Skip to contents

Renders a human-readable review interface from a rectangular dataset.

Usage

betwixt_render(
  claim,
  range = NULL,
  evidence_media_file = NULL,
  evidence_url = NULL,
  descriptive = NULL,
  assertions = NULL,
  context = NULL,
  con = NULL,
  title = "Betwixt Review",
  description = "Please review the following claims.",
  betwixt_id = "P1234",
  template = "long_review"
)

Arguments

claim

A data frame containing the material to render.

range

Optional data frame defining proposed values for reviewable assertions.

evidence_media_file

Optional character string naming the column containing the media resource presented as evidence.

evidence_url

Optional character string naming the column containing a URL to supporting evidence.

descriptive

Optional character vector naming columns presented as descriptive metadata.

assertions

Optional character vector naming columns presented as reviewable assertions.

context

Optional character vector naming columns presented as contextual information.

con

Optional output connection or file path. If NULL, return the rendered HTML as a character string.

title

Character string giving the review title.

description

Character string describing the review task.

betwixt_id

Character string identifying the review artefact.

template

Character string identifying the review template. Must be "long_review" or "wide_review".

Value

If con is NULL, a character string containing the rendered HTML. Otherwise, con is returned invisibly.

Details

A review interface may contain four blocks:

  • evidence, containing media or a source URL inspected by the reviewer;

  • descriptive metadata, containing information that may be displayed or edited without constituting an epistemic review decision;

  • semantic assertions, containing the information under review; and

  • context, containing additional information that supports the review.

The columns belonging to these blocks are specified explicitly. Their roles are not inferred from column names.

Long and wide reviews are alternative human-readable projections of semantic assertions. They do not define different semantic models.

In a long review, assertions identifies the assertion components exposed as columns. For example, c("subject", "predicate", "value") exposes a complete elementary assertion, while c("predicate", "value") may be used when the subject is implicit in the evidence under review.

A long review may also expose multiple proposed values, for example c("predicate", "value1", "value2").

In a wide review, each column named by assertions represents a predicate and its cells contain the corresponding reviewable values.

Sequential row numbers are generated by the renderer. They are part of the review interface and do not need to be present in claim.