emhass Adjust PV forecasts

This PR will propose a new method to adjust the current prediction for PV power production forecast. This can be very useful to adapt the prediction to local conditions. This should work for either specific conditions with your solar panel such as special shading problems, or even provide a correction based on your specific micro-climate conditions of your environment. We will use the already available methods using machine learning

Feb 27 '25 00:02 davidusb-geek

Codecov Report

Attention: Patch coverage is 88.33333% with 14 lines in your changes missing coverage. Please review.

Project coverage is 65.80%. Comparing base (ced5467) to head (bd12696). Report is 23 commits behind head on master.

Files with missing lines	Patch %	Lines
src/emhass/command_line.py	72.41%	8 Missing :warning:
src/emhass/forecast.py	92.98%	4 Missing :warning:
src/emhass/utils.py	93.10%	2 Missing :warning:

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #476      +/-   ##
==========================================
+ Coverage   65.24%   65.80%   +0.56%     
==========================================
  Files           8        8              
  Lines        3070     3150      +80     
==========================================
+ Hits         2003     2073      +70     
- Misses       1067     1077      +10

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

:rocket: New features to boost your workflow:

:snowflake: Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
:package: JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

Mar 01 '25 22:03 codecov[bot]

What is still needed for PV adjust: (UPDATE)

[x] ~~Need to create a common method to retrieve data from HA in command_line.py~~
- ~~Before we retrieved only for perfect optim and MPC (for the now functionality)~~
- ~~Now we will need to retrieve on the 3 cases: perfect, dayahead and MPC~~
- ~~See if the current implementations on perfect and MPC matches~~
[x] ~~Define the new needed parameters >> see test_forecast.py >> test_pv_forecast_adjust~~
[x] ~~The adjust_pv_forecast method in forecast.py still needs a data preparation method, similar to what is done on test_pv_forecast_adjust~~

Mar 01 '25 23:03 davidusb-geek

@davidusb-geek A thing that popped during an exeptionally sunny March week here in Belgium last week. If you set your inverter to curtail, the value in your sensor_power_photovoltaics will go down (for instance if you set your inverter to zero grid export when curtailmant has been foreseen by EMHASS). I guess that is something you have to consider when doing ML predictions on the difference between measured and predicted solar yield. I noticed I had to change my pv_power_forecast runtime parameter for mpc for this. I started with the actual value of sensor_power_photovoltaics as first value in the list but had to change that to solar+actual consuption (which should normally be the same as actual solar + predicted curtailment) to prevent flip flopping (curtailment on/off)