Sinergym

Patch Notes

- The name of this project and tool will be **Sinergym**!
- All documentation and script have been adaped to this new name.
- Third party accounts and tools have been adapted accordingly as well such as Google Cloud Platform, Docker Hub, container registry, etc.

GitHub Actions is here!

:x: **Travis** has been deactivated due to its temporal interruption service. (in any case, travis conf will be in repository)
:heavy_check_mark: **GitHub Action** will do this job instead and more!

Patch Notes
- There is a bot checking **autopep8** format in python code during pull request (automated commits).
- There is another bot **building** repository environment with *Dockerfile*, Executing **tests** with that and **updating Docker Hub the latest tag** if all tests have been passed.
- Release event is caught in order to **update Docker Hub** with specified tag for repository. (repared)

**EnergyPlus 9.5.0** can be used by Energym!

Patch Notes

- IDF files included in Energym have been updated using [EnergyPlus IDF converter](https://bigladdersoftware.com/epx/docs/9-5/auxiliary-programs/idf-version-updater.html). Thanks, miguems and jajimer for that information.
- This update to **Energyplus 9.5.0** changes some variable names such as `Facility Total HVAC Electric Demand Power`; its new name is `Facility Total HVAC Electricity Demand Rate`. This has been changed where needed.
- `opyplus` library doesn't contain the new IDD file for **Energyplus 9.5.0**, so this file has been manually specified in the code (using an ENV variable).
- The IDF update changes runperiods, so year start and end are specified now. The simulator has been updated in order to manage this correctly.
- The **Dockerfile** and **devcontainer** now use Energyplus version 9.5.0.
- Some **tests** have been updated in order to respect compatibility.

This is the first version of Energym!

*close 35, 37, 40, 41, 42, 43, 44*

Patch Notes:
- **Range getter** is developed in *energym/utils/common.py* in order to obtain absolute ranges in a set of Energym outputs. This information is used to normalize observation according to each type of environment.
- Added **normalized observation** to Tensorboard when environment is correctly wrapped.
- Improved **Tensorboard custom callback** code and developed compatibility with *OnPolicy/OffPolicy* algorithms
- Fixed some bug issues.
- More **docstring documentation**.
- Modified **gym action space** in continuous environments. Now is [-1,1] in all environments. Energym has **setpoint_ranges** for each value which will be transformed into Energym simulation action space correctly.
- New gym action space brings some bugs with **Rule Based Controller**. They have been solved.
- Adapted tests to this new repository version.

- Rule controller and Multi-observation wrapper
- Initial Documentation
- Initial tests
- Datacenter environments
- Dynamic observation/action spaces
- Agent-environment interaction logger