Current summary adds all calls from any function to any other, which can include repeats of same from-to pairs. The current n_edges values extracted in summary statistics reports all of these, while the resultant visualisations generated by pkgcheck ultimately reduce them to unique from-to pairs only:
Does this reduced number of unique edges provide a better metric of network size? Could also provide summary statistics of n_edges and n_edges_unique, but that's likely to get confusing. Probably only need to be more explicit about what the summary statistic means (all edges, including repeated ones), and what the plot represents (unique edges only). Thoughts @noamross?
Current summary adds all calls from any function to any other, which can include repeats of same from-to pairs. The current
n_edges
values extracted in summary statistics reports all of these, while the resultant visualisations generated bypkgcheck
ultimately reduce them to unique from-to pairs only:https://github.com/ropensci-review-tools/pkgstats/blob/5f8e82b7c1e84424175fffe3ab1c38e60670acc2/R/plot.R#L60
Does this reduced number of unique edges provide a better metric of network size? Could also provide summary statistics of
n_edges
andn_edges_unique
, but that's likely to get confusing. Probably only need to be more explicit about what the summary statistic means (all edges, including repeated ones), and what the plot represents (unique edges only). Thoughts @noamross?