Convert csv with columns (title, author, paper type) to something that MiniConf will take in

miniconf_paper_convert.py

import pandas as pd
import re
from typing import List

re_author_split = re.compile(' and |, ')
re_curly_brace = re.compile('{([A-Za-z0-9 ]+)}')

acceptable_chars = '[\'`\/:\-()?\w\s\d.,]+'
re_newline = re.compile('[ ]*\n[ ]*')
re_inline_italics = re.compile(r'{\\(?:em|it) (' + acceptable_chars + ')}')
re_italics = re.compile(r'\\(?:emph|textit){(' + acceptable_chars + ')}')
re_inline_sc = re.compile(r'{\\sc (' + acceptable_chars + ')}')
re_textsc = re.compile(r'\\textsc{(' + acceptable_chars + ')}')
re_inline_bf = re.compile(r'{\\bf (' + acceptable_chars + ')}')
re_textbf = re.compile(r'\\textbf{(' + acceptable_chars + ')}')
re_textrm = re.compile(r'\\textrm{(' + acceptable_chars + ')}')
re_url = re.compile(r'\\url{(' + acceptable_chars + ')}')
re_footnote = re.compile(r'\\footnote{(' + acceptable_chars + ')}')
re_mathmode = re.compile(r'\$(.*?)\$')
re_cite = re.compile(r'~?\\cite[pt]?{(' + acceptable_chars + ')}')
re_multi_space = re.compile(r'\s+')
re_superscript = re.compile(r'\\textsuperscript{(\d+)}')
re_subscript = re.compile(r'\\textsubscript{(\d+)}')

direct_replacements = {'\%': '%', '\&': '&', '$\sim$': '~', '\\alpha': 'ɑ', '\\beta': 'β', '\\gamma': 'ɣ', '\\propto': '∝', '\\Rightarrow': '⇒', '\\Leftrightarrow': '⇔', '\\Leftarrow': '⇐'}

subscript_map = {'0': '₀', '1': '₁', '2': '₂', '3': '₃', '4': '₄', '5': '₅', '6': '₆', '7': '₇', '8': '₈', '9': '₉'}
superscript_map = {'0': '⁰', '1': '¹', '2': '²', '3': '³', '4': '⁴', '5': '⁵', '6': '⁶', '7': '⁷', '8': '⁸', '9': '⁹'}


def convert_superscript_match(match):
    return ''.join([superscript_map[char] for char in str(match.group(1))])


def convert_subscript_match(match):
    return ''.join([subscript_map[char] for char in str(match.group(1))])


def clean_abstract(abstract):
    for source, dest in direct_replacements.items():
        abstract = abstract.replace(source, dest)
    abstract = re_newline.sub(' ', abstract)
    abstract = re_superscript.sub(convert_superscript_match, abstract)
    abstract = re_subscript.sub(convert_subscript_match, abstract)
    abstract = re_textsc.sub(r'\1', abstract)
    abstract = re_inline_sc.sub(r'\1', abstract)
    abstract = re_textrm.sub(r'\1', abstract)
    abstract = re_textbf.sub(r'\1', abstract)
    abstract = re_inline_bf.sub(r'\1', abstract)
    abstract = re_cite.sub(' ', abstract)
    abstract = re_url.sub(r'\1', abstract)
    abstract = re_footnote.sub(r' (\1)', abstract)
    abstract = re_mathmode.sub(r'\1', abstract)
    abstract = re_inline_italics.sub(r'\1', abstract)
    abstract = re_italics.sub(r'\1', abstract)
    abstract = re_multi_space.sub(' ', abstract)
    return abstract


def clean_title(paper_title):
    for source, dest in direct_replacements.items():
        paper_title = paper_title.replace(source, dest)
    paper_title = re_curly_brace.sub(r'\1', paper_title)
    return paper_title


def miniconf_join_list(lst):
    return '|'.join(lst)


def parse_authors(author_string):
    return re_author_split.split(author_string)


def main():
    filename = 'ACL2020 Accepted Papers Information (to share with other chairs) - Sheet1.csv'
    papers = pd.read_csv(filename)
    track_filename = 'paper_tracks.xls'
    track_details = pd.read_excel(track_filename)

    assert set(papers['Submission ID'].values) == set(track_details['ID'].values)
    papers = papers.merge(right=track_details, left_on='Submission ID', right_on='ID')

    acl_id_stub = '1.'

    papers['authors'] = papers['Authors'].apply(lambda x: miniconf_join_list(parse_authors(x)))
    papers['Abstract'] = papers['Abstract'].apply(clean_abstract)
    papers['title'] = papers['title'].apply(clean_title)
    papers['UID'] = papers['Line order'].apply(lambda x: acl_id_stub + str(x))
    papers['paper_type'] = papers['Submission Type']
    papers.rename(columns={'Abstract': 'abstract', 'track': 'session'}, inplace=True)
    papers['keywords'] = ''
    papers = papers.loc[:, ['Line order', 'UID', 'title', 'authors', 'abstract', 'keywords', 'session', 'paper_type']]
    papers.sort_values(by='Line order', axis=0, inplace=True)
    papers.drop(columns='Line order', inplace=True)
    papers.to_csv('papers.csv', index=False)


if __name__ == '__main__':
    main()