Darts Forecasting

WHENWhen constructing a TimeSeries with values array input

ACTIONExpand the values array to exactly 3 dimensions (time×component×sample) using numpy array shape operations

CONSEQUENCETimeSeries operations will fail with shape mismatches, causing downstream models to receive incorrectly formatted data tensors

WHENWhen setting the TimeSeries values dtype

ACTIONCast values to float64 or float32 to verify numerical precision for financial calculations

CONSEQUENCENumerical operations may produce incorrect results with integer or non-floating-point dtypes, leading to inaccurate forecasts

WHENWhen setting the TimeSeries time index

ACTIONVerify time index is strictly monotonically increasing with a well-defined frequency and no holes by default

CONSEQUENCEDownstream models will produce incorrect forecasts due to temporal misalignment, and gap detection methods will fail