Make extracting verses more robust

import logging

import re

import os.path

import traceback

from bs4 import BeautifulSoup

from unidecode import unidecode

def reconstruct_verse_text_from_reading(reading):

    last_non_punct_token = [t for t in reading.tokens if not t.is_punct()]

    codepoints = [' ' for _ in range(last_non_punct_token.span[1])]

    try:

        codepoints = [' ' for _ in range(reading.tokens[-1].span[1])]

        for token in reading.tokens:

            codepoints[token.span[0]:token.span[1]] = token.text

    except Exception:

        print('ERROR reconstructing verse from reading {!r}'

              .format(reading))

        traceback.print_exc()

        codepoints = []

    return ''.join(codepoints)

import argparse

import json

import os

import traceback

import allzweckmesser as azm

def main(hypotactic_dir, top_out_dir, meters=['hexameter']):

    corpus = azm.corpus.HypotacticCorpus.from_directory(hypotactic_dir)

    for document in corpus.documents:

        print('Processing {}'.format(document.title))

        try:

            doc_out_dir = os.path.join(top_out_dir, document.title)

            os.makedirs(doc_out_dir, exist_ok=True)

            for meter in meters:

                    for line in document.get_lines_with_meter([meter])

                ]

                if verses_for_meter:

                with open(os.path.join(doc_out_dir, '{}.json'.format(meter)),

                    with open(os.path.join(doc_out_dir,

                                           '{}.json'.format(meter)),

                              'w') as f:

                        json.dump(verses_for_meter, f)

        except Exception:

            print('ERROR at document {}'.format(document.title))

            traceback.print_exc()

def parse_args_and_main():