add checkboxes for dose response curves for log transformation and increasing assumption

[tonywu1999]

Summary by CodeRabbit

• New Features
  • Added adjustable plot height slider for customized visualization sizing.
  • Introduced configurable response curve fitting options including logarithmic x-axis transformation and increasing trend specification.
  • Enhanced plot output rendering with improved UI architecture.

[coderabbitai]

📝 Walkthrough

Walkthrough

This change introduces new namespace constants for response-curve fitting options and plot visualization controls, refactors plot rendering from insertUI to renderUI, and converts hardcoded fitting parameters (log x-axis and increasing trend flags) to dynamic user-controlled inputs driven by UI elements.

Changes

Cohort / File(s)	Summary
Constants & Namespace R/constants.R	Added five new public namespace keys: modeling_response_curve_fitting_options, modeling_response_curve_increasing_trend, modeling_response_curve_log_xaxis, visualization_plot_output, and visualization_plot_height_slider.
Server Logic & Helpers R/module-statmodel-server.R	Introduced get_response_curve_fitting_options() helper function; refactored plot rendering from insertUI to renderUI with centralized output handling; updated dose-response fitting to read dynamic user inputs for log x-axis and increasing trend flags instead of hardcoded values; added plot height slider integration.
Modeling UI Options R/statmodel-ui-options-modeling.R	Added UI placeholder for response-curve fitting options; introduced two new checkbox UI helper functions (create_response_curve_log_xaxis_checkbox() and create_response_curve_increasing_trend_checkbox()) with tooltip support.
Results UI R/statmodel-ui-results.R	Updated plot output ID reference from literal string "comparison_plots" to namespaced constant NAMESPACE_STATMODEL$visualization_plot_output.
Utilities R/utils.R	Modified fitResponseCurves call to read increasing and transform_dose parameters from user inputs via namespace constants instead of hardcoded boolean values.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• feat(response-curve): Enable modeling and plotting of dose response curves #139: Modifies response-curve feature code by adding/using fitResponseCurves and related UI constants, overlapping in visualization and response-curve plotting infrastructure.
• feature(response-curve): Enable setting up metadata for response curves #133: Implements response-curve analysis support with constants, UI, and server wiring changes to the same files.
• refactor(statmodel): Set up a constants file for all namespace IDs #130: Extends the same NAMESPACE_STATMODEL constant and updates module UI/server code to use namespaced IDs across the codebase.

Suggested reviewers

• sszvetecz

Poem

🐰 A hop through constants, new and bright,
Log axes, trends—now user's might!
From hardcoded flags to choices free,
The response curve dances, wild and spry! ✨

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the primary changes: adding checkboxes for dose response curves with log transformation and increasing trend options.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch feat-log-axis

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

Failed to generate code suggestions for PR

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

R/module-statmodel-server.R (1)

433-449: ⚠️ Potential issue | 🟡 Minor

Provide a fallback default for the height parameter in groupComparisonPlots calls.

The height slider is conditionally rendered only when plot_type == Heatmap (line 843–844), and even then it doesn't exist on the first render. For Volcano and ComparisonPlot types, the slider never exists. This means input[[NAMESPACE_STATMODEL$visualization_plot_height_slider]] will be NULL in both cases. While MSstats::groupComparisonPlots has a default height = 600, passing NULL explicitly does not trigger that default and will likely cause an error.

Apply the fix to both line 433 and line 449:

Proposed fix

-        height = input[[NAMESPACE_STATMODEL$visualization_plot_height_slider]],
+        height = input[[NAMESPACE_STATMODEL$visualization_plot_height_slider]] %||% 500,

(Using %||% from rlang, or if (!is.null(...)) ... else 500 if rlang is unavailable)

🧹 Nitpick comments (3)

R/statmodel-ui-options-modeling.R (1)

17-17: Stale comment — the feature it requests is now implemented by this PR.

The comment # need option for increasing or decreasing trend for dose response describes exactly what this PR adds. Consider removing it.

Proposed fix

-    # need option for increasing or decreasing trend for dose response

R/module-statmodel-server.R (2)

797-848: Nested render* inside renderUI — functional but worth noting the trade-offs.

Creating renderPlot/renderPlotly outputs inside renderUI (lines 802, 814, 829) produces anonymous outputs. This works in Shiny, but has a caveat: every time the renderUI re-evaluates (e.g., when visualization_view_results changes), the anonymous outputs are destroyed and recreated, losing any client-side state.

Two additional notes on this block:

1. Line 810: matrix <- contrast$matrix shadows base::matrix. Not a bug here but a readability concern — consider renaming to e.g. contrast_mat.

2. Line 819: ratio_response = TRUE in visualizeResponseProtein, while fitResponseCurves (utils.R:1313) uses ratio_response = FALSE. If these are intentionally different (fitting vs. visualization), a brief comment explaining why would help future maintainers.

---

836-847: Height slider is created inside the same renderUI that reads its value.

The slider for heatmap height is emitted as part of the renderUI output, but the plot (output_plot) already read input[[visualization_plot_height_slider]] during this same render pass. On the very first render the slider doesn't exist yet so height is NULL. Once the slider appears, changes to it will re-trigger renderPlotly (since it's a reactive dependency inside the render function), so subsequent interactions work fine — it's only the initial render that gets NULL.

If you want the slider to drive height from the start, consider placing the sliderInput in a separate renderUI or in the static UI, rather than co-locating it with the plot output.