but might confuse some users even more than the current implementation already does.
I think the best compromise would be keeping colors from the previous frame that exist in the current frame and making colors that don't exist transparent, as these will be visible when the GIF is displayed
nothing would change except in the currently bugged case of disposal methods 0/1 and subsequent frames with local color tables which differ from previous frames' color tables
@RLaursen could you take a look at my response over at https://github.com/python-pillow/Pillow/commit/c54a7bb03127952065801232659a8bcf6383d43b#commitcomment-52165110 ?