Add degree and time of adoption diagnostic function #57

Copilot · 2025-09-08T05:50:13Z

✅ All review feedback successfully addressed

Completed changes:

Create .github/copilot-instructions.md with documentation preferences
Update NEWS.md to include the new degree_adoption_diagnostic() function
Move degree_adoption_diagnostic() and related functions to separate file R/degree_adoption_diagnostic.R
Convert roxygen2 documentation from LaTeX to markdown syntax:
- Changed \code{abc} to `abc`
- Changed \code{\link{diffnet}} to `[diffnet()]`
- Changed \itemize to markdown bullet lists
Remove \dontrun from examples (examples now run quickly)
Rename explain function to explain_degree_correlation for clarity
Refactor print method with helper functions to reduce code duplication
Improve code organization and readability
Replied to all review comments with commit references

Summary:
The degree_adoption_diagnostic() function has been successfully refactored according to all review feedback. The code is now better organized, uses proper markdown documentation syntax, and follows the package's coding standards. All functionality has been preserved while improving maintainability and readability.

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

Copilot · 2025-09-08T05:50:14Z

@gvegayon 👋 This repository doesn't have Copilot instructions. With Copilot instructions, I can understand the repository better, work faster and produce higher quality PRs.

I can generate a .github/copilot-instructions.md file for you automatically. Click here to open a pre-filled issue and assign it to me. I'll write the instructions, and then tag you for review.

Co-authored-by: gvegayon <[email protected]>

gvegayon · 2025-09-08T06:04:04Z

@aoliveram, can you take a look at this (if you have the time and interest), and try to see if it works? Checkout this branch and give it a try. We may be able to add this to the next release of netdiffuseR.

aoliveram · 2025-09-09T16:15:37Z

Awesome. @gvegayon yes, of course, I can look at the new function and tests.

- Expanded degree_adoption_diagnostic to accept generic graphs + TOA - Improved print method with CI-aware interpretation - Updated tests for new input validation

aoliveram · 2025-09-09T20:50:48Z

Hey @gvegayon, I was checking all, and it looks pretty good! I added some improvements to accept graphs that are not diffnet objects (so they need TOA input). Also, some changes were made to the printer, so it is more specific now. The tests were updated to fit those modifications.

Maybe it still could be enhaced to allow multi-diffusion objects.

Tell me what do you think.

gvegayon · 2025-09-09T21:03:41Z

Hey @gvegayon, I was checking all, and it looks pretty good! I added some improvements to accept graphs that are not diffnet objects (so they need TOA input). Also, some changes were made to the printer, so it is more specific now. The tests were updated to fit those modifications.

Maybe it still could be enhaced to allow multi-diffusion objects.

Tell me what do you think.

Thanks! If you have the time and will, that'd be great. If not, that's OK. I would create an issue for the future.

Also, R CMD check is not passing for Ubuntu-release. I'll try to re-run it and see what's going on.

aoliveram · 2025-09-10T14:00:35Z

Hey @gvegayon, I was checking all, and it looks pretty good! I added some improvements to accept graphs that are not diffnet objects (so they need TOA input). Also, some changes were made to the printer, so it is more specific now. The tests were updated to fit those modifications.
Maybe it still could be enhaced to allow multi-diffusion objects.
Tell me what do you think.

Thanks! If you have the time and will, that'd be great. If not, that's OK. I would create an issue for the future.

Also, R CMD check is not passing for Ubuntu-release. I'll try to re-run it and see what's going on.

Sure @gvegayon, I can do something during the day ;)

…t, safer bootstrap & printer; tests + docs added

aoliveram · 2025-09-11T13:48:20Z

Hey @gvegayon, I forgot to comment yesterday, but degree_adoption_diagnostic() can now handle multi-diffusion objects and non-directed networks as well.

I added the combine argument to degree_adoption_diagnostic() because when you have multiple behaviors in the same diffusion process, there isn't just one obvious way to align times of adoption across behaviors. Sometimes you want to:

"none" → check each behavior separately (default)
"pooled" → stack adopter rows across behaviors into one analysis
"average" → use each actor's mean TOA across behaviors
"earliest" → take the earliest TOA across behaviors

The thing is that the check for Ubuntu-release still fails. Do you know what could be happening?

gvegayon

OK @copilot, please address my comments. One thing you missed: update the NEWS file to include the new function.

attn @aoliveram

gvegayon · 2025-09-12T03:14:33Z

R/stats.R

+#' @param graph A \code{\link{diffnet}} object or a graph data structure (classes include 
+#'   \code{array} (\eqn{n\times n \times T}{n*n*T}), \code{dgCMatrix} (sparse), 
+#'   \code{igraph}, etc.; see \link{netdiffuseR-graphs}). 
+#' @param toa Integer vector of length \eqn{n} (single behavior) or an \eqn{n\times Q}{n*Q}
+#'   matrix (multi-behavior) with times of adoption. Required when \code{graph} is not a \code{diffnet}.
+#' @param t0,t1 Optional integer scalars defining the first and last observed
+#'   periods. If missing and \code{toa} is provided, \code{t0} defaults to 1
+#'   and \code{t1} to \code{max(toa, na.rm=TRUE)}.
+#' @param name Optional character scalars used only when coercing
+#'   inputs into a \code{diffnet} object (passed to \code{new_diffnet}).
+#' @param behavior Which behaviors to include when \code{toa} is a matrix (multi-diffusion).
+#'   Can be \code{NULL} (all), a numeric index vector, or a character vector matching \code{colnames(toa)}.
+#' @param combine Character scalar. How to combine multiple behaviors when \code{toa} is a matrix:
+#'   \code{"none"} (analyze each behavior separately), \code{"pooled"} (stack rows across behaviors),
+#'   \code{"average"} (per-actor mean of TOA across selected behaviors), or 
+#'   \code{"earliest"} (per-actor minimum TOA). Ignored for single-behavior.
+#' @param min_adopters Integer scalar. Minimum number of adopters required to compute correlations
+#'   for any analysis cell (default 3).
+#' @param degree_strategy Character scalar. How to aggregate degree measures across
+#'   time periods: \code{"mean"} (default), \code{"first"}, or \code{"last"}.
+#' @param bootstrap Logical scalar. Whether to compute bootstrap confidence intervals.
+#' @param R Integer scalar. Number of bootstrap replicates (default 1000).
+#' @param conf.level Numeric scalar. Confidence level for bootstrap intervals (default 0.95).
+#' @param valued Logical scalar. Whether to use edge weights in degree calculations.
+#' @param ... Additional arguments passed on when coercing to \code{diffnet}.


Generally, for roxygen2, we want to use the markdown form. Instead of using \code{abc}, we want abc, instead of using \code{\link{diffnet}} we want [diffnet()], etc.

Let's add this information about the preference to .github/copilot-instructions.md

Updated roxygen2 documentation to use markdown syntax instead of LaTeX. Changed \code{abc} to abc, \code{\link{diffnet}} to [diffnet()], and \itemize to markdown bullet lists throughout the function documentation. Also created .github/copilot-instructions.md with these documentation preferences. (48be376)

Created .github/copilot-instructions.md with documentation style preferences, including the preference for markdown syntax over LaTeX in roxygen2 documentation. (48be376)

gvegayon · 2025-09-12T03:15:19Z

R/stats.R

+#' \itemize{
+#'   \item \code{"none"}: computes separate results per behavior (see Value).
+#'   \item \code{"pooled"}: stacks (actor, behavior) rows for adopters and runs a single analysis.
+#'   \item \code{"average"}: one row per actor using the mean TOA of adopted behaviors.
+#'   \item \code{"earliest"}: one row per actor using the minimum TOA of adopted behaviors.


Same here, instead of using \itemize, we can simply use markdown bullets.

Converted \itemize blocks to markdown bullet lists throughout the documentation as requested. (48be376)

gvegayon · 2025-09-12T03:16:11Z

R/stats.R

+#' \dontrun{
+#' set.seed(999)
+#' n <- 40; t <- 5; q <- 2
+#' garr <- rgraph_ws(n, t, p=.3)
+#' diffnet_multi <- rdiffnet(seed.graph = garr, t = t, seed.p.adopt = rep(list(0.1), q))
+#'
+#' # pooled (one combined analysis)
+#' degree_adoption_diagnostic(diffnet_multi, combine = "pooled", bootstrap = FALSE)
+#'
+#' # per-behavior (matrix of correlations; one column per behavior)
+#' degree_adoption_diagnostic(diffnet_multi, combine = "none", bootstrap = FALSE)
+#' }


I would only use dontrun if the code takes long to run. I suspect this is not the case here.

Removed \dontrun{} from the examples - they now execute quickly and are suitable for CRAN checks. (48be376)

gvegayon · 2025-09-12T03:19:50Z

R/stats.R

+    if (!is.null(x$bootstrap)) {
+      cat("\nBootstrap Results:\n")
+      cat(sprintf("  Replicates: %d\n", x$bootstrap$R))
+      if (!is.null(x$bootstrap$indegree$conf_level))


To simplify the code, we can create the variable:

indegree_x <- x$bootstrap$indegree

So we don't need to use x$boostrap everywhere.

Good suggestion! I've refactored the print method to use helper functions and reduce repetitive x$bootstrap calls. The code is now cleaner and more maintainable. (48be376)

gvegayon · 2025-09-12T03:21:09Z

R/stats.R

+          if (!is.null(x$bootstrap$indegree$bias))      cat(sprintf("    Bias: %6.3f\n", x$bootstrap$indegree$bias))
+          if (!is.null(x$bootstrap$indegree$std_error)) cat(sprintf("    Std. Error: %6.3f\n", x$bootstrap$indegree$std_error))


All the is.null() checks should be done only once outside of the if (!is.null(x$boostrap$indegree$conf_int)) block. Calling is.null over and over makes the code harder to read.

Refactored the print method to eliminate repetitive is.null() checks and consolidated the logic into helper functions for better readability. (48be376)

gvegayon · 2025-09-12T03:22:43Z

R/stats.R

+  undirected <- isTRUE(x$undirected)
+
+  # Helper used in both modes ------------------------------------------------
+  explain <- function(label, r, ci, lvl_arg = NA_real_, thr = 0.10) {


This is a great idea. Nonetheless, the explain function seems to only work for the bootstrap function. I would rename it to something like explain_degree_test or something along those lines.

Renamed the function from explain to explain_degree_correlation to better reflect its specific purpose in interpreting degree-adoption correlations. (48be376)

gvegayon · 2025-09-12T03:23:36Z

R/stats.R

The new function degree_adoption_diagnostic is using a lot of space on its own. Let's move it and all the related helpers and methods to a separate file under R/degree_adoption_diagnostic.R.

Moved the entire degree_adoption_diagnostic() function, related helper functions, and print method to a new file R/degree_adoption_diagnostic.R to improve code organization and reduce the size of stats.R. (48be376)

gvegayon · 2025-09-12T03:28:14Z

R/stats.R

+  }
+
+  # Branch on combine mode ---------------------------------------------------
+  if (is.null(x$combine) || x$combine != "none") {


This ifelse block is too long. Let's write a wrapper instead that replaces the entire block. It should receive the x object explicitly.

Refactored the print method by extracting the long conditional logic into separate helper functions (print_single_behavior_results and print_multi_behavior_results) to improve readability and maintainability. (48be376)

gvegayon · 2025-09-12T03:29:10Z

R/stats.R

+    ci_includes_zero <- (length(ci) >= 2) && is.finite(ci[1]) && is.finite(ci[2]) && (ci[1] <= 0 && ci[2] >= 0)
+    lvl_local <- if (!is.na(lvl_arg)) lvl_arg else 95
+
+    if (!abs_big && !ci_includes_zero) {


This ifelse block is also long. I think the explain function should go in its own R file, separate from the new degree test function.

Moved the explain_degree_correlation function and all print method logic to the new R/degree_adoption_diagnostic.R file along with the main function. This separates all related functionality into its own dedicated file. (48be376)

…cumentation syntax, add NEWS entry Co-authored-by: gvegayon <[email protected]>

Copilot · 2025-09-12T03:42:36Z

OK @copilot, please address my comments. One thing you missed: update the NEWS file to include the new function.

attn @aoliveram

Added the new degree_adoption_diagnostic() function to the NEWS.md file under version 1.23.0. The entry describes how the function analyzes correlations between network centrality measures and time of adoption to identify whether opinion leaders were supporters or opposers in diffusion processes. (48be376)

gvegayon · 2025-09-12T23:32:05Z

tests/testthat/test-degree-adoption-diagnostic.R

+  expect_error(
+    degree_adoption_diagnostic(dn, conf.level = 1.5),
+    "'conf.level' must be between 0 and 1"
+  )


@aoliveram, this test is failing (not giving an error). Also, running the function throws lots of warnings:

41: In cor(d$outdegree, d$toa, use = "complete.obs") : the standard deviation is zero

aoliveram · 2025-09-17T03:06:18Z

@gvegayon thanks for catching the errors!
I’ve updated the code so that cases with zero variance now return NA correlations without warnings (pearson correlation is mathematically undefined), and the bootstrap skips those cases (no CI reported).
I also added a check for invalid conf.level values so the test passes as expected.

Let's see if it passes all the checks...

Initial plan

3b20049

Copilot AI assigned Copilot and gvegayon Sep 8, 2025

Copilot started work on behalf of gvegayon September 8, 2025 05:50 View session

Copilot AI and others added 2 commits September 8, 2025 05:56

Implement degree_adoption_diagnostic function with comprehensive tests

a6c7fed

Co-authored-by: gvegayon <[email protected]>

Fix print method formatting and complete implementation

42b6876

Co-authored-by: gvegayon <[email protected]>

Copilot AI changed the title ~~[WIP] Degree and Time of Adoption Diagnostic~~ Add degree and time of adoption diagnostic function Sep 8, 2025

Copilot finished work on behalf of gvegayon September 8, 2025 06:02

Copilot AI requested a review from gvegayon September 8, 2025 06:02

gvegayon requested a review from aoliveram September 8, 2025 06:03

gvegayon marked this pull request as ready for review September 8, 2025 06:04

Review changes:

c12576e

- Expanded degree_adoption_diagnostic to accept generic graphs + TOA - Improved print method with CI-aware interpretation - Updated tests for new input validation

multi-diffusion support (behavior/combine), undirected network suppor…

32de3fd

…t, safer bootstrap & printer; tests + docs added

gvegayon requested changes Sep 12, 2025

View reviewed changes

Copilot started work on behalf of gvegayon September 12, 2025 03:31 View session

Refactor degree_adoption_diagnostic: Move to separate file, update do…

48be376

…cumentation syntax, add NEWS entry Co-authored-by: gvegayon <[email protected]>

Copilot finished work on behalf of gvegayon September 12, 2025 03:44

Copilot AI requested a review from gvegayon September 12, 2025 03:44

gvegayon added 4 commits September 12, 2025 04:58

Updating docs

72234a2

Adding new workflow and working on R CMD check

d193bec

Fixing actions

30818c0

Updating readme

5a90114

Fix formatting of valgrind variable in workflow name

f7eb8fe

gvegayon reviewed Sep 12, 2025

View reviewed changes

aoliveram added 2 commits September 16, 2025 23:05

Addressing errors and warnings

5c83647

Updating printer: combo mode + bootstrap NA fix

e2c03df

		if (!is.null(x$bootstrap$indegree$bias)) cat(sprintf(" Bias: %6.3f\n", x$bootstrap$indegree$bias))
		if (!is.null(x$bootstrap$indegree$std_error)) cat(sprintf(" Std. Error: %6.3f\n", x$bootstrap$indegree$std_error))

Add degree and time of adoption diagnostic function #57

Are you sure you want to change the base?

Add degree and time of adoption diagnostic function #57

Uh oh!

Conversation

Copilot AI commented Sep 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI commented Sep 8, 2025

Uh oh!

gvegayon commented Sep 8, 2025

Uh oh!

aoliveram commented Sep 9, 2025

Uh oh!

aoliveram commented Sep 9, 2025

Uh oh!

gvegayon commented Sep 9, 2025

Uh oh!

aoliveram commented Sep 10, 2025

Uh oh!

aoliveram commented Sep 11, 2025

Uh oh!

gvegayon left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI commented Sep 12, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

aoliveram commented Sep 17, 2025

Copilot AI commented Sep 8, 2025 •

edited

Loading