chore: update python requirements

feat: add lstm rl notebook example

fix: unify trade processing to correctly settle active orders and PnL

chore: git ignore train outputs

fix: rename dir and add requirements for python

fix: add L2Levels type alias to satisfy clippy::type_complexity

fix: update run_tests.sh

docs: add agent.md API reference

chore: rename project to TradingGYM

feat: add RL training examples (PPO & GRPO)

feat: add gymnasium env, ws l2 feed, and terminal dashboard

feat: enhance engine with market orders and L2 snapshot support

docs: remove release badge

docs: remove changelog, use github releases

fix: ci python job use maturin build, add badges