#!/usr/bin/env python

#

# Add unique ID attributes to para tags.  This script should only be

# run by one person, since otherwise it introduces the possibility of

# chaotic conflicts among tags.

import glob, os, re, sys

tagged = re.compile('<para[^>]* id="x_([0-9a-f]+)"[^>]*>', re.M)

untagged = re.compile('<para>')

script_folder='/home/hg/repos/SEES-hacks/temp/'

names = glob.glob('/home/hg/repos/SEES-hacks/temp/ch*.docbook') 

# First pass: find the highest-numbered paragraph ID.

chapter=None

seen = set()

errs = 0

beginning="p_list= "

id_file=open(script_folder+'p_list.py','w') 

dictionary={}

id_list=[]

for name in names:

    for m in tagged.finditer(open(name).read()):

        i = int(m.group(1),16)

        if i in seen:

            print >> sys.stderr, '%s: duplication of ID %s' % (name, i)

            errs += 1

        seen.add(i)

        if i > biggest_id:

            biggest_id = i

def retag(s):

    global biggest_id

    global chapter   

    biggest_id += 1

    id_name="%s_%x" % (chapter.split('/')[-1],biggest_id)   

    id_list.append(id_name)    

    return '<para id="%s">' %id_name

# Second pass: add IDs to paragraphs that currently lack them.

for name in names:

    biggest_id=0 

    chapter=name.split('.')[0]    

    id_list=[]    

    f = open(name).read()

    f1 = untagged.sub(retag, f )

    dictionary[chapter.split('/')[-1]]=id_list       

    if f1 != f:

        tmpname = name + '.tmp'

        fp = open(tmpname, 'w')

        fp.write(f1)

        fp.close()

        os.rename(tmpname, name)

p_lists_string=beginning+str(dictionary)

id_file.write(p_lists_string)

sys.exit(errs)