Filtering noise from time series data (precipitation/evaporation data)-Python

I have an annual data set of 3 soil tanks weights that were measured on minute basis. I need to calculate the precipitation and evaporation from these data but it contains gaps/ noise/ sensor failures. I need to prepare this excel table for the calculating program.
the sharp jumps should be removed and the gaps should be interpolated (an example of how it should look like at the end is attached on the picture).