Bulbul's Python Tips

Caveat: I am not a professional programmer. I do this as a hobby. The following notes are strategies which i have either laboured to discover on my own, or found through my readings, in which case there will be due ackowledgements. I welcome constructive comments and criticism. E-mail me at bulbul@ucla.edu.

Detecting Forms

If your HTML uses both the 'GET' and 'POST' methods, it is important to detect which request method is used before calling cgi.FieldStorage(), because calling that method will fail to return a value (i.e., get stuck in an infinite loop) if the 'GET' method is used with an empty query string. The cgi.FieldStorage() method will not exhibit this undesired behaviour even when the form is empty. Therefore, code similar to the following should do the trick:

# This code sets the variable "form" to None when the form is non-existent or empty, 
# and to a cgi.FieldStorage object in all other cases.

import os, cgi

if ( os.environ['REQUEST_METHOD'] == 'GET' ) :

        if  os.environ['QUERY_STRING'] :
                # This will always yield a non-empty FieldStorage object.
                # That is, form.keys() will never be [].
                form = cgi.FieldStorage()

        else : 
                # The query string is empty.
                # Calling cgi.FieldStorage() here would enter an infinite loop.
                form = None

elif ( os.environ['REQUEST_METHOD'] == 'POST' ) :

        # This will yeild a non-empty FieldStorage object whose keys() 
        # may or may not be [].
        form = cgi.FieldStorage()

        # If the form is blank, it's probably easiest to make this the equivalent 
        # of there being no form at all.
        if not form.keys() :
                form = None

Adding a Missing `string.replace` Function

The Python 1.4 string module lacks the replace( str, old, new [, max]) function. The following code adds it to the imported string module just in case it isn't there:

import string     # Of course, don't import string again if you've already done so.

def replace( str, old, new, max=-1 ) :

	index = string.find( str, old )

	count = 0

	while ( index >= 0 ) and ( ( count < max ) or ( max < 0 ) ) :
		head = str[:index]
		tail = str[index+len(old):]
		str = head + new + tail
		count = count + 1
		index = string.find( str, old )

	return str


# We only want to use this new function if string.replace 
# doesn't already exist.

if 'replace' not in dir( string ) :

	string.replace = replace

Change in the Syntax of `append()`

Somewhere between versions 1.5 and 2.2 there has been a change in the syntax of the append method of a list. You can no longer append an arbitrary number of arguments to a list, so you can no longer do numberlist.append( 1, 2, 3 ). Instead you have to append() each item separately or use the + operator.

This fact is easy to discover in most scripts, but it can lead to mysterious failures in CGI scripts which used to work, where such an error is harder to pinpoint. (Two hours of my life i could have spent otherwise.)

Using `os.path.walk()`

I hate the fact that there's no example in the standard documentation on how to use os.path.walk(path, visitfunct, arg). I've written a little example in the form of a program that recursive calculates the size of a directory. View it here: dirsize.py.

To check the size of the current directory, just type dirsize.py. To check a different directory, type dirsize.py dirname, substituting the directory name.

Last updated June 18, 2005.