Python script to pull a bibtex reference from arxiv.

arxiref.py

#!/usr/bin/python3
import arxiv
import sys
import PyPDF4
import re

## Usage: arxiref.py 1711.07059 
##    or: arxiref.py paper.pdf
## If given something that looks like an arXiv id, searches for something with that name
## If given a filename, treats it as a pdf, looks for an arXiv id on page 1, then proceeds as in first case
## The produced bibtex code has "**" where you need to choose an id for the reference.


regex = "arXiv:([0-9]+\.[0-9]+)"
p = re.compile(regex)

def check_id(arg):
  """Test if input is an arXiv id"""
  dot = "." in arg
  numv = True
  for char in arg:
    if not(char in "1234567890v."):
      numv = False
  return dot and numv

def print_ref(arxiv_id):
  paper = arxiv.query(id_list=[arxiv_id])[0]
  print('''@Article{**
  Title = {%s},
  Author = {%s},
  Year = {%s},
  archivePrefix = "arXiv",
  eprint = {%s}
}''' % (paper['title']," and ".join(paper['authors']),paper['published'][:4],arxiv_id))

def get_id_file(filename):
  with open(filename, 'rb') as f:
    pdf = PyPDF4.PdfFileReader(f)
    text = pdf.getPage(0).extractText()
    m = p.search(text)
    return m.group(0)[6:]



argument = sys.argv[1]
if check_id(argument):
  print_ref(argument)
else: #argument is a filename
  print_ref(get_id_file(argument))