Parser for Wikimania 2019 xml to tsv (the tsv output format might since then have become outdated)

xml_parser.py

import sys
from xml.etree import ElementTree
import datetime
from collections import defaultdict


DAY_TITLE_FORMAT = '%A'  # '%A %d %B'
BLOCK_FORMAT = '{day}-session-block-{num}'  # saturday-session-block-1
TSV_FORMAR = '{id}\t{day}\t{title}\t"{abstract}"\t"{persons_list}"\t{persons}\t{room}\t\t{block}\t{space}\t{start}\t{end}\t{link}\t\tTRUE\t"{tags}"\t\t\t\n'

BLOCK_ENDS = {
    '2019-08-16': {'index': 1, 'end_times': ('12:00', '14:30')},
    '2019-08-17': {'index': 2, 'end_times': ('11:00', '13:00', '15:30')},
    '2019-08-18': {'index': 3, 'end_times': ('11:00', '13:00', '15:30')}
}


def get_next_day(days):
    next_day = None
    for day in days:
        if next_day is None:
            next_day = day
        else:
            next_date = strptime(next_day.get('date'), '%Y-%m-%d')
            date = strptime(day.get('date'), '%Y-%m-%d')
            if date < next_date:
                next_day = day
    return next_day


def strptime(*args):
    return datetime.datetime.strptime(*args)


def build_day_string(day):
    date = strptime(day.get('date'), '%Y-%m-%d')
    return date.strftime(DAY_TITLE_FORMAT)


def build_block_string(day, end, block_data):
    block_num = len(block_data) + 1
    for i, cmp_time in enumerate(block_data, 1):
        if end <= cmp_time:
            block_num = i
            break

    weekday_lc = build_day_string(day).lower()

    return BLOCK_FORMAT.format(day=weekday_lc, num=block_num)


def get_next_event(events):
    """
    Get the next from a list of event.

    Checks first for start time and then title alphabetically.
    """
    next_event = None
    for event in events:
        if next_event is None:
            next_event = event
        else:
            start = strptime(event.find('start').text, '%H:%M')
            next_start = strptime(next_event.find('start').text, '%H:%M')
            title = event.find('title').text
            next_title = next_event.find('title').text
            if start < next_start or \
               start == next_start and title < next_title:
                next_event = event
    return next_event


def create_csv_item(event, day, block_data):
    parameters = defaultdict(str)
    parameters['id'] = event.attrib.get('id')
    parameters['day'] = build_day_string(day)
    parameters['title'] = event.find('title').text
    parameters['abstract'] = event.find('title').text.replace('\n', '\r')

    presenters = get_all_presenters(event.find('persons'))
    parameters['persons_list'] = '\r'.join(presenters)
    parameters['persons'] = ', '.join(presenters)

    parameters['room'] = event.find('room').text

    space_element = event.find('space')
    if space_element is not None:
        parameters['space'] = space_element.text
    else:
        parameters['space'] = ''

    start = event.find('start').text
    parameters['start'] = start
    start_time = strptime(start, '%H:%M')
    duration_string = event.find('duration').text
    duration_time = strptime(duration_string, '%H:%M')
    end_time = start_time + datetime.timedelta(
        hours=duration_time.hour, minutes=duration_time.minute
    )
    parameters['end'] = end_time.strftime('%H:%M')

    parameters['block'] = build_block_string(day, end_time, block_data)

    parameters['link'] = event.find('links/link').get('href')

    parameters['tags'] = '\r'.join(event.find('identifiers').attrib.keys())  # not the right place for this, maybe abuse pathways
    return TSV_FORMAR.format(**parameters)


def get_all_presenters(persons_element):
    if persons_element is None:
        return []
    person_elements = persons_element.findall('person')
    return [p.text for p in person_elements]


def make_block_data():
    block_data = {}
    for v in BLOCK_ENDS.values():
        block_data[v.get('index')] = [strptime(time, '%H:%M')
                                      for time in v.get('end_times')]
    return block_data


if __name__ == '__main__':
    out_file = '{}.tsv'.format(sys.argv[1].rsplit('.')[0])
    with open(out_file, 'w') as f:
        block_data = make_block_data()
        tree = ElementTree.parse(sys.argv[1])
        root = tree.getroot()
        days = root.findall('.//day')
        day_index = 1
        while days:
            day = get_next_day(days)
            days.remove(day)
            events = day.findall('.//event')
            while events:
                event = get_next_event(events)
                events.remove(event)
                f.write(create_csv_item(event, day, block_data.get(day_index)))
            day_index += 1