title & authers of NIPS

NIPSscraping.py

# -*- coding: utf-8 -*-
"""
Created on Tue Dec 16 00:16:31 2014

@author: xiangze
"""
import glob

from BeautifulSoup import BeautifulSoup 

fns=["advances-in-neural-information-processing-systems-8-1995"]

for fname in fns:
    f=open(fname)
    soup=BeautifulSoup(f.read())
    papers=[]
    for l in  soup.findAll("li"):
        p={}
        p["author"]=[]
        for i in l.contents:
            si=str(i)

            if "author" in si:
                txt=i.text.encode("utf-8")
                p["author"].append(txt)
            elif "href" in si:
                txt=i.text.encode("utf-8")
                p["title"]=txt
        papers.append(p)
    print papers