search

ProvTools / provR

GNU General Public License v2.0
7 stars 3 forks source link

Capture parentage of graphics calls #62

Open MKLau opened 6 years ago

MKLau commented 6 years ago

This is currently being worked on by B Lerner in RDataTracker.

blernermhc commented 6 years ago

I'm not sure what this is referring to. The development version of RDT supports this. Are there specific things missing from that?

MKLau commented 6 years ago

If this is implemented, that's great. Perhaps this has changed after I last tested. Previously, provenance was not monitoring the graphics device, but recording specific functions known to be involved in plotting (e.g. dev.off). If it's still the case that provenance is not being captured by monitoring their graphics environment, then I think we should implement this.

blernermhc commented 6 years ago

You are right that the current implementation looks at the function calls. It looks for every function in the default plotting package and could easily be extended to support ggplot2 or other plotting packages.

I do not know how to track the graphics device itself. I think the actual plotting is done in C code.

MKLau commented 6 years ago

This package (called recordr, but different from the one developed by NCEAS) that I've been reviewing is aimed at tracking R graphics: https://github.com/gmbecker/recordr

I haven't dug down into their code yet.

blernermhc commented 6 years ago

Looks interesting and non-trivial. I think it would be good to run their vignette to see what kind of output they get before diving in to try to understand their code.

https://github.com/gmbecker/recordr/blob/master/vignettes/index.Rmd

Is this the project that CRAN rejected?