Basketball-reference-web-scraper

This bug fix, resolved by 111, updates the column index for the `game_score` value when parsing daily player box scores.

Basketball Reference added a plus/minus column _before_ the game score column, which caused the web scraper to return plus/minus values instead of the game score values.

Authored-by: ecallahan5 <ecallahan5gmail.com>

Resolves 99.

BBref added stats tables for quarters, which also had `basic` in the `id` value.


This meant that the number of matching footers was much higher.

Instead, I check for `game-basic` `id` values.

This probably requires a more robust solution in the future, but this is my strategy for now.

* Adds play-by-play stats scraping for pages like https://www.basketball-reference.com/boxscores/pbp/201810160BOS.html (#86) (DaiJunyan)
* Rewrites internal writing interface (89)
* Updates `README` to reflect new `play_by_play` client method and adds `Contributors` section (94)

Adds a `players_advanced_season_totals` method (84 and 85).

This method returns various advanced statistics associated with players in a given season.

Related to 81 and 82.

I had bumped `urllib3` to the latest version at the time (`1.25.2`) in 80 to resolve a security vulnerability in `urllib3 < 1.24.2`.

However, this meant that the required `requests` version, `2.20.0` had the wrong `urllib3` version.

`urllib31.24.3` both avoids the security vulnerability and fulfills `requests`' `urllib3` version requirements.

This release includes 79 and 80.

79 attempts to handle empty or blank string values for numerical fields while 80 upgrades the `urllib3` library to `1.25.2` due to a security vulnerability with versions less than `1.24.2`.