tests/UNIT_TESTS.md

In statuser: Statistical Tools Designed for End Users

Unit test index

This document lists testthat files under tests/testthat/, the R source files they primarily exercise, and short notes on what groups of tests cover. Test titles are quoted as in code (test_that("...", { ... })).

Test IDs

Each test_that() block is preceded by a comment of the form #<function>_<NNN> (for example #t.test2_001). The name is the R function under test (t.test2, plot_freq, scatter.gam, print.table2, …). Numbers are unique per function across the whole suite (assigned in sorted file order, with test-backend_equivalence.R processed last so primary test files keep the lowest IDs). To refresh IDs after edits, run Rscript tests/testthat/_annotate_test_ids.R from the package root.

test-backend_equivalence.R → cross-cutting (R/t.test2.R, R/lm2.R, R/interprobe.R, …)

Regression-style checks that statuser outputs stay aligned with reference implementations (base stats, estimatr, marginaleffects).

| Subset | What it checks | |--------|----------------| | t.test2 vs stats::t.test | Default and formula interfaces agree on core two-sample scenarios. | | lm2 vs estimatr::lm_robust | Coefficients and inference match for a representative linear model. | | interprobe | Johnson–Neyman / regions of significance and internal interaction p-values agree with marginaleffects-based references (binary x case). |

Helper code: tests/testthat/helper-backend_equivalence.R.

test-convert_to_sql.R → R/convert_to_sql.R

| Subset | What it checks | |--------|----------------| | File output | Writes a SQL file; optional INSERT generation. | | Schema | create_table behaviour. | | Types | Date and numeric column handling. | | Safety | Escaping single quotes in string data. | | Return value | Invisible return. |

test-desc_var.R → R/desc_var.R

| Subset | What it checks | |--------|----------------| | Basics | Correct summary stats; digits; expected column set. | | Inputs | Data frames; columns named like arguments (y, group); formula and multi-way grouping; sorting of results. | | Missing / empty | NA handling; empty groups; all-NA y; single observation per group. | | Diagnostics | Missing factor combinations; perfectly overlapping groupers. | | Errors | Non-numeric y; truly missing variables; evaluation order for df$col vs validation. |

test-format_pvalue.R → R/format_pvalue.R

| Subset | What it checks | |--------|----------------| | Formatting | Ordinary p-values; vector input; digits; include_p / “includes_p” variants. | | Rules | Threshold vs digits; boundary values; values outside ([0,1]); negative inputs. | | Edge cases | NA; empty vectors; single NA. |

test-gam_functions.R → R/scatter.gam.R, R/plot_gam.R, R/twolines.R

| Subset | What it checks | |--------|----------------| | scatter.gam | Runs with data-frame and formula inputs; data.dots, three.dots, k; invalid data errors; ylim via ... without duplicate-argument issues; sensible fitted values. | | plot_gam | Errors for non-GAM objects; predictor and quantile.others validation; structure; reasonable predictions; multiple predictors. | | twolines | graph, quiet, covariates; structure; U-shaped vs inverted-U detection. |

test-interprobe.R → R/interprobe.R (and helpers in R/interprobe_helpers.R)

| Subset | What it checks | |--------|----------------| | lm input | Structure; fitted model returned when estimated internally; APA printing unless quiet=TRUE. | | Binary x | ti-based GAM testing path (gam_results_testing). | | lm2 input | Works with lm2 fits. | | model / data routing | Fits lm2 when model = linear (or sentinel) with data; vector interface with model = linear. | | Keyword vs object | linear in calling env vs bare keyword; lm as keyword when appropriate. | | Errors | Messages preserve intended model name. |

test-list2.R → R/list2.R

| Subset | What it checks | |--------|----------------| | Naming | Names from symbols; explicit names; mixed named/unnamed. | | Edge cases | Single argument; empty call. | | Values | Preserved values; non-trivial expressions. |

test-lm2.R → R/lm2.R

| Subset | What it checks | |--------|----------------| | Defaults | Statuser-flavoured output; HC3 SEs by default. | | Coefficients | Standardized betas; agreement with lm / lm_robust; classical vs robust SE differences. | | ... | Extra args passed through to lm_robust. | | Predict | With/without newdata; with/without clustering; se.fit; alignment with lm_robust. | | Ecosystem | marginaleffects (predict, avg_slopes, comparisons); broom (tidy, glance, augment); modelsummary (single and multi-model). | | Diagnostics in print | Heteroskedasticity flags (! / !!!); collinearity on interactions (X); SE.cluster when clustered; cluster SE > robust when clustering matters. | | print.lm2 | notes toggles; Call line; summary stats; interaction notes. | | Mean column | Numeric vs factor level percentages. | | Data | Missing values; single/many predictors; with/without data (env vs explicit). | | Snapshots | Stable print output for representative cases. |

test-message2.R → R/message2.R

| Subset | What it checks | |--------|----------------| | Rendering | Default message; colours; font; stop; multi-part messages. | | Return | Invisible return. |

test-plot_cdf.R → R/plot_cdf.R

| Subset | What it checks | |--------|----------------| | Smoke / inputs | Runs; data frame; formula; df$col on LHS/RHS without bare names in env; missing values. | | Options | Custom args; show.ks; show.quantiles; group count variants. | | Return / errors | Object structure; dataset name in errors. | | Edge cases | Single group; all-NA; very small n per group. | | Order / factors | order, order = -1, factor levels when order is NULL. | | Two-vector mode | Basic and custom params; order; unequal lengths; legend margin; xlim padding. |

test-plot_density.R → R/plot_density.R

| Subset | What it checks | |--------|----------------| | Smoke / inputs | Same themes as plot_cdf: runs, data, formula, df$col, missing. | | Options | Custom parameters; show.means; passthrough to density(). | | Groups | Different counts; single vs many groups. | | Return / errors | Structure; dataset name in errors. | | Order / factors | order, -1, factor levels. | | Two-vector mode | Custom params; order; length mismatch; legend space. |

test-plot_freq.R → R/plot_freq.R

| Subset | What it checks | |--------|----------------| | Smoke / inputs | Runs; data; grouping; formula / df$col; freq and general custom args; value.labels, show.x.value, ticks.max, show.legend. | | Return / errors | Structure; dataset name in errors; missing values. | | Order | Named order, validation, three groups, default sort, order = -1 (including with factors). | | Two-vector mode | Structure, params, order, missing, validation, name deduction, data columns, continuous x, col/freq. | | Layout | xlim covers bars; pretty ticks for many levels; legend space. |

test-plot_means.R → R/plot_means.R

| Subset | What it checks | |--------|----------------| | Integration | Formula + data returns desc_var-like object; multi-way formulas (+); order; three-way grouping with sparse cells. | | tests = "auto" | Three scenarios (binary x1 only; binary x1 and x2; x2 with >2 levels). | | Numerical agreement | Cell means/CIs vs t.test (no clusters); Welch t in scenario 1; 2×2 interaction p vs lm_robust HC3; clustered CIs follow ci.level. |

test-print_functions.R → R/print_table2.R (methods for table2 objects)

| Subset | What it checks | |--------|----------------| | print.table2 | Frequency tables; proportion tables; three-way tables; tables without dimension names. |

test-resize_images.R → R/resize_images.R

| Subset | What it checks | |--------|----------------| | Robustness | Missing folder; empty folder. | | Parameters | width accepted. |

test-table2.R → R/table2.R

| Subset | What it checks | |--------|----------------| | Core | One- and multi-way tables; data columns; prop, digits, useNA, exclude. | | Chi-square | chi output; correction; match chisq.test / prop.test where applicable. | | Proportions | Row/column/ALL vs prop.table; row/column sums. | | Single vector | Structure; print; prop+chi prints chi once. | | Comparisons in cells | Logical expressions in 1D/2D; truncation; with prop and chi. | | Equivalence | Frequencies vs table(); freq snapshot stability for several print styles. |

test-text2.R → R/text2.R

| Subset | What it checks | |--------|----------------| | Layout | Alignments (incl. vectorized); padding; cex; extra graphics::text args. | | Styling | Per-label col; recycled bg; multiple labels. | | Edge cases | Small/special cases covered in the last block. |

test-t.test2.R → R/t.test2.R (and printing in R/print_t.test2.R)

| Subset | What it checks | |--------|----------------| | API | Export; class; column set. | | Inputs | Formula with/without data; df$col formula; two vectors; paired; one-sample. | | Column labels | Short names; long names; 0/1 as var=value; long names with 0/1. | | Missing data | Reporting for two-sample and paired. | | Equivalence | Matches stats::t.test (incl. paired, one-sample, formula, small n, extremes, factor/numeric grouping). | | Environment | Variables resolved from env when data omitted. | | Print snapshots | Stable printed output across designs (incl. missing and long names). |

test-validate_plot.R → R/validate.R (validate_plot())

| Subset | What it checks | |--------|----------------| | Happy paths | Standard (y, group); data; formula with/without data; optional group; df$var. | | Errors | Invalid inputs; messages include variable name, function name, and length mismatch detail. | | Edge cases | y ~ 1; explicit data_name. |

tests/testthat.R

Loads testthat and the package; not a test script itself.

Maintenance

After adding or renaming tests, regenerate this index (or extend the table for new test-*.R files). The canonical list of test files is:

list.files("tests/testthat", pattern = "^test-.*\\.R$").

statuser documentation built on April 25, 2026, 5:06 p.m.