Stats on ICML 2017 accepted papers

stats.py

import requests
from bs4 import BeautifulSoup
import re
import collections

# use lowercase
CANONICAL_NAMES = {
}

# use lowercase
CANONICAL_INSTITUTIONS = {
  'google deepmind': 'deepmind',
  'deep mind': 'deepmind',
  'Google Research, NY'.lower(): 'google',
  'Google Research'.lower(): 'google',
  'Google Brain'.lower(): 'google',
  'Google Inc.'.lower(): 'google',
  'cmu': 'carnegie mellon university',
  'microsoft research, india': 'microsoft research',
  'ucl': 'university college london',
  'iit kanpur': 'indian institute of technology kanpur',
  'duke': 'duke university',
  'mit': 'massachusetts institute of technology',
  'MIT CSAIL'.lower(): 'massachusetts institute of technology',
  'nyu': 'New York University'.lower(),
  'Georgia Tech'.lower(): 'Georgia Institute of Technology'.lower(),
  'Ecole Polytechnique de Montreal'.lower(): 'École Polytechnique de Montréal'.lower(),
}

def canonicalize(name, canonical):
  return canonical.get(name.lower(), name.lower())

class LazyLoadingPage(object):
    def __init__(self, url):
        self.url = url
        self._req = None
        self._soup = None

    @property
    def req(self):
        if self._req is None:
            self._req = requests.get(self.url)
        return self._req

    @property
    def soup(self):
        if self._soup is None:
            self._soup = BeautifulSoup(self.req.text, "html.parser")
        return self._soup

    def __str__(self):
        return "{cls}(url: {url})".format(
            cls=self.__class__.__name__,
            url=self.url)


class PapersList(LazyLoadingPage):
  @property
  def papers(self):
    return (Paper(p) for p in self.soup.find('main').find('div', class_='col-xs-9').find_all('p'))


class Paper(object):
  def __init__(self, soup):
    self.soup = soup

  @property
  def title(self):
    return self.soup.find('b').text

  @property
  def authors(self):
    return (Author(s.strip()) for s in self.soup.find('i').text.split('·'))

  @property
  def first_author(self):
    return next(self.authors)

  @property
  def last_author(self):
    return list(self.authors)[-1]


class Author(object):
  def __init__(self, name_inst):
    match = re.match(r'([^)]*) \(([^)]*)\)', name_inst)
    self.name = canonicalize(match.group(1), CANONICAL_NAMES)
    self.institution = canonicalize(match.group(2), CANONICAL_INSTITUTIONS)
    

def non_empty(seq):
  for x in seq:
    if len(x) != 0:
      yield x

def show_top_n(counter, n=10):
  for value, count in counter.most_common(n):
    print("{: 3} {}".format(count, value))

def take_n(seq, n=10):
  for i, x in enumerate(seq):
    if i > n:
      break
    yield x

ICML_URL = "https://2017.icml.cc/Conferences/2017/AcceptedPapersInitial"

def main():
  icml = PapersList(ICML_URL)

  print("------------------")
  print("Top first authors:\n")
  show_top_n(collections.Counter(
    paper.first_author.name for paper in icml.papers), 10)

  print("------------------")
  print("Top last authors:\n")
  show_top_n(collections.Counter(
    paper.last_author.name for paper in icml.papers), 10)

  print("------------------")
  print("Top institutions by first author:\n")
  show_top_n(collections.Counter(
    non_empty(paper.first_author.institution for paper in icml.papers)), 10)

  print("------------------")
  print("Top institutions by last author:\n")
  show_top_n(collections.Counter(
    non_empty(paper.last_author.institution for paper in icml.papers)), 10)

  print("------------------")
  print("Top institutions by all authors:\n")
  show_top_n(collections.Counter(
    non_empty(author.institution for paper in icml.papers for author in paper.authors)), 10)

  print("------------------")
  print("Top papers by number of authors:\n")
  for paper in take_n(sorted(icml.papers, key=lambda paper: -len(list(paper.authors)))):
    print("{: 3} {}".format(len(list(paper.authors)), paper.title))


if __name__ == '__main__':
  main()