retrieve images from a raw discourse post

discourse_images.py

def get_dimensions(dimensions):
    # either 123x456 or 123x456,78%
    if ',' in dimensions:
        dim, perc = dimensions.split(',')
        perc = int(perc.replace('%', ''))
    else:
        dim = dimensions
        perc = 100
    
    x, y = dim.split('x')
    x = int(int(x)*perc / 100)
    y = int(int(y)*perc / 100)
    return "{x}x{y}".format(x=x, y=y)

def fix_uploads(doc, basehref, outputdir):
    uploads = re.findall('!\[(.*?)]\(upload://([A-Za-z0-9\.]*)\)', doc)
    for alt, upl in uploads:
        if '|' in alt:
            alttext, dimensions = alt.split('|')
            dimensions = get_dimensions(dimensions)
        else:
            alttext = alt
            dimensions = ''
            
        base, ext = upl.split('.')
        rebase = hex(base62.decode(base, base62.CHARSET_INVERTED))[2:].zfill(40)
        
        upload_location = '{basehref}/{c1}/{c2}/{rebase}.{ext}'.format(
            basehref=basehref, 
            c1=rebase[0], 
            c2=rebase[1], 
            rebase=rebase,
            ext=ext)
        filename = '{base}.{ext}'.format(
            base = hashlib.sha256(upload_location.encode('utf-8')).hexdigest(),
            ext = ext)
        fullfilename = '{outdir}/{filename}'.format(outdir = outputdir, filename = filename)
        urllib.request.urlretrieve(upload_location, fullfilename)
        
        if dimensions:
            cmd = "mogrify -resize {dimensions} {filename}".format(
                dimensions = dimensions,
                filename = fullfilename
            )
            os.system(cmd)
         
        search =  "![{alt}](upload://{upl})".format(alt=alt, upl=upl)
        replace = "\n![]({upl})\ \n".format(alt=alttext, upl=filename)
        doc = doc.replace(search,replace)
    return doc