#!/usr/local/bin/python
"""
mkdatabase - Recursively descend the texts and images directories
and build a dbm database mapping spunk-numbers to filenames.
Also check that all numbers exist.
"""

import sys
import os
import dbm
import regex
import string

DIRS=["texts", "images"]
DATABASE="indexdb"
MATCHPROG=regex.compile("sp\([0-9][0-9][0-9][0-9][0-9][0-9]\)\..*")
IGNORE=[".", "..", "index.txt"]

def main():
    dict = {}
    # Collect dictionaries
    for d in DIRS:
	os.path.walk(d, walker, dict)
    # Ckech that all keys exist
    keys = dict.keys()
    keys.sort()
    lastkey = string.atoi(keys[-1])
    for i in range(1, lastkey+1):
	thiskey = '%06.6d'%i
	if not dict.has_key(thiskey):
	    print 'Missing: ', thiskey
    # Copy dict to dbm database
    dbmdb = dbm.open(DATABASE, 'n')
    for k in keys:
	dbmdb[k] = dict[k]

def walker(dict, dirname, names):
    for name in names:
	fullname = os.path.join(dirname, name)
	if name in IGNORE or os.path.isdir(fullname):
	    continue
	if MATCHPROG.match(name) == len(name):
	    id = MATCHPROG.group(1)
	    if dict.has_key(id):
		print 'Duplicate:', dict[id]
		print '          ', fullname
	    else:
		dict[id] = fullname
	else:
	    print 'Unexpected:', fullname

main()