Reads a standalone Betwixt review HTML file and reconstructs the original candidate state, the persisted reviewed state, review metadata, and review provenance.
Value
A named list with four elements:
- metadata
Review-level metadata including
title,description,project_id,sequence, andstatus.- provenance
Review provenance including the original and saved filenames, creation and lifecycle timestamps, reviewer name and email, and the optional review-level comment.
- candidate
A data frame reconstructing the candidate state presented for review. It contains row number, evidence fields, descriptive fields, and candidate semantic assertion values.
- reviewed
A data frame reconstructing the persisted reviewed state. It contains the current descriptive and semantic values, assertion qualifications, row finalisation state, row outcome, and optional row-level comments.
Empty optional evidence fields and comments are returned as NA_character_.
row_number is returned as integer and reviewed$finalised as logical.
Details
read_review() does not finalise semantic knowledge or construct a
subsequent stabilised semantic state. It reconstructs the review artefact
as saved by Betwixt.
The returned candidate and reviewed data frames preserve the same
evidence, descriptive, and semantic fields. The reviewed data additionally
contains assertion qualifications, row finalisation state, row outcome, and
optional row-level comments.
Review-wide information such as reviewer identity, lifecycle timestamps, and
the review-level comment is returned separately in provenance.
Betwixt review files preserve both original candidate values and the current
reviewed values in machine-readable HTML attributes and controls.
read_review() uses these persisted values to reconstruct the two states
independently.
A finalised Betwixt review is a completed review artefact. Reading such a file does not promote reviewed assertions into a subsequent stabilised semantic state or apply them to an external knowledge system.