Version 0.4.1

Enhancements

• Added further documentation on how sportsreference can be used.
• Included explicit package versions to ensure a working configuration is always tested and verified.
• Added support for PyUp to automatically modify package versions when vulnerabilities or updates are identified.

Fixes

• Fixed an issue where the NBA schedule was only pulling games in a team's schedule that were already completed instead of the entire list of games.

Deprecations

• Python 3.4 has officially been dropped from the list of supported versions of Python due to it's end-of-life on March 16, 2019. Going forward, all users will be asked to switch to a supported version of Python.

Version 0.4.0

Enhancements

• Individual player stats are now included with boxscores to allow users to see metrics for every player that participated in a game.
• Added an optional slim parameter to the Roster classes to retrieve just the names and player ID's for every member of a team instead of pulling all of their stats. If just the names and ID's are desired, this significantly reduces the amount of time required to retrieve results.
• Updated documentation to include more advanced examples of ways to use sportsreference.

Fixes

• Fixed issues with integration tests attempting to test features that were out-of-scope. Now, tests are significantly faster.

Version 0.3.2

Fixes

• Fixed an issue where the Roster class would throw an error while attempting to pull a player that has no stats associated with him/her.
• Fixed an issue where NCAAF boxscore information, such as the date and location of a game, was stored in the incorrect property.

Version 0.3.1

Fixes

• Fixed an issue where an error was thrown in the Boxscores classes for any game that doesn't have a score associated with it, such as games in the future or games that occur on the current day but haven't been updated yet.
• Added missing parameters to the NCAA Boxscores classes to indicate the names and abbreviations of both the winners and losers, if applicable.

Version 0.3.0

Enhancements

• Expanded the amount of metrics of the Boxscores classes by including additional properties such as the final score, plus the names and abbreviations of both the winners and losers. For NCAAB and NCAAF, team rankings are also included when applicable, as well as a Boolean value to indicate whether the game includes at least one Top 25 team.
• The Boxscores class for all sports can now be iterated over a range with one call instead of just returning a single day's or week's worth of games. This new method is both easier to use and more performant by reducing the number of times the Boxscores class needs to be instantiated to get multiple dates.

Fixes

• Fixed an issue with the NFL documentation where an example on using the Boxscores class used invalid parameters which would throw an error when used in practice. The example has been updated with the correct parameters for calling the NFL Boxscores class.

Known Issues

• An error is thrown when pulling boxscores that don't have a score associated with it yet. The error is only thrown for all future boxscores as well as any games that occur on the current day. To avoid this issue, either avoid querying future boxscores or upgrade to a newer version of sportsreference.

Version 0.2.2

Fixes:

• Fixed an issue where NCAAB rankings always returned an empty dictionary

Version 0.2.1

Enhancements

• Improved the repository's community accessibility and availability by adding the following:
  • A Code of Conduct that everyone who interacts with this repository is required to follow
  • Pull Request, Issue, and Feature Request templates
  • A Contributing file which provides guidelines that all contributors should follow when creating updates
• Updated code style checker from PEP8 to pycodestyle

Fixes

• Fixed an issue parsing boxscore information for older games where less details are displayed, causing incorrect or mismatching information to be saved in various game information in the Boxscore class.
• Updated escape sequences for regex statements to enable the highest compatibility with future Python releases.
• Added specificity to all error handlers to ensure the code will provide meaningful information, or throw an error whenever an unexpected issue occurs.

Version 0.2.0

Enhancements

• Added the Player class for all sports which allows individual players to be queried for season and career stats.
• Added the Roster class which instantiates a Player for every member of a team's roster to allow entire teams to be searched.
• Added a Rankings class for NCAAB and NCAAF which provides weekly rankings from the Associated Press.
• Added a Conferences class for NCAAB and NCAAF which contains the name of the conference each college team participates in.
• Now calculating more stats in the NCAAB Team class to provide more parity with the Boxscore class.
• Added a property in the Game class with the boxscore's URI to allow all URI's to be pulled without instantiating the Boxscore class.
• Re-arranged the layout of the docs to enable easier navigation.

Fixes

• Fixed spelling issues in documentation.
• Fixed an issue where an error was thrown while parsing the NHL boxscore meta-information when no information is provided.
• Fixed an issue parsing an NCAAB team's ranking in the Boxscore class when the boxscores aren't present on the requested boxscore page.

Known Issues

• For more historic seasons (generally 5+ years ago), the boxscore pages contains less game meta-information (such as date, location, attendance, etc.) which causes the incorrect values to be parsed. This does not affect the statistical values (points, shots, fouls, yards, hits, etc.) for the boxscores. Until fixed, it is recommended to not rely on the game meta-information for seasons prior to the most recent for all sports.

Version 0.1.1

Enhancements

• Now using a universal PIP wheel instead of wheels specific to Python2 and Python3 as both use the same codebase.

Fixes

• Added error handling to all properties to prevent errors when incomplete/missing/invalid data is present on sports-reference.com.
• Replaced default value of all invalid properties with None to allow missing data to be distinguished from valid data where the result is zero (ie. A team that scored 0 touchdowns vs. a team that didn't have the number of touchdowns recorded).
• Fixed an issue where the dataframe_extended property would throw an error when no boxscores were created for a given year (seasons prior to 2010-2011 for NCAAB for example).
• Fixed an issue where NCAAB team records wouldn't be parsed when the advanced stats table wasn't included with the boxscore.
• Fixed an issue where missing NHL game meta-information on a boxscore would throw an error.

Known Issues

• For more historic seasons (generally 5+ years ago), the boxscore pages contains less game meta-information (such as date, location, attendance, etc.) which causes the incorrect values to be parsed. This does not affect the statistical values (points, shots, fouls, yards, hits, etc.) for the boxscores. Until fixed, it is recommended to not rely on the game meta-information for seasons prior to the most recent for all sports.

Version 0.1.0

Initial release of sportsreference

• Support for MLB, NBA, NCAAB, NCAAF, NFL, and NHL
• Supports queries for team, schedule, and boxscore information for all sports
• Supports Pandas DataFrames for easy parsing of all statistics for a team and/or game