Discharge slot clipping

• Adjust discharge slots to have the % just below the expected
• Default to 1 pass discharge calculations (2 isn't needed)
• Add new metric_min_improvement_discharge which should normally be 0. Allows different value for charging.
• Adjust defaults in config template and update readme

Minor fixes and optimisations for discharge

• Align reported data to the correct simulated time (was + 5 minutes)
• Fix bug with soc_min handling
• Retain discharge window start where possible to avoid keep setting it
• Once discharge stops don't restart unless there is at least 5 minutes of SOC left
• Reset Charge rate during charge slots
• Reset Discharge rate during discharge slots
• Understand charge/discharge rate settings

Fix problem when on a fixed rate tariff

If there are no charge windows then avoid optimising then (and hence don't crash)

Fix problem with 100% charge when discharge was enabled

Fix problem with 100% charge when discharge was enabled, GivTCP treats 100% reserve as Eco (Paused) which confused the discharge code and overwrote the reserve back to 4%

Automatic conversion of text to boolean

It's possible to use a boolean config option connected to a switch or sensor, automatically convert the type

Add more configurability around charge and discharge calculation algorithms

• Fix bug with non-integer discharge %
• Stop discharge if soc% target is hit, and count that in the simulation
• Add more configurability around charge and discharge calculation algorithms
• Discharge calculates by default oldest slot first, while charge on newest first

Publish min, max, average and threshold rates into rates

Publish min, max, average and threshold rates into rates
Some code tidy up around configuration
Fix rounding to use round to nearest not ceiling

Fix bugs in REST mode

• Battery scaling not working correctly in REST mode
• If charge_limit was missing from config in REST mode then there was a crash

Fix bug in discharge only mode

In discharge only mode there was a crash, now fixed

Fix bug in workaround for incorrect battery size reporting

In the case where an incorrect battery sized was detected there was a typo leading to a crash