Normalize the CSV exports generated by the ConQuest overlap template

Normalize the CSV exports generated by the ConQuest overlap template

Usage

normalize_conquest_overlap_exports(
  bundle,
  parameter_file,
  regression_file,
  covariance_file,
  case_file,
  delimiter = c("auto", "comma", "tab", "semicolon", ",", "\t", ";"),
  conquest_version = NULL,
  conquest_edition = NULL,
  run_date = NULL,
  report_file = NULL
)

Arguments

bundle

Output from build_conquest_overlap_bundle().

parameter_file

Path to the ConQuest export parameters CSV.

regression_file

Path to the ConQuest export reg_coefficients CSV.

covariance_file

Path to the ConQuest export covariance CSV.

case_file

Path to the ConQuest show cases EAP CSV.

delimiter

Delimiter used by all four files. The default, "auto", detects comma-, tab-, or semicolon-delimited files.

conquest_version

Optional ConQuest version recorded for provenance.

conquest_edition

Optional edition label, such as "demo/free", recorded for provenance.

run_date

Optional external-run date coercible to Date. Supply the date of the ConQuest run, not the normalization date.

report_file

Optional generated English show parameters review text. The estimation summary supplies stopping status and checks the person count, item model, covariate and ITEMS constraint. Missing reports leave external convergence unknown. This is not a general report parser.

Value

An object of class mfrm_conquest_overlap_tables with normalized population, item, and case-EAP tables.

Details

This helper is intentionally limited to the native CSV files requested by the command template from build_conquest_overlap_bundle(). It maps the exported regression rows to the bundle's intercept and covariate, maps the unidimensional covariance row to sigma2, trims padded person identifiers, and reconstructs the final item location under ConQuest's default sum-to-zero item constraint. It does not parse arbitrary ConQuest reports.

After normalization, pass the returned object directly to review_conquest_overlap(). Parameter and regressor indices must be unique integers in the generated model's coordinate range; row order may differ. Present dimension columns must identify dimension 1. Population variance must be finite and nonnegative; a zero boundary value is retained for descriptive review, not accepted as evidence of regular inference. Malformed exports are rejected before labels are assigned. This validation does not establish successful estimation. Present item labels must agree with the item map. IDs are read as text, including leading zeros, long integers and the literal ID NA. New bundle aliases are restored through person_map; unknown aliases are rejected. Keep all files from the same run: checks cannot establish that response cells or covariate values were unchanged. criterion_met describes the reported stopping rule, not integration accuracy or inferential readiness.

See also

normalize_conquest_overlap_files(), normalize_conquest_overlap_tables(), review_conquest_overlap()

Examples

if (FALSE) { # \dontrun{
bundle <- build_conquest_overlap_bundle(output_dir = "conquest-overlap")
# Run bundle$conquest_command in ConQuest, then:
cq <- normalize_conquest_overlap_exports(
  bundle,
  "conquest-overlap/conquest_overlap_conquest_parameters.csv",
  "conquest-overlap/conquest_overlap_conquest_reg_coefficients.csv",
  "conquest-overlap/conquest_overlap_conquest_covariance.csv",
  "conquest-overlap/conquest_overlap_conquest_cases_eap.csv",
  report_file = "conquest-overlap/conquest_overlap_conquest_parameters_review.txt"
)
review_conquest_overlap(bundle, cq)
} # }