Fixed the NIPS crawler to work with the new format on the website

[gwenniger]

with the changed format of the newlines.

TODO: It seems like still further fixes to get/parse the reviews may be
necessary.

modified:   code/data_prepare/crawler/NIPS_crawl.py

[emaadmanzoor]

The NeurIPS crawler doesn't work on my end either (Python 2), the parsing logic (specifically, matching on <p>(.*)</p>) is incorrect. I was able to fix it by replacing that portion of the scraper with the following (uses BeautifulSoup4):

data = str(get_url(http, reviews_url))
...
soup = BeautifulSoup(data, 'html.parser')
...
if year < 2016:
...
else:
        review_divs = [div.text for div in
                       soup.find_all("div", style="white-space: pre-wrap;")]
        for review_div in review_divs:
            r = Review(None, review_div, None, None, None, None, None, None)
            reviews.append(r)
...