Package index • CausalStress

Run benchmarks

Discover, configure, and execute governed benchmark tasks.

cs_run_single(): Run a single DGP/estimator combination for one seed
cs_run_seeds(): Run a DGP x estimator combination over multiple seeds
cs_run_grid(): Run a DGP × estimator grid over multiple seeds
cs_run_suite(): Run a suite of DGPs with given estimators
cs_run_campaign(): Run a full campaign across DGPs, estimators, and seeds
cs_plan_campaign(): Plan a batched campaign
cs_run_batch(): Run a single batch from a campaign plan
cs_get_dgp(): Get a DGP descriptor by ID/version/status
cs_get_estimator(): Get an estimator descriptor by ID
cs_get_suite(): Get a suite definition
cs_suite_registry(): Suite registry
cs_register_estimator(): Register an additional estimator at runtime
cs_register_grf_dr_att(): Register GRF-based ATT estimator in the estimator registry
cs_set_rng(): Set CausalStress RNG state

Collect and interpret evidence

cs_collect_scores(): Collect canonical typed score records
cs_collect_att(): Collect ATT-level results from tidy runs
cs_collect_qst(): Collect QST-level results from tidy runs
cs_tidy(): Tidy CausalStress run results
cs_tidy_run(): Convert a single CausalStress run to a one-row tibble
cs_tidy_batch(): Tidy a batch of run results
cs_summarise_runs(): Summarise Monte Carlo runs for a DGP × estimator combination
cs_summarise_qst(): Summarise QST performance across runs
cs_summarise_gatekeeper(): Gatekeeper summary for placebo suites
cs_plot_att_error(): Plot ATT errors by estimator and DGP
cs_plot_qst(): Plot QST curves with confidence bands and truth overlay
cs_plot_placebo(): Plot placebo results (sharp-null checks)
cs_science_payload(): Extract the science payload from a run result
cs_provenance(): Extract provenance from a run result
cs_meta_flatten(): Flatten minimal identifiers for analysis

Persistence and audit

cs_consolidate(): Consolidate staged batch artifacts into a pins board
cs_gather_results(): Gather staged results and pin them to a board
cs_stage_result(): Stage a run result to the filesystem (atomic persistence)
cs_read_batch(): Read a batch pin from a board
cs_audit(): Audit pinned CausalStress results on a board
cs_delete_result(): Delete a single result pin
cs_delete_campaign(): Delete all persisted results for a DGP/estimator pair

Truth and validation

cs_true_att(): True ATT calculator
cs_true_qst(): QST truth from potential outcomes
cs_tau_oracle: Oracle quantile grid
cs_validate_tau_grid(): Validate that a truth tibble uses the canonical tau grid
cs_validate_dgp(): Validate a synthetic DGP for schema, determinism, and sanity
cs_validate_registry(): Validate all registered DGPs
cs_validate_dgp_registry(): Validate the DGP registry structure
cs_dgp_executable_meta(): Deterministic executable metadata for DGPs (manual mapping)

Included estimators

est_bart_att(): BART ATT estimator (bartCause)
est_gengc(): GenGC ATT + QST estimator (soft dependency)
est_gengc_dr(): Doubly robust GenGC estimator (soft dependency)
est_grf_dr_att() est_grf_dr(): GRF-based doubly-robust ATT estimator (causal forest)
est_ipw_att(): Inverse-probability weighted ATT estimator (IPW-ATT)
est_lm_att(): Linear outcome regression ATT estimator
est_oracle_att(): Oracle ATT estimator using structural treatment effects
est_tmle_att(): TMLE ATT estimator

Included DGP generators

dgp_synth_baseline_v130(): Baseline synthetic DGP for CausalStress (v1.3.0)
dgp_synth_baseline_v160(): Baseline synthetic DGP for CausalStress (v1.6.0)
dgp_synth_hd_sparse_plm_v130(): High-dimensional sparse partially linear DGP (v1.3.0)
dgp_synth_hd_sparse_plm_v140(): High-dimensional sparse partially linear DGP (v1.4.0)
dgp_synth_hd_sparse_plm_v150(): High-dimensional sparse partially linear DGP (v1.5.0)
dgp_synth_heavytail_v130(): Heavy-tailed synthetic DGP for CausalStress (v1.3.0)
dgp_synth_heavytail_v160(): Heavy-tailed synthetic DGP for CausalStress (v1.6.0)
dgp_synth_baseline() dgp_synth_hd_sparse_plm() dgp_synth_heavytail() dgp_synth_nonlinear_heteroskedastic() dgp_synth_overlap_stressed() dgp_synth_placebo_heavytail() dgp_synth_placebo_kangschafer() dgp_synth_placebo_nonlinear() dgp_synth_placebo_tau0() dgp_synth_placebo_tilted() dgp_synth_qte1() dgp_synth_tilt_mild(): Latest-version synthetic DGP wrappers
dgp_synth_nonlinear_heteroskedastic_v130(): Nonlinear heteroskedastic synthetic DGP for CausalStress (v1.3.0)
dgp_synth_nonlinear_heteroskedastic_v140(): Nonlinear heteroskedastic synthetic DGP for CausalStress (v1.4.0)
dgp_synth_nonlinear_heteroskedastic_v150(): Nonlinear heteroskedastic synthetic DGP for CausalStress (v1.5.0)
dgp_synth_nonlinear_heteroskedastic_v160(): Nonlinear heteroskedastic synthetic DGP for CausalStress (v1.6.0)
dgp_synth_overlap_stressed_v130(): Overlap-stressed synthetic DGP for CausalStress (v1.3.0)
dgp_synth_overlap_stressed_v140(): Overlap-stressed synthetic DGP for CausalStress (v1.4.0)
dgp_synth_overlap_stressed_v160(): Overlap-stressed synthetic DGP for CausalStress (v1.6.0)
dgp_synth_placebo_heavytail_v130(): Placebo heavy-tail synthetic DGP (sharp null, v1.3.0)
dgp_synth_placebo_kangschafer_v140(): Kang-Schafer placebo synthetic DGP (sharp null, v1.4.0)
dgp_synth_placebo_nonlinear_v130(): Placebo nonlinear synthetic DGP (sharp null, v1.3.0)
dgp_synth_placebo_tau0_v130(): Sharp-null placebo synthetic DGP (tau = 0, v1.3.0)
dgp_synth_placebo_tilted_v130(): Placebo tilted propensity synthetic DGP (sharp null, v1.3.0)
dgp_synth_placebo_tilted_v140(): Placebo tilted propensity synthetic DGP (sharp null, v1.4.0)
dgp_synth_qte1_v130(): Sign-flip QTE synthetic DGP (v1.3.0)
dgp_synth_qte1_v160(): Sign-flip QTE synthetic DGP (v1.6.0)
dgp_synth_tilt_mild_v130(): Mildly tilted propensity synthetic DGP for CausalStress (v1.3.0)
dgp_synth_tilt_mild_v160(): Mildly tilted propensity synthetic DGP for CausalStress (v1.6.0)

Package concepts

cs_ci_methods: Confidence interval methods (ci_method)
`%||%`: Null-coalescing helper