Parse some XML

parsexml.py

from lxml import etree
import re
import csv
import os.path as op
import sys
import glob

class NLRPParse(object):

    def __init__(self,fname):
        
        prefix = op.splitext(op.basename(fname))[0]        
        loc = op.split(fname)[0]

        f = open(fname,'r')
        tree = etree.parse(f)

        fw = open(op.join(loc,prefix+'.csv'),'w')
        
        allcolumns = []
        allrows = tree.xpath('//row')

        # just get the column names from the first row
        # ASSUMES THEY ARE THE SAME FOR EACH ROW!
        for col in allrows[0]:
            allcolumns.append(col.tag)
                
        self.dw = csv.DictWriter(fw,allcolumns)
        self.dw.writeheader()

        allrows = tree.xpath('//row')
        for r in allrows:
            d = dict()
            for col in r:
                if col.text is not None:
                    bettertext = re.sub('\r','',col.text)
                    bettertext = re.sub('\n','',bettertext)
                    d[col.tag] = bettertext.strip().encode('utf-8')
                else:
                    d[col.tag] = ""
            self.dw.writerow(d)

        f.close()
        fw.close()

for a in glob.glob(sys.argv[1]):
    suffix = op.splitext(op.basename(a))[1]
    if suffix=='.xml':
        print "Parsing file: ", a
        p = NLRPParse(a)