Remap action to fit gym's action space

[ChenDRAG]

See #312 for details.

[ChenDRAG]

The validity of this implementation is verified by my version of PPO as stated in #307.
I didn't check tests and docs, so it won't pass tests. Will fix this tomorrow.

[danagi]

Is the action of SAC mapped in forward and then mapped again in map_action?

[Trinkle23897]

Is the action of SAC mapped in forward and then mapped again in map_action?

just once, the code in forward is to correct the log_prob for entropy term

[danagi]

great job

[codecov-io]

Codecov Report

Merging #313 (ecff62b) into master (0c7117d) will increase coverage by 0.01%.
The diff coverage is 97.22%.

❗ Current head ecff62b differs from pull request most recent head 60dee40. Consider uploading reports for the commit 60dee40 to get more accurate results

@@            Coverage Diff             @@
##           master     #313      +/-   ##
==========================================
+ Coverage   93.91%   93.92%   +0.01%     
==========================================
  Files          51       51              
  Lines        3270     3278       +8     
==========================================
+ Hits         3071     3079       +8     
  Misses        199      199              

Flag	Coverage Δ
unittests	93.92% <97.22%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
tianshou/policy/modelfree/a2c.py	86.20% <ø> (ø)
tianshou/policy/modelfree/sac.py	84.88% <91.66%> (-0.84%)	⬇️
tianshou/data/collector.py	94.93% <100.00%> (+0.04%)	⬆️
tianshou/policy/base.py	79.56% <100.00%> (+2.32%)	⬆️
tianshou/policy/modelfree/ddpg.py	98.68% <100.00%> (-0.10%)	⬇️
tianshou/policy/modelfree/discrete_sac.py	87.69% <100.00%> (ø)
tianshou/policy/modelfree/pg.py	97.36% <100.00%> (ø)
tianshou/policy/modelfree/ppo.py	97.59% <100.00%> (+1.07%)	⬆️
tianshou/policy/modelfree/td3.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0c7117d...60dee40. Read the comment docs.