You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some serious thought needs to go into how this might work, but the idea is that updating the series data with push! or setindex! would first apply applicable recipes before updating the data. To get this right, we would need to store:
a history of transformations to apply to the data
when multiple series are created, the series that the final data subset should be mapped to.
As an example, if we push to a histogram series, then it should know that the transformations (recipes) were "histogram --> bar --> filled path". We'd need to update the original data with a push/set and then recompute the transformations.
In short, this is doable and very powerful, but needs some real thought and some additional machinery to get it right.
Brainstorm: can we simply store the original args/kw with the series, update the args with the values from the push/append/setindex call, and then re-generate the series? Then we might be able to generalize to a update!(series::Series, args...; kw...), and might solve things like #62 in the process.
Moving from JuliaPlots/RecipesBase.jl#8.
Some serious thought needs to go into how this might work, but the idea is that updating the series data with
push!
orsetindex!
would first apply applicable recipes before updating the data. To get this right, we would need to store:As an example, if we push to a
histogram
series, then it should know that the transformations (recipes) were "histogram --> bar --> filled path". We'd need to update the original data with a push/set and then recompute the transformations.In short, this is doable and very powerful, but needs some real thought and some additional machinery to get it right.
Brainstorm: can we simply store the original args/kw with the series, update the args with the values from the push/append/setindex call, and then re-generate the series? Then we might be able to generalize to a
update!(series::Series, args...; kw...)
, and might solve things like #62 in the process.cc: @joshday
The text was updated successfully, but these errors were encountered: