v1.2.0

Refactor

• Step: Return fees and cost from step info (f395d95)
• Step: Return fees and cost from step info (4e5b779)
• Render: Show figure is optional with env parameter (0a61245)
• reward: Better reward penalty for holding (f43b616)
• reward: Fees doesn't effect reward (64b23e9)
• step: Initial money as penalty then illegal move (ae4b31a)
• Observation Space: -INF lower bound (bb39b87)
• Reward: Get high reward on good trade and penalty on holding (04006bd)
• Reward: No total reward. Better reward on hold (8a719bc)
• Reward: Return accumulated total reward instead of step reward (02ba466)
• Step: Change return done into two separate done (71e579e)
• Step Function: Implementing legal action mask on every step (11f8e5c)
• Reward Function: Changed for better strategy (fca1e04)
• Observation Space: Added future price for window size (5d6043f)
• Env: Added custom feature size (ea48de5)

Feature

• Continuous Action space: Now action space is a regression space (cd86c6f)

v1.1.1

Fix

• Env: Datatype of obs should be float32 and info should be dict (fe6ef1b)

v1.1.0

Refactor

• Reward: Remove reward from plot/render (dbb4143)

Feature

• StockEnvV1: New environment for training (4c3b3f5)

Fix

• Setup: Typo of readme.md (1f60b8d)
• Profit Calculation: Latest profit not getting reset (688d38e)
• Max share: Not limiting stock holding (6be5f3f)
• sell reward: Should be multiplied not divided (f9474c1)
• Gym Make: Compatible for MuZero (6731737)

v1.0.1

Full Changelog: v1.0.0...v1.0.1

v1.0.0

Full Changelog: https://github.com/drlove2002/TradingGym/commits/v1.0.0