Updated CSS Selectors to websites and made scraping code slightly mor…

[jhchang]

…e robust.

I changed the html parser to html5lib, since I noticed html.parser was having quite some difficulty parsing certain articles. I also updated some CSS selectors for reuters.com and and oreilly.com because the original code had css selectors that did not select the right elements because website layout probably have changed since this code was last updated.

I noticed a few articles had some body text split up into multiple children to represent the paragraphs. So I made changes so the code loops through the children text instead of just the first child.

This is my first pull request and appreciate any feedback. Thank you