Fix S1/S2 correction and gas gain when simulating S1/S2PE

[mhliu0001]

This pull request proposes two changes:

1. Use S1/S2 correction calculated from TRUE positions rather than RECONSTRUCTED ones when we simulate PEs;
2. Add gas gain (x, y) dependence.

Should resolve #121.

Maybe someone should update the notebooks as well. Since the notebooks contain benchmark runs, I can't update them directly.

[github-actions]

Pull Request Test Coverage Report for Build 7276943356

• 44 of 44 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.2%) to 84.188%

---

Totals
Change from base Build 7267421814:	0.2%
Covered Lines:	1970
Relevant Lines:	2340

---

💛 - Coveralls

[zihaoxu98]

Thanks @mhliu0001 Only one suggestion. I think it's better to define the gas gain map as the relative map to parameter['gas_gain']. Then the placeholder map should be all identities. Then it won't have any conflicts with previous usage.