Cutting out images: arch_cut_images.py

#!/usr/bin/python
# -*- coding: utf-8 -*-
# This file is meant to cut images out of whole page images in order to
# form a directory of figures so that they can be displayed. The input
# format is a list of URLs containing digilib coordinates.
#
# Imagemagick is then called as an external program to cut out the images.
# 
# Example: http://echo.mpiwg-berlin.mpg.de/ECHOdocuView?pn=3&ws=1&wx=0.0543&wy=0.251&ww=0.7991&wh=0.5829&url=/mpiwg/online/permanent/library/PUBSU9QD&viewMode=images&tocMode=thumbs&tocPN=1&searchPN=1
#
# Feature Request: Commented lines to have notes that some pictures should not be cut out

import string, re, os, subprocess, sys

# figfile = "Alvarus_1509_YHKVZ7B4.fig"
# ID = "PUBSU9QD"

archfigfile=sys.argv[1]

arch_ID = str(re.findall(r"_[0-9]{3}_",archfigfile))[3:-3]

arch_dir = "/Volumes/online_permanent/archimedes_repository/large/%s" % archfigfile[:-4]
arch_pi = "%s/%s-01-pageimg" % (arch_dir, arch_ID)
arch_fig = "%s/figures" % arch_dir

arch_URL = "http://echo.mpiwg-berlin.mpg.de/ECHOdocuView?url=/mpiwg/online/permanent/archimedes/%s&tocMode=figures" % archfigfile[:-4]

def leadZero(num,length=2):
    '''Print leading zeros, very useful, length is also more
    contrallable now, but we have a default value.'''
    tmp = str(num)
    return "0"*(length-len(tmp)) + tmp
# def leadZero ends here


def identify(dim,filename):
    """Short function for call to imagemagick's identify. Big problem
    here is how you can actually save the output of subprocess."""
    # This is it: I write the output of the shell into a variable
    output = subprocess.Popen(["identify","-format", dim, filename],stdout=subprocess.PIPE).communicate()[0]
    return output
# def identify ends here

def getpageimg(pidir):
    """Returns a list of the contents of the pageimg directory"""
    images = os.listdir(pidir)
    images.sort()
    return images
# def getpageimg ends here

# create a figures directory, if necessary
if os.path.exists(arch_fig):
    print "Figures directory %s already exists, please check." % arch_fig
    sys.exit()
else:
    print "Creating direcory %s" % arch_fig
    os.mkdir(arch_fig)

# if not os.path.exists(arch_fig):
#     os.mkdir(arch_fig)

# Get the pageimages and put them in a list
imglist = getpageimg(arch_pi)

# Three counters:
# Counter is for checking if there is more than one figure on one page.
imagecounter = 1

# Use linecount to count the lines, this one is used so that the
# old_pn is activated after the first iteration
linecount = 0

# The cutcount is for counting cut out figures
cutcount = 0

# Checker for comments
comment = 0

# read lines
#filein = open(figfile,"r")
filein = open(archfigfile,"r")

print "Getting images from", arch_pi

for line in filein:
    line = line.rstrip('\n') # removes line break.

    if line.startswith("#"):
        comment = 1
        print "Commented line"
        linecount = linecount + 1
        old_pn = pn
    else:
        # ws = str(re.findall(r"ws=[0-9.]*?&",line))[5:-3]
        pn = str(re.findall(r"pn=[0-9]*?&",line))[5:-3]
        wx = str(re.findall(r"wx=[0-9.]*?&",line))[5:-3]
        wy = str(re.findall(r"wy=[0-9.]*?&",line))[5:-3]
        ww = str(re.findall(r"ww=[0-9.]*?&",line))[5:-3]
        wh = str(re.findall(r"wh=[0-9.]*?&",line))[5:-3]
        # print leadZero(pn,3), wx, wy, ww, wh

        if linecount > 0:
            if pn == old_pn:
                imagecounter = imagecounter + 1
            elif pn != old_pn:
                imagecounter = 1
    
        # filename = "%s.jpg" % (leadZero(pn,3))
        # For local use, temporary. Normally, this one would point the tif
        # folder of the page images
        # filename = "/Users/kthoden/ECHO_pictures/apian_1550/%s.jpg" % (leadZero(pn,3))

        # Decrease pn by 1 to adjust to counting of imglist, not anymore?
        pnkorr = int(pn)-1
        # Construct filename based on foxridge directory and index in pageimg directory
        filename = "%s/%s" % (arch_pi,imglist[pnkorr])
    
        # Calculate size of image
        width_img = float(identify("%w",filename)) 
        height_img = float(identify("%h",filename)) 

        # absolute coordinates of upper left corner
        abs_x = float(wx) * width_img
        abs_y = float(wy) * height_img
    
        # print abs_x, abs_y

        # absolute size of region we want to extract
        abs_w = float(ww) * width_img
        abs_h = float(wh) * height_img

        # print abs_w, abs_h

        # Build a string for the figure's filename, based on pn-index 
        pagename = str(imglist[pnkorr])[:-4]

        # Still to be tested: can I use following notation also in the call itself?
        coord_for_im = "%sx%s+%s+%s" % (abs_w,abs_h,abs_x,abs_y)
        output = "%s/%s-%s.jpg" % (arch_fig,pagename,leadZero(imagecounter))

        subprocess.call(["convert", "-extract", coord_for_im,filename, output])
    
        print "Extracting image %s on page %s" % (leadZero(imagecounter),imglist[pnkorr])

        # Count line, store old page number.
        cutcount = cutcount + 1        
        linecount = linecount + 1
        old_pn = pn
    
    # Controls
    # print "linecount", linecount
    # print "old_pn", old_pn
    # raw_input()                 # Press Enter to continue

print "Wrote %d figures" % (cutcount)
if comment == 1:
    print "There were commented lines. This might mean that you have to edit the source xml file to remove erroneous image tags. "


answer = raw_input("Do you want to view the result in a browser? [Y/n] ").lower()
if answer == "y":
    subprocess.call(["open",arch_URL])
else:
    sys.exit()