tipresias/candystore

# Releases

## v0.3.4

### Chores

- Remove dependency version limits: It's better to leave version limiting to end users to avoid conflicts with other packages. This also reduces maintenance work for the package.

## v0.3.3

### Chores

- Update dependencies

## v0.3.2

### Bug fixes

- Fixed typing for `numpy` >= 1.20

### Chores

- Update dependencies

## v0.3.1

### Chores

- Update dependencies

## v0.3.0

### Features
- Made round_type column in match_results data more realistic, with later rounds being labeled "Finals" instead of all rounds being labeled "Regular".

### Breaking Changes
- Changed the following data sets to match data generated by the fitzRoy R package
  - `betting_odds`
    - Removed `round_number` column, and made `round` column `int` type.
  - `players`
    - Made `round` `str` type, with text labels for finals rounds.
    - Fixed column names for away team quarterly score events (e.g. `aw1g` should be `aq1g`).
    - Made `pd.DataFrame` the default return type for data methods, because it's the data structure that most Python data people are used to working with, as well as being easier to work with than a list of dictionaries.

## v0.2.1

## Features
- `players` method to generate fake player data.
- `match_results` method to generate fake match results data.

## v0.2

## Features

- **BREAKING CHANGE**: Moved data factory functions to methods of the CandyStore class. This will make data sets compatible with each other, because they'll all be based on the same set of base match data.
- Added `betting_odds` data.

## v0.1.1

### Bug fixes

- Only include one Brisbane team per round to avoid potential duplicates based on how team names are normalised.

## v0.1

### Features

- Added factory function for fixture data.