Not able to reproduce lubridate values with vignette

[jayqi]

We have the following copied example for lubridate in the vignette.

Function	In-Subgraph Size	Coverage Ratio	Total Lines
divide_period_by_period	39	1	2
days	22	1	1
check_duration	15	0	1
as.POSIXt	13	0	1
eweeks	13	0	2
check_interval	12	0	11
date<-	12	NA	NA
add_months	10	1	4
ceil_multi_unit	10	1	1
am	6	1	1

I am not able to reproduce this result.

---

Installing the current CRAN version (0.3.2), and using the version of lubridate specified in the vignette, I get the following for the top 10 nodes by inSubgraphSize:

            node inSubgraphSize coverageRatio totalLines
 1:        month             82             1          1
 2:           tz             80             1          1
 3: reclass_date             69             1          1
 4:         date             68             1          1
 5:      is.Date             61             1          1
 6:    is.POSIXt             58             1          1
 7:         wday             57             1          1
 8:  .deprecated             56             0         10
 9:   is.POSIXct             56             1          1
10:      as_date             53             1          1

If I look at the same 10 nodes as the vignette, I get

                       node inSubgraphSize coverageRatio totalLines
 1: divide_period_by_period              1             1          2
 2:                    days             10             1          1
 3:          check_duration              1             0          3
 4:               as.POSIXt              1             0          1
 5:                  eweeks              1             0          2
 6:          check_interval              1             0         13
 7:                  date<-              1            NA         NA
 8:              add_months              3             1          4
 9:         ceil_multi_unit              5             1          1
10:                      am              2             1          1

I suspect what happened was two things:

1. Because we added support for non-exported functions (per FunctionReporter handling of non-exported internal functions #122, in PR FunctionReporter to handle Non-exported Functions #125), there are now a lot more connections in the graph, so the top ten nodes by inSubgraphSize are much bigger.
2. The same ten nodes have way lower inSubgraphSize because of bug Network Statistics are off #165 that was fixed in PR Bugfix Nodes Table Merging #166. In fact, divide_period_by_period is even mentioned in that issue.

I think we should just replace this table in the vignette with the new correct top 10 nodes. I will do it as part of PR #181.

[jayqi]

Additionally, because of the new non-exported functions from #125, the Jaccard Similarity example looking at groups with identical out-neighborhoods returns 20 groups instead of seven. It looks clunkier, but I'm just going to paste it in for now in #181 in order to have a more faithful reproduction of what the vignette is doing.

[bburns632]

Jay, as I recall, the PR that added UML conventions also added additional rigor to finding functional relationships. so it makes sense this PR is outdated. Let's version tag both the example package and pkgnet. We could use [devtools::install_version()]( https://www.rdocumentation.org/packages/devtools/versions/1.13.6/topics/install_version). Then, the vignette should remain reproducible. I'd prefer to keep the actual vignette static, rather that regenerated each time, as to avoid silent formatting errors.

…

On Fri, Feb 22, 2019 at 11:33 PM Jay Qi ***@***.***> wrote: Additionally, because of the new non-exported functions from #125 <#125>, the Jaccard Similarity example looking at groups with identical out-neighborhoods returns 20 groups instead of seven. It looks clunkier, but I'm just going to paste it in for now in #181 <#181> in order to have a more faithful reproduction of what the vignette is doing. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#189 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AXZRy2s_2492j6EE56q8TaJ-2mF2OJpdks5vQNK9gaJpZM4bNwCV> .

[jayqi]

I'm cool with keeping the results static. I was just saying we should update the static results with the new values that you'd get with our bug-fixed-new-functionality version of pkgnet. At a minimum, the old static results in the vignette are just plain wrong because of bug #165.

[bburns632]

Agreed. Let's update.

…

On Sat, Feb 23, 2019, 12:25 PM Jay Qi ***@***.***> wrote: I'm cool with keeping the results static. I was just saying we should update the static results with the new values that you'd get with our bug-fixed-new-functionality version of pkgnet. At a minimum, the old static results in the vignette are just plain wrong because of bug #165 <#165>. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#189 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AXZRywqwhVaDxAzKAXaXtwhlOtH54Ccxks5vQYeHgaJpZM4bNwCV> .