In python extending the string.Template class for easy regular expressions

Question

I like everything about python mostly thanks to its simplicity. I feel regular expressions have taken me away from my love for python far too many times. As such I wanted to extend the already existing string.Template class which lets me set variables in a string so that I can get values of an already existing string.

My first attempt works pretty well but has a few drawbacks:

import re
from string import Template
class TemplateX(Template):
    def getvalues(self,Str):
        regex = r""
        skipnext = False
        for i in self.template:
            if skipnext == False:
                if i != "$":
                    regex += i
                else:
                    regex += r"(.+)"
                    skipnext = True
            else:
                skipnext = False
        values = re.search(regex,Str).groups()
        return values

temp = TemplateX("  Coords;     $x;$y;$z;\n")
newstring = temp.substitute(x="1",y="2",z="3")
print newstring

values = temp.getvalues(newstring)
print values

newstring prints as: " Coords; 1;2;3;\n"

values prints as: ("1","2","3")

I am fine with losing some functionality of re for this simpler approach. My question is how can i add a little more functionality to getvalues to allow for variables in TemplateX to be more than 1 character (like the Template class and substitute allows). i.e. so that this works:

temp = TemplateX("  Coords;     $xvar;$yvar;$zvar;\n")
newstring = temp.substitute(xvar="1",yvar="2",zvar="3")
print newstring

values = temp.getvalues(newstring)
print values

temp2 = TemplateX("  FindThese:    $gx = $gy - $gz")
values2 = temp2.getvalues("  FindThese:    $10000 = $10 - $5x")

seems like i almost need to pass the variables with getvalues so i can write an re that can determine what the template is and what the variables are.. especially in the second case, am i right? — Symon
– Symon, Commented Aug 31, 2011 at 20:28

rczajka · Accepted Answer · 2011-08-31 21:05:05Z

3

You're pretty much trying to turn a template into a regex reading that template, right? That's not always doable, for example if you have a template "$x$y" and a string "abc", you can't know whether it's ("ab", "c"), ("abc", "") or some other case. And it's not just the case with template variables sitting next to each other.

So if you want to this kind of stuff, you have to know what's in the strings you feed to it. That is – that there are some characters between the variables that don't occur in the variables themselves (i.e. variables are all digits and there are semicolons between them in the template). If you're just parsing some input, you'd probably be safer if you just wrote your own regexes.

That said, you can do something like this:

class TemplateX(Template):
    def getvalues(self,Str):
        """ Reads a string matching the template to find the original values.

            >>> temp = TemplateX("Blah: $xx;$y;")
            >>> newsting = temp.substitute(xx="1",y="2")
            >>> temp.getvalues(newstring)
            ('1', '2')

        """
        regex = re.sub(self.pattern, "(.*)", self.template)
        m = re.match(regex, Str)
        return m.groups()

It uses Template's own mechanism for finding the placeholders and replaces them with regex wildcards. Then it runs the regex on the given string.

You could also try to get the initial keyword arguments:

    def getvalues(self,Str):
        regex = re.sub(self.pattern, r"(?P<\2>.*)", self.template)
        m = re.match(regex, Str)
        return m.groupdict()

And then choke it with:

temp = TemplateX("$a$a")
newstring = temp.substitute(a='a')
print temp.getvalues(newstring)

edited Aug 31, 2011 at 21:05

answered Aug 31, 2011 at 20:51

rczajka

1,85014 silver badges13 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Symon Over a year ago

that is exactly what i was looking for, thanks a lot! I realize i can't get the full functinality out of re with this extension of template but it's nice to have an easier more pythonic tool for doing simple regular expressions! :)

Symon Over a year ago

I think it might even make a good feature request for PEP 42.. if you feel like submitting it. (even though there are a few limitations to it).

rczajka Over a year ago

I don't really think it's more Pythonic. It's not shorter than regex, it gives you way less control, and it's too easy to silently return bad values instead of failing loudly. Reversing Python's .format strings might be fun, though. :)

grisevg · Accepted Answer · 2011-08-31 20:20:58Z

0

Use UserString for a super class http://docs.python.org/library/userdict.html#module-UserString

answered Aug 31, 2011 at 20:20

grisevg

2504 silver badges20 bronze badges

Collectives™ on Stack Overflow

In python extending the string.Template class for easy regular expressions

2 Answers 2

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related