CODEWITHSUNDEEP
rubyimagedownloadcorruptioncorrupt-data
Nakilon
Nakilon

Reputation: 35074

Validating downloaded image

This downloads image to a disk:

image = open permalink_url, "rb", &:read
...
File.binwrite "images/#{hash}", image

Sometimes it comes corrupted:

enter image description here

while there was no exception.

  1. How do I check if the image has been downloaded correctly (to otherwise retry the procedure)?
  2. How much is it ok, that there was no exception? How did it happen? Was the network exception silenced on some intermediate server?

UPD: Imagemagick says identify "reports if an image is incomplete or corrupt" but it does not:

$ identify temp.png
temp.png PNG 1080x1080 1080x1080+0+0 8-bit sRGB 2.126MB 0.000u 0:00.049

Here are two corrupted images:

  1. https://drive.google.com/file/d/0B3BLwu7Vb2U-MnNqdHV4MzFSX2s/view?usp=sharing
  2. https://drive.google.com/file/d/0B3BLwu7Vb2U-d3Fab2lmT1hvZlE/view?usp=sharing

UPD: I redownloaded the image and did some analysis -- the bad variation has 300000 extra bytes somewhere in the middle broken in a lot of pieces. Garbage is not just 0x00 but looks random.

Upvotes: 0

Views: 276

Answers (1)

kxmh42
kxmh42

Reputation: 3830

Use any of the image handling gems, e.g. chunky_png:

require 'chunky_png'
begin
  ChunkyPNG::Datastream.from_file('bad.png')
rescue ChunkyPNG::CRCMismatch
  puts "png corrupted!"
end

Edit: Datastream is more efficient than Image in this case.

Edit 2: If you want to be able to validate any format that ImageMagick can handle and don't mind calling external binaries, this should work:

unless system('identify', '-verbose', 'bad.jpg', out: IO::NULL, err: IO::NULL)
  puts "the file can't be opened or is corrupted"
end

Upvotes: 1

Related Questions