Python - Map decode routine problem?

[james.duckworthy]

Python programmers may find the application to decoding an encrypted 
map 
image format known as Memory Map to produce a standard PNG image file 
interesting.

Someone obviously very well versed in Python and in the intricacies 
of image 
files has written a routine to decode originally the UK Landranger 
maps, so 
that they can be used with other mapping and GPS programs apart from 
the 
restricted Memory Map viewer.

You can get the details in this RAR archive.
http://www.mediafire.com/?3twm5xbozet

The application was obviously written with a specific job in mind, 
and has 
no GUI and is not very user friendly for people to use who are not 
Python 
experts.

One of the problems is that the original application was intended for 
UK 
Landranger map squares which are 20000 pixels square, and when 
decoded from 
Memory Map qct format to PNG format the files produced are around 50-
60 
megabytes which is quite manageable as a PNG image. PNG has to be 
fully 
loaded into RAM to manipulate, can't be paged from disk. But with a 
much 
larger qct image file, to decode it into a single PNG file would 
produce too 
big a file for most computers to handle when a say 200Mb PNG file on 
disk is 
decompressed into RAM.

So it is necessary to be able to decode large qct (Memory Map Image) 
files 
into map squares of a manageable size.

Easy enough when you know how! Problem is the existing Python routine 
starts 
decoding at the top left of the qct image file, and the only control 
in the 
routine that I have been able to decipher is the number of pixels 
along an 
edge of the square that are decoded which are controlled in the code 
line

extractsquare(row[2],row[1],int(row[3]),int(row[4]),20000)

near the end of the file qct2png.py

The last term 20000 sets the edge length of the decoded square.

The question is how can this routine be modified to start the 
decoding at 
pixel x,y, instead of at pixel 1,1 in the qct file, and to then 
decode a map 
square of say 20000 pixels edge length from that point, so that a qct 
image 
perhaps 60000 x 60000 pixels could be decoded into 9 squares of 20000 
x 
20000?

The original programmer has solved all of the complex problems of 
decrypting 
the encrypted file format and the routine works as intended, but with 
a much 
larger qct file than the originally intended target you are limited 
to 
decoding a square of about 30000 pixels edge depending on how much 
RAM your 
computer has to later handle the resultant PNG image file. Usually 
the PNG 
file needs to be further converted to a format which can be paged 
from disk 
such as MrSID or ozf2 which is a format used by a very popular GPS 
application called OziExplorer.

Does anyone know any "Python Pirates" who could solve this?

Could you post a suggested code alteration to do this to this 
newsgroup 
please?