Changelog:
- Added timeout on curl requests to miners API in case their API gets stuck or the miner is stuck
- Increased tricking timeout to 5 seconds, just to allow more time to fix the gpu-stats.json file in case of curls failures mentioned in previous point
- Switched to PCIE Addresses instead of using gpu indexes id. This will ensure 100% correct temperatures in all mixed AMD/NVIDIA setups
- Small improvements / refactoring to source code