Rename arxiv papers with their original title. You have to put and run this script in the same folder as the pdfs to be renamed.

rename_arxiv.py

#!/usr/bin/env python3
"""
Rename arxiv papers with their original title. You have to put and run this script in the same folder as the pdfs to be renamed.

Author: Ignacio Heredia
Date: October 2017
"""

import os
import requests
import xmltodict
import re

homedir = os.path.dirname(os.path.realpath(__file__))

pattern = re.compile('^[0-9]{4}.[0-9]{5}$') #pattern of arxiv titles eg. '1708.07120'
for f in sorted(os.listdir(homedir)):
    if f.endswith(".pdf"):
        fname = f.split('.pdf')[0]
        fname = fname[:10]  # forget version information

        if pattern.match(fname) is None:
            print('{} is not an arxiv paper number format'.format(fname))
            continue

        try:
            url = 'http://export.arxiv.org/api/query?search_query={}'.format(fname)
            r = requests.get(url)
            d = xmltodict.parse(r.content)
            if type(d['feed']['entry']) is list:
                title = d['feed']['entry'][0]['title']
            else:
                title = d['feed']['entry']['title']
            os.rename(os.path.join(homedir, f), os.path.join(homedir, title + '.pdf'))
            print('{} --> {}'.format(fname, title))
            
        except:
            print('{} is not found in arxiv'.format(fname))
            pass


## Using the title metadata (problem: some PDFs don't have set correctly the metadata)

#from PyPDF2 import PdfFileWriter, PdfFileReader
#def get_pdf_title(pdf_file_path):
#    with open(pdf_file_path) as f:
#        pdf_reader = PdfFileReader(f) 
#        return pdf_reader.getDocumentInfo().title

#for f in sorted(os.listdir(homedir)):
#    if f.endswith(".pdf"):
#        try:
#            pdf_path = os.path.join(homedir, f)
#            print '{} --> {}'.format(f, get_pdf_title(pdf_path))
#        except:
#            pass