Re: [Gnumed-devel] help with trimming $

[catmat]

Horst Herb wrote:

> On Tue, 1 Mar 2005 14:05, Richard Terry wrote:
>  
>
> > eg 10/12/2004 Hypertension to return just Hypertension
> > or 10/2000 Appendix
> > 0r 2001 Cancer Bowel to return Cancer of the bowel
> >
> > Anyone volunteer the python code to use.
> >    
>
> For this I wouldn't even use regular expressions (which cost a lot of 
> processor time and memory)
> - all examples stated start with a number, so would just parse:
> if (line starts with number):
> 1.) read character until character not in [0-9, '.', '/', '-'] and append to 
> parsestr
> 2.) try split parsestr with date separators ('.', '/', '-')
> 3.) if only one split str and length ==4: this is the date (check plausibility 
> of date)
> 4.) else, if two split strings: assume first junk is months, last is year, 
> check for plausibility
> 5.) else, ...
>
> hey, it's quick and easy to do in Python, and since you are learning Python 
> just now, ... a good exercise in string manipulation / parsing?
>
> Horst
>  
I initially tried to use regex for parsing twiki markup, but in the end, 
I found
it was easier just thinking about characters, and lines, and using
a state machine state variable e.g.  INIT=0,  IN_NUMBERS=1 ,  
AFTER_NUMBERS=2
and using a big IF statement.
e.g. 
      state = INIT
      accumulators = ....

      for x in line:
         if state = INIT:
               do ...
               if condition : change state = IN_NUMBERS
         elif state= IN_NUMBERS:
               do...   if condition : change state=....
         elif  state= AFTER_NUMBERS:
                do

      process_accumulators