Motivation

When tring to build an autonomous site planning model, I was shocked by the backwardness of sunlight analysis technology in the Chinese real estate design industry:

  1. all available products are distributed as close-sourced "plugins" [用过的都知道~] in commercial desktop softwares, such as AutoCAD and Rhino Grasshopper. Worse yet, many are slow and not upgraded for years.

  2. no python package available makes it impossible to utilize DNN models, which nowadays are usually implemented on either tensorflow or pytorch, and nothing else.

The problem of sun position involves not only solar astronomy about space-time transformation between celestial and topocentric coordinates, but also convention in specific industries. It needs great effort for a single person to comprehend related knowledge scattered in various papers and documents. Thus FSEG is initiated as:

  1. a container of algorithms and know-how on sun position in different industries
  2. an easy-to-use API to relieve users from underlying astronomical and computational complexities
Back to top