Array Computing

PVGIS makes it easy to select among different array backends and floating point precision. For example, really massive time series may be processed in a distributed computing context. Using PVGIS with the default array-backend options, an end-user does not need to do anything.

Default array backend and data type

DATA_TYPE_DEFAULT='float32'

ARRAY_BACKEND_DEFAULT='numpy'

Array backend

Under development

The agnostic array-backend system is still incomplete. We are still evaluating the benefits of this implementation. Currently only the default NUMPY system is supported.

Higher precision

While the default output is of type float32

pvgis power broadband \
    8.628 45.812 214 180 45 \
    '2010-01-27 12:00:00'

789.574

if we ever need higher precision, we can ask for float64 through

pvgis power broadband \
    8.628 45.812 214 180 45 \
    '2010-01-27 12:00:00' \
    --dtype float64

789.574