get_citations.py

get_citations.py

import os
import urllib.request
import subprocess
import pandas as pd # pip install pandas
import betterbib # pip install betterbib
import bibtexparser  # pip install bibtexparser
from bibtexparser.bparser import BibTexParser


def fix_duplicated_entries(lines):
    out = ''
    entries = set()
    for line in lines:
        line = line.decode("ISO-8859-1")
        if '@article{' in line or '@phdthesis' in line:
            entry = line.split('{')[1].split(',')[0]
            orig = entry
            while entry in entries:
                entry += '_dup'
            entries.add(entry)
            if entry != orig:
                print(f'Duplicate entry: {orig} -> {entry}')
                line = line.replace(orig, entry)

            entries.add(entry)
        out += line
    return out


def download_bib(fname='citations.bib', url=None):
    if url is None:
        'URL not specificied. Default to downloading Jean-Remi King'
        url = 'https://scholar.googleusercontent.com/citations?view_op=export_citations&user='
        url += 'XZOgIwEAAAAJ&citsig=AMD79ooAAAAAYQJVOIWP0ghV_gx88PW2T3FK25bf0z9K&hl=en'
    data = urllib.request.urlopen(url)
    data = fix_duplicated_entries(data)
                
    with open(fname, 'w', encoding="ascii", errors='ignore') as f:
        f.write(data)
        

def enrich_bib(src, dest=None):
    print(f'Enrich {src}...')
    if dest is None:
        dest = src
    assert os.path.isfile(src)
    command = f'betterbib {src} {dest}'
    process = subprocess.Popen(command.split(), stdout=subprocess.PIPE)
    output, error = process.communicate()
    if error is not None:
        print(error.decode('utf-8'))
    else:
        output.decode('utf-8')

        
def bib_to_csv(src, dest=None):
    if dest is None:
        dest = src.replace('.bib', '.csv')
    
    parser = BibTexParser(common_strings=True)

    with open(src) as f:
        bib = bibtexparser.load(f, parser=parser)
        bib = pd.DataFrame(bib.entries)
    bib.to_csv(dest)
    return bib


def main():
    download_bib('citations.bib')
    enrich_bib('citations.bib', 'citations_rich.bib')
    csv = bib_to_csv('citations_rich.bib', 'citations_rich.csv')
    return csv


def lsp_format(bib, year=2021):
    
    bib.year = bib.year.fillna(0).astype(int)
    bib = bib.query(f'year>={year}')
    arxiv = bib[bib.journal.fillna('').str.contains('arXiv')]
    bib.loc[arxiv.index, "doi"] = arxiv.journal.fillna('').apply(lambda x: x.split(' preprint ')[1])
    bib.loc[arxiv.index, "journal"] = arxiv.journal.fillna('').apply(lambda x: x.split(' preprint ')[0])

    replacements = [
        (' and ', '; '),
        ('JR', r'Jean-Rémi'),
        ('Remi', r'Rémi'),
        ("{\\'e}", r'é'),
        ('{\\"u}', r'ü'),
        ('{\\"e}', r'ë'), 
        ('{\\c{c}}', r'ç'),
        ('{\^\i}', r'î'),
        ('{\`e}', r'è')
    ]

    for input, output in replacements:
        bib.author = bib.author.apply(lambda authors: authors.replace(input, output))


    def bold(t):
        return '\033[1m' + t + '\033[0m'


    for _, pub in bib.iterrows():
        authors = list()
        for author in pub.author.split('; '):
            if author == '':
                continue
            if author == 'others':
                authors.append('et al.')
            else:
                last, first = author.split(', ')
                first = '-'.join([f[0] for f in first.split('-')])
                authors.append(last +', ' + first +'.')
        authors = ', '.join(authors)
        
        title = pub.title.replace('{', '').replace('}', '')
        line = f'• {authors} ({pub.year}). {title}.'
        journal = 'journal' if not pd.isna(pub['journal']) else 'organization'
        for k in (journal, 'volume', 'pages', 'doi'):
            if pd.isna(pub[k]) or pub[k] == '':
                continue
            value = str(pub[k])
            if k == journal:
                value =  bold(value)
            if k == 'pages':
                value = value.replace('--', '-')
                value = ''.join([c for c in value if not c.isalpha()])
                start = [i for i, c in enumerate(value) if c.isdigit()]
                if not len(start):
                    continue
                value = 'pp.' + value[start[0]:]
            if k == 'doi':
                value = 'doi:' + value
            line += ', '+ value 
        print(line)
        print('')
        
if __name__ == '__main__':
    df = main()
    lsp_format(df)