Trying to get cleaner XML output from a text file

[Gabriel Genellina]

En Fri, 29 May 2009 14:09:10 -0300, iainemsley <iainemsley at googlemail.com>  
escribió:

> I'm using Python2.5 to try and convert some text files into XML using
> xml.minidom. I'm currently doing some plays which have a structure
> like
> Scene 1
> Act 1
> blah blah
> Act2
> blah blah
> Scene 2
> Act 1
> and so on.

(I think you get the hierarchy wrong: usually a play contains some Acts;  
each act contains several Scenes)

> I'm trying to turn it into
> <div type="scene">1
>   <div type="act">1
>    <speech />
>   </div>
>   <div type="act">2
>    <speech />
>   </div>
> </div>
> (or ideally <div type="scene" id="1"> bit I can always come back to
> this later)

Using ElementTree is a lot easier than minidom:

import sys
 from itertools import groupby, count
import xml.etree.ElementTree as ET
import re

class Seq:
     "Automatic sequencer for acts/scenes"
     def __init__(self):
         self.act_nr = count(1)
         self.scene_nr = count(1)
     def next_scene(self):
         return self.scene_nr.next()
     def next_act(self):
         self.scene_nr = count(1)
         return self.act_nr.next()
seq = Seq()

def add_act(body, act_text):
     act = ET.SubElement(body, "div", type="act", id="a%s" % seq.next_act())
     for scene_text in scene_sep.split(act_text):
         add_scene(act, scene_text)

def add_scene(act, scene_text):
     scene = ET.SubElement(act, "div", type="scene", id="%ss%s" %  
(act.get("id"), seq.next_scene()))
     for p in paragraphs(scene_text.splitlines(True)):
         add_speech(scene, p)

def add_speech(scene, p):
     speech = ET.SubElement(scene, "speech")
     speech.text = p

body = ET.Element("body")
scene_sep = re.compile(r"Scene\s+[1-9]+", re.I)
act_sep = re.compile(r"Act\s+[1-9]+", re.I)
for act_text in act_sep.split(text):
     add_act(body, act_text)

doc = ET.ElementTree(body)
doc.write(sys.stdout)


-- 
Gabriel Genellina