I changed the predict_quality() function as we discussed it. The data is loaded for the day(s) one day before the day(s) to be predicted. A "fake" record (full of zeros) is added for the (last) day to be predicted if it does not yet exist. This fixes the error that was reported by @fabianmoronzirfas. It occurred because rain data was available for today but river flow and wwtp flow were available only for yesterday. The merge led to NA values within the newdata data frame.
Feel free to add your "last" change (that you wanted to integrate before we make a new release) to this branch. Then, merge this pull request. Contact me if you need assistance.
Hi @wseis,
I changed the
predict_quality()function as we discussed it. The data is loaded for the day(s) one day before the day(s) to be predicted. A "fake" record (full of zeros) is added for the (last) day to be predicted if it does not yet exist. This fixes the error that was reported by @fabianmoronzirfas. It occurred because rain data was available for today but river flow and wwtp flow were available only for yesterday. The merge led toNAvalues within thenewdatadata frame.Feel free to add your "last" change (that you wanted to integrate before we make a new release) to this branch. Then, merge this pull request. Contact me if you need assistance.
Regards, Hauke
This change is