#!/usr/bin/env python
"""web2png -- convert a web hierarchy from using GIFs to using PNGs.

This script is a front end for gif2png that assists you in converting an
entire website.  Requires gif2png 1.0.5 or later.  Requires python 1.5.2
or later. Requires Unix.

by Eric S. Raymond <esr@thyrsus.com>

Slightly modified by Aaron Isotton <aaron@isotton.com> for Debian.
"""
__version__="$Revision: 1.25 $" 

# Future library code -- I'm going to submit this for Python 1.6
import sys, os, sys
from stat import *

if sys.version[:5] < "1.5.2":
    print "This program requires Python 1.5.2 or later."
    sys.exit(1)

FTW_FILE    = 0	# Item is a normal file
FTW_DIR     = 1	# Item is a directory
FTW_CHRDEV  = 2	# Item is a character special device file
FTW_BLKDEV  = 3	# Item is a block special device file
FTW_FIFO    = 4	# Item is a FIFO
FTW_SYMLINK = 5	# Item is a symbolic link
FTW_SOCKET  = 6	# Item is a socket
FTW_NOSTAT = -1	# stat failed on item
FTW_DNR	   = -2	# item was an unreadable directory

class FTWException(Exception):
	pass

def ftw_inner(func, path=".", extra=[]):
    try:
        st = os.stat(path)
    except:
        return FTW_NOSTAT
    mode = st[ST_MODE]

    if S_ISDIR(mode):
        try:
            subdirs = os.listdir(path)
        except:
            return FTW_DNR

    if S_ISREG(mode):  ftype = FTW_FILE
    if S_ISDIR(mode):  ftype = FTW_DIR
    if S_ISBLK(mode):  ftype = FTW_BLKDEV
    if S_ISCHR(mode):  ftype = FTW_CHRDEV
    if S_ISFIFO(mode): ftype = FTW_FIFO
    if S_ISLNK(mode):  ftype = FTW_SYMLINK
    if S_ISSOCK(mode): ftype = FTW_SOCKET

    terminate = apply(func, tuple([path, st, ftype] + list(extra)))
    if terminate < 0:
        raise FTWException, terminate

    if S_ISDIR(mode) and not terminate:
        for f in subdirs:
            ftw_inner(func, os.path.join(path, f), extra)

def ftw(func, path=".", *extra):
    """Python tree-walker -- like ftw(3), but more capable."""
    try:
        ftw_inner(func, path, extra)
    except FTWException, val:
        return val
    return 0

def find(dir=".", regexp=None):
   """Return a tuple list of entries beneath dir matching a given pattern."""
   def findhook(path, st, ftype, r, fl):
       if not r or r.search(path):
           fl.append((path, ftype))
       return 0
   flist=[]
   ftw(findhook, dir, regexp, flist)
   return flist

def findfiles(dir=".", regexp=None):
    """Return a list of files beneath dir matching a given pattern."""
    return map(lambda x: os.path.normpath(x[0]),
               filter(lambda t: t[1]==FTW_FILE, find(dir, regexp)))

def cmp(f1, f2, blocksize=1):
    # print "Comparing %s with %s" % (f1, f2)
    # st1 = os.stat(f1); print st1
    # st2 = os.stat(f2); print st2
    f1 = open(f1, "r")
    f2 = open(f2, "r")
    blocknum = 0
    try:
        while 1:
            blocknum = blocknum + 1
            s1 = f1.read(blocksize);
            s2 = f2.read(blocksize);
            if not s1 and not s2:
                break
            elif s1 != s2:
                return blocknum
    finally:
        f1.close()
        f2.close()
    return None

# End future library code

import re, getopt, commands, string, shutil, urllib

nochange = all = None
gifre = re.compile(r"\.gif$", re.IGNORECASE)
pngre = re.compile(r"\.png$", re.IGNORECASE)
pagere = re.compile(r"\.s?html$|\.s?htm$|\.php$|\.inc$|\.css$|\.js$", re.IGNORECASE)
bakre = re.compile(r"\.bak$", re.IGNORECASE)

# Possible left delimiters of these matches are =, ", (, \s
# Possible right delimiters of these matches are ", >, ), \s
# .?"? parts deal with the possibility that the optional " might be backslashed
delim = r'(?:\\?")?'	# Optional double quote, possibly preceded by backslash
gif = delim + '([^">]*\.gif)' + delim
imgre = re.compile(r'SRC=' + gif, re.IGNORECASE)
hrefre = re.compile(r'<A HREF=' + gif + '>', re.IGNORECASE)
backre = re.compile(r'BACKGROUND=' + gif, re.IGNORECASE)
basere = re.compile(r'<BASE HREF=' + gif + '>', re.IGNORECASE)
cssurlre = re.compile(r'url\(' + gif + '\)', re.IGNORECASE)

def version_controlled(page):
    # Is given page under version control?
    return os.path.exists(page + ",v") \
           or os.path.exists(os.path.join("RCS", page) + ",v")

def web2png(directory):
    # Convert a web hierarchy rooted on the given directory
    gifs = findfiles(directory, gifre)
    htmls = findfiles(directory, pagere)

    # There's a standard max on the number of arguments we can feed gif2png.
    # if we see more than these, 
    if len(gifs) > 5120:
        "web2png: Too many GIFs.  Try converting some subtrees first."
        sys.exit(1)

    print "This web subtree has", len(gifs), "GIFs and", len(htmls), "pages."

    # Display information on files we won't convert
    rejects = \
	commands.getoutput("gif2png -w "+transparency+string.join(gifs," ")+" >/dev/null")
    rejects = re.sub("gif2png: ", "	", rejects)
    if rejects:
        if all:
            print "You are forcing conversion of these GIFs:"
        else:
            print "The following GIFs will not be converted:"
        print rejects

    # Figure out which files are eligible for conversion
    if all:
        giflist = gifs
    else:
        giflist = \
                string.split(commands.getoutput("gif2png -w "+transparency+string.join(gifs," ")+" 2>/dev/null"))

    if verbose: print "Giflist: " + repr(giflist)

    convert_gifs = []
    for gif in giflist:
        png = re.sub(r"\.gif$", r".png", gif)
        if os.path.exists(png):
            print "\t%s already has a PNG equivalent" % (gif,)
        else:
            convert_gifs.append(gif)

    # Display information on files we will convert
    if convert_gifs:
        print "The following GIFs will be converted:\n\t" + \
		string.join(convert_gifs, "\n\t")
    if not convert_gifs:
        print "All eligible GIFs seem to have been converted already."

    # Create a dictionary mapping pages to sets of references to be mapped
    print "Checking for HTML, PHP JavaScript and include pages that need conversion..."
    pagecount = 0
    page_conversions = {}
    for file in htmls:
        fp = open(file, "r")
        contents = fp.read()
        fp.close()

        basedir = basere.search(contents)
        if basedir:
            basedir = base.group(1)
        else:
            basedir = os.path.dirname(file)

        matches = imgre.findall(contents) + hrefre.findall(contents) + backre.findall(contents) + cssurlre.findall(contents)
        if verbose:
	    print "Found %d .gif image(s) references in %s."%(len(matches),file)
        convert_refs = []
        gif_basenames = map(os.path.basename, giflist)
        for ref in matches:
            withbase = os.path.join(basedir, ref)
            # Simple case -- it's a filename, we presume it's local
            if withbase[:5] != "http:" and withbase[1:] != os.sep:
		
                target = os.path.normpath(withbase)
                if not target in giflist:
                    if verbose: print "Won't replace %s. Not on my list."%target
                    continue
            # Tricky case -- compare by basename and content
            else:
                basename = os.path.basename(withbase[7:])
                if not basename in gif_basenames:
                    print "No local match by basename for %s" % (ref,)
                    continue
                target = giflist[gif_basenames.index(basename)]
                (data, headers) = urllib.urlretrieve(withbase)
                not_equal = cmp(data, target)
                os.remove(data)
                if not_equal:
                    print "Data of %s and %s don't match at offset %d" % (ref, target, not_equal)
                    continue
            ref = ref[:-4]
            convert_refs.append((ref, target))
        if convert_refs:
            print "\tIn %s, I see: %s" % (file, string.join(map(lambda x: x[0]+".gif", convert_refs)," "))
            page_conversions[file] = convert_refs
	    pagecount = pagecount + 1
    print "%d HTML or PHP page(s) need conversion." % (pagecount,)

    # Unless user is willing to make changes, we're done now
    if nochange:
        return

    # Convert gifs verbosely
    if convert_gifs:
        print "GIF conversions begin:"
        os.system("gif2png -v -O " + string.join(convert_gifs, " "))
        # print "GIF conversions complete"

    # Now check to see which conversions did not take
    failures = []
    for gif in convert_gifs:
        png = re.sub(r"\.gif$", r".png", gif)
        if not os.path.exists(png):
            failures.append(gif)
    if failures:
        print "Some conversions failed:", string.join(failures)

    # Now hack the references in the web pages
    for page in page_conversions.keys():
        print "Converting %s..." % (page,)
        if version_controlled(page):
            os.system("co -l " + page)
        else:
            os.system("cp -pf " + page + " " + page + ".bak")
        fp = open(page, "r")
        contents = fp.read()
        fp.close()
        basedir = basere.search(contents)
        if basedir:
            basedir = base.group(1)
        else:
            basedir = ""
        for (ref, target) in page_conversions[page]:
            if target in giflist and not target in failures:
                source = '(?P<g1>[="(\s])' + ref + r'\.gif(?P<g2>[\s"\\>)])'
                target = r"\g<g1>" + ref + r".png\g<g2>"
                print "Source: %s, target %s" % (source, target)
                contents = re.sub(source, target, contents)
        fp = open(page, "w")
        fp.write(contents)
        fp.close()
    print "Page conversions complete."

def cleanup(directory):
    # Clean up superfluous .gif and .bak files left over after a conversion
    map(os.unlink, findfiles(directory, bakre))
    pnglist = findfiles(directory, pngre)
    for png in pnglist:
        gif = png[:-4] + ".gif"
        if os.path.exists(gif):
            os.remove(gif)

def unconvert(directory):
    # Reverse a conversion    
    pnglist = findfiles(directory, pngre)
    for png in pnglist:
        gif = png[:-4] + ".gif"
        if os.path.exists(gif):
            os.unlink(png)
    htmls = findfiles(directory, pagere)
    for page in htmls:
        if os.path.exists(page + ".bak"):
            os.rename(page + ".bak", page)
        elif version_controlled(page):
            os.system("rcs -u " + page)

if __name__ == '__main__': 

    delete = nochange = reverse = verbose = 0
    transparency = ""

    (options, arguments) = getopt.getopt(sys.argv[1:], "adnrtv")

    if not arguments:
        arguments = ['.']

    for (switch, val) in options:
        if switch == '-a':
            all = 1
        elif switch == '-d':
            delete = 1
        elif switch == '-n':
            nochange = 1
        elif switch == '-r':
            reverse = 1
        elif switch == '-t':
            transparency = " -t "
        elif switch == '-v':
            verbose = 1

    if delete:
        map(cleanup, arguments)
    elif reverse:
        map(unconvert, arguments)
    else:
        map(web2png, arguments)

# The following sets edit modes for GNU EMACS
# Local Variables:
# mode:python
# End: