内存转换器中的Python Wand.image PDF至JPG

Question

I am trying to write some code that will convert a PDF that resides on the web into a series of jpgs.

I got working code that:

1) takes pdf

2) saves it to disk

3) converts it to JPGs, which are saved to disk.

Is there a way to write the same code (attempt at code below, that throws an error), that would take the PDF from internet, but keep it in memory (to keep the program from writing to disk/reading from disk), then convert it to JPGs (which are to be uploaded to AWS s3)?

I was thinking this would work:

f = urlopen("https://s3.us-east-2.amazonaws.com/converted1jpgs/example.pdf") #file to process

But i get the following error:

"Exception TypeError: TypeError("object of type 'NoneType' has no len()",) in > ignored"

Full code, along with proper PDF file that i want converted. Note: the code works if i replace f= with the location of a PDF saved on disk:

from urllib2 import urlopen
from wand.image import Image

#location on disk
save_location = "/home/bob/Desktop/pdfs to convert/example1"

#file prefix
test_id = 'example'
print 1
f = urlopen("https://s3.us-east-2.amazonaws.com/converted1jpgs/example.pdf")
print 2
print type(f)

with Image(filename=f) as img:
    print('pages = ', len(img.sequence))
    with img.convert('jpg') as converted:
        converted.save(filename=save_location+"/"+test_id+".jpg")

Answer 1

The result of urlopen obviously isn't a filename, so you can't pass in filename=f and expect it to work.

I don't have Wand installed, but from the docs , there are clearly a bunch of alternative ways to construct it.

First, urlopen is a file-like object. Of course "file-like object" is a somewhat vague term, and not all file-like objects work for all APIs that expect file-like objects (eg, the API may expect to be able to call fileno and read from it at the POSIX level…), but this is at least worth trying (note file instead of filename ):

with Image(file=f) as img:

If that doesn't work, you can always read the data into memory:

buf = f.read()
with Image(blob=buf) as img:

Not as ideal (if you have giant files), but at least you don't have to store it on disk.