Skip to content

Discuss: Configuration (take 2)

We have gradually organised and improved the handling of configuration and metadata into yaml files. Most recently metadata was divided into index metadata .yaml files, a separate one for index_function, and another for input variable specification.

We should consider to have one file collecting various overall configurations that now reside at various places in the code. Candidates might be:

  • MISSVAL (now a global?)
  • minimum allowed number, or proportion, of data points to calculate a percentile, which is relevant at least for index functions percentile and thresholded_percentile
  • the np.inf and -np.inf used to indicate "condition not met" in index functions first/last_occurrence et al.
  • "Instructions" for processing of global metadata (#36 (closed))? There are two main aspects to consider:
    • basic functionality should be working before climix v.1.0 is released (cf. attempt in !72 (closed))
    • progress on iris #3325 that now has been pushed to iris v.3.1
  • what schema is used for output tracking_id (cf. #16 (moved))
  • ...
Edited by Lars Bärring