feat: add log-points for NCA user input changes

[Gero1999]

Issue

Closes #1219

Description

Adds INFO-level logging for user input changes across the NCA setup modules so that exported session_log.txt files capture what configuration choices the user made — not just that a calculation happened.

New log-points added:

• Settings — analyte, specimen, profile, extrapolation method, min half-life points, flag rule enable/disable and threshold changes
• Data Imputation — BLQ strategy changes, C0 imputation toggle
• Parameter Selection — full parameter list at DEBUG level (count was already logged at INFO)
• Slope Selector — rule add/remove upgraded from TRACE→INFO, DEBUG summary of rule counts on table change
• General Exclusions — exclusion add (with row count, type, reason), removal, settings restore

Definition of Done

• Log messages appear in the R console when users change NCA settings, parameters, slope rules, or exclusions
• Messages are included in the ZIP-exported session_log.txt
• No logging on reactive invalidation noise — only meaningful user actions

How to test

1. Run aNCA::run_app()
2. Upload data and navigate to the NCA tab
3. Change settings (method, analyte, specimen, flag rules, etc.)
4. Add/remove slope rules and exclusions
5. Verify log messages appear in the R console for each action
6. Export ZIP and check session_log.txt contains the entries

Contributor checklist

• Code passes lintr checks
• Code passes all unit tests
• New logic covered by unit tests
• New logic is documented
• App or package changes are reflected in NEWS
• Package version is incremented
• R script works with the new implementation (if applicable)
• Settings upload works with the new implementation (if applicable)
• If any .scss change was done, run data-raw/compile_css.R

Notes to reviewer

This PR targets 1210-enhancement/replace-logger-with-console (#1212), not main. It will auto-retarget to main once #1212 is merged.

All log calls use the same log_info/log_debug functions introduced in #1212 — no new dependencies.

Version bump and NEWS entry still needed before merge.

[Copilot]

Pull request overview

This PR adds INFO/DEBUG logging across the Shiny NCA setup modules so exported session_log.txt captures user configuration changes (settings, imputation, parameter selection, slope rules, exclusions), improving post-hoc traceability of an NCA run.

Changes:

• Added log points for selection changes (analyte/specimen/profile, extrapolation method, min half-life points, flag rules).
• Added log points for BLQ and C0 imputation setting changes.
• Added more detailed logging for slope rule and exclusion add/remove/restore actions, plus DEBUG summaries.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
inst/shiny/modules/tab_nca/setup/slope_selector.R	Adds DEBUG summary log when manual slope rules table updates.
inst/shiny/modules/tab_nca/setup/settings.R	Adds INFO logs for key NCA settings and flag-rule changes.
inst/shiny/modules/tab_nca/setup/parameter_selection.R	Adds DEBUG log listing selected parameters per study type.
inst/shiny/modules/tab_nca/setup/manual_slopes_table.R	Upgrades add/remove slope rule logs from TRACE to INFO.
inst/shiny/modules/tab_nca/setup/general_exclusions.R	Adds INFO logs for exclusions restored/added/removed.
inst/shiny/modules/tab_nca/setup/data_imputation.R	Adds INFO logs for BLQ strategy and C0 imputation toggle changes.

[js3110]

Looks good, but there seems to be a lag on some of the widgets, especially the analyte pcspec and profile widgets- the log only appears after I change something else, not when I change the condition

[Gero1999]

@js3110 reactive is now on time also for those, let me know if there are others

[Shaakon35]

ignoreInit = TRUE prevents firing on first render, but input$method is still updated programmatically during settings restore via .restore_non_filter_settings → updateSelectInput(session, "method", ...). This will log "Extrapolation method changed" when the user uploads settings — which is a restore, not a user change.

The analyte, specimen, and profile observers already guard against this with filters_initialized(). Consider applying the same pattern here:

observeEvent(input$method, {
  if (filters_initialized()) {
    log_info("Extrapolation method changed: ", input$method)
  }
}, ignoreInit = TRUE)

observeEvent(input$min_hl_points, {
  if (filters_initialized()) {
    log_info("Min. half-life points changed: ", input$min_hl_points)
  }
}, ignoreInit = TRUE)

If logging during settings restore is intentional, add a comment noting that.

[Shaakon35]

Finding 6 — No unit tests for log output

The "New logic covered by unit tests" checklist item is unchecked. Unit testing Shiny observeEvent log calls is difficult without a test harness for the modules, so this isn't a blocker. However, if the logging functions (log_info, log_debug, log_trace) from #1212 support a capture/sink mechanism, a few smoke tests verifying that key log messages are emitted (e.g., via capture.output or a custom log collector) would add confidence against silent regressions.

Suggested approach if feasible:

it("logs analyte selection change", {
  # Use shiny::testServer or capture.output around the log function
  msgs <- capture.output(log_info("Analyte selection changed: AnalyteA"), type = "message")
  expect_true(any(grepl("Analyte selection changed", msgs)))
})

Not blocking — just noting for consideration.

[js3110]

Slope rules updates too many times- gives the same message wehn I:

• add exlcusion
• add reason (to same exclusion) (where it gives message 3 times)

[Gero1999]

@js3110 I think this one finally fixes it. It will show the message only whenever you make a change on the table (1 time if you just add a reason)

[js3110]

much better, thank you!