Envpool

Enhancement

1. add cleanrl example (48, 53)
2. add appo example (51)
3. export raw reward in atari env with `info["reward"]` (50)

New Env

Add toy_text (41, 42, 43, 44, 45, 46)

Enhancement

1. Add SB3 example (39)
2. Add gray_scale option in atari (37)

New Env

1. finish porting classic_control (25, 34)

Bug Fix

1. Atari speed and correctness (21, 35)
2. avoid overwriting config when registering multiple envs (29)
3. remove -1 in python spec, otherwise it cannot use `action_spaec.sample()` (34)
4. change abc_meta in py_envpool.h to support multiple pybind11 registrations in one file (34)

Documentation

1. how to integrate new env into envpool (31, 32)

Enhancement

1. add ci pipeline for generating wheels (36)
1. update issue templates
1. add a warning when calling `envpool.seed` function (22)
2. raise ValueError when `batch_size > num_envs` (22)
3. move `reward` to `common_state_spec` (22)
4. add `reward_threshold` in config and will auto pop up in `env.spec.reward_threshold` (24)
5. split registration to per-env (26)

Bug Fix

1. fix atari env cv2::resize bug: previous code doesn't use INTER_AREA correctly (13)
2. fix black screen in atari env (15)

Enhancement

1. add dockerfile for develop env and release env: `make docker-dev` and `make docker-release` respectively (11)
2. upgrade ALE to 0.7.3 (15)
3. fix auditwheel purelib warning (12)
4. add repeat_action_probability (15)

First open-source EnvPool version. Only include Atari environments but reserve multi-agent APIs for further integration.