artesiawater / waterbalans Goto Github PK

Request to extent e.plot.wq.loading():

1. if we measured the output waterflux at water stations (uitlaten), distinguish the measured output flux and the error (sluitfout) in the graph (we temporarily added the measured output flux in our last meeting as a red line).
   

2. visualize the lower limit of the assumed concentration together with the increment in the plot
   

See codacy for hotspots.

Remove FEWS related code from coverage calculating if possible (cannot test).

Add documentation, the basic configuration was set up, but there is very little content at the moment...

The idea is to help users link the output at different levels in the waterbalance package to data in the Excel water balance sheets.
Also helpful would be a mapping between the names.

Not sure if this is the way to go yet, but something along these lines:

python_excel_names = {
    "wb.buckets.Verhard": {
        "q_s" : "kwel (+) / wegzijging (-)",
        "q_no": "neerslag - verdamping"
        # ...
    }
}

python_excel_data = {
    "wb.buckets.Verhard": { 
        "q_s": "Rekenblad!A1", # note this is not correct, just a random example
        "q_no": "Rekenblad!B1", 
        # ...
   }
}

Replace with REST API, replace function get_fews_series() in timeseries.py

When adding timeseries (i.e. Constant timeseries, or ValueSeries or Local csv files) to the full model (as opposed to just one bucket) currently a value of -9999 has to be filled in as the BakjeID (a term we maybe should also update to something more english...). This -9999 value isn't particularly intuitive. Perhaps we can come up with something better.

Currently there isnt a clear method to simulate only a short period of time starting from custom storage levels. The initial storage parameters are applied at the first time step of any simulation. It would make sense to be able to store the current state of the system, and load that state back in when running a simulation starting from a certain date.

Perhaps this is as simple as exporting a parameter dataframe with all the hInits updated to the values of the storage at time step t.

These are not issues with the waterbalans package but issues that cause potential differences between the Excel and Python versions of the water balance calculations.

• hTargetMin and hTargetMax (i.e. min/max water levels over specific periods) are not always copied into the Rekenblad sheet correctly, occasionally (in some years) starting a day later than the data suggests. Perhaps this differs per Excel, but it was observed at least once.
• EFacMin and EFacMax are often listed in the Excel Rekenblad sheet for "Gedraineerd" buckets but are not used. The waterbalance package does not pick up this parameter if it is provided. Just a note for people comparing the Excel input to the Python parameters.

Pumping timeseries area plot should be stacked.

The Makkink to Penman factors in the original Excel balances have a much higher precision. These were updated with 27e37e8 . This should improve comparison between Excel balances and Python balances.

The flag to use the Excel factors or the original published factors is hardcoded. Perhaps something that should be exposed to the user...

Keeping issue open as a reminder to take a look at that.

When simulating EAG with QOutMax, this parameter is not passed to the numba calc_waterbalance method. When passing it, the result seems off. Probably a sign problem in the code.

Fix using Snelrewaard example.

Contact W+B about code to do this developed in Waterbalansen 't Hol project.

• Iteration to close waterbalance errors between EAGs
• Method to specify order, including branching?
• ...

Especially when reading data from files and using timeseries from files.

• Note when a timeseries/parameter is not used by a simulation.
• Indicate what type a timeseries is considered to be when it is processed.
• ...

E.g. for fraction plot, copy stacking order from Excel balance. Also apply similar color scheme.

Users starting with the Python waterbalance module are not always aware of the fact that all parameters have a default value.

I think this could be improved by no longer allowing parameters to be missing (by default):

1. Raise error on missing parameters unless use_default_params is set to True (default would be False).
2. Method to compare parameter input dataframe with all parameters in the model. I.e. e.parameter_comparison(params)
3. On use_default_params=True, add log message on simulate showing which parameters were not passed by user and for which default values were selected.

Add a release to PyPI to allow installation with pip