Open bjohare opened 8 years ago
Comment by adrivsh Thursday Jun 16, 2016 at 13:36 GMT
Leeway Let the user go a little below min or above max when changing the inputs. I suggest 0.1(M-m) as the margin, where m=min(param) *in the data and M=max(param) in the data
Single values For parameters for which m=M, that is there is only one value in the data, use the bounds provided in _inputsinfo for bounding the x axis, and draw an ad hoc rectangle or whatever to bypass the failure of the kde function to render single value "distributions"
Should the user be able to select a value using these 'fake' distributions?
well, yes. but i don't understand the question. in phl, "avoided losses with early warning" is a single value. the rendering of the distro is ugly, but it works and we can pick a value.
ASAIC, this is good enough for now, so not a prioty for you and you can focus on the docs
It seems that it does not work when the input can take negative values (it's only the case of exposure bias now). In that case, the leeway can be moved to a negative value, but when we run the model, it goes back to the zero position.
Issue by adrivsh Tuesday May 31, 2016 at 00:57 GMT Originally opened as https://github.com/worldbank-climate-group/resilience-indicator-tool-old/issues/9
Hazard and exposure bias seem to lack one