webnlg.py

import os
import xml.etree.ElementTree as ET
import pandas as pd

tree = ET.parse('../data/webnlg2017/challenge_data_train_dev/train/2triples/2triples_Airport_train_challenge.xml')
root = tree.getroot()

all_elem = list(root.iter('entry'))

# entries
entries = [{
    "category": elem.attrib['category'],
    "eid": elem.attrib['eid'],
    "size": elem.attrib['size']
} for elem in all_elem]

entries_df = pd.DataFrame(entries)

# original tripleset
otriples = [
    {'eid': elem.attrib['eid'],
     'text': e.text} for e in elem.find('originaltripleset').findall('otriple') for elem in all_elem
]
otriples_df = pd.DataFrame(otriples)

# modified tripleset
mtriples = [
    {'eid': elem.attrib['eid'],
     'text': e.text} for e in elem.find('modifiedtripleset').findall('mtriple') for elem in all_elem
]

mtriples_df = pd.DataFrame(mtriples)

# lexes
lexes = [
    {'eid': elem.attrib['eid'],
     'text': e.text,
     'comment': e.attrib['comment'],
     'lid': e.attrib['lid']} for e in elem.findall('lex') for elem in all_elem
]

lexes_df = pd.DataFrame(lexes)