
Validate real eye-tracking exports and compatibility corpora
eyeprocess-format-validation.RdBuild reproducible evidence for source-format compatibility. The validation framework distinguishes declared support, synthetic-fixture testing, and empirical validation with real exports. It audits source structure, adapter detection, canonical import, field coverage, native timestamp preservation, coordinate semantics, provenance, and canonical round-trip fidelity.
Details
validate_eye_source() validates one file or folder. Use
validation_manifest() and validate_eye_corpus() to evaluate a
versioned collection of exports from multiple vendors, devices, and software
versions. init_validation_corpus() is idempotent and preserves an existing
manifest and case files unless overwrite = TRUE is supplied. A source
can be retained temporarily in the result for creation of an anonymized
validation bundle; raw vendor data are never included in bundles by default.
anonymize_eye_dataset() replaces participant, recording, and session
identifiers, removes raw tables and source paths by default, and records the
operation in provenance. Automated anonymization cannot guarantee removal of
all study-specific free text, so exported bundles still require human review.