Offline report comparison example

This directory is a labelled document-processing fixture for termivar report compare. It is not a pair of real assessments and does not allege that a target changed. The two inputs were derived from the public, credential-free first-use document solely to exercise the four comparison groups with independently selected expected counts.

The example was generated with development package version 0.10.0-alpha.2. Exact source and binary identity, input/output hashes, and the generation command are recorded in provenance.json.

Expected groups:

Group	Count	Fixture meaning
only_in_after	1	A synthetic identity exists only in after.json
only_in_before	1	A different synthetic identity exists only in before.json
changed	1	One shared identity has a deliberately edited display summary
unchanged	1	One shared identity has equal comparable content

The generated comparison summary is therefore:

only_in_after:  1
only_in_before: 1
changed:        1
unchanged:      1

Run from the repository root:

termivar report compare \
  --before docs/examples/report-compare/before.json \
  --after docs/examples/report-compare/after.json \
  --same-scope

termivar report compare \
  --before docs/examples/report-compare/before.json \
  --after docs/examples/report-compare/after.json \
  --same-scope \
  --format json \
  --output comparison.json

termivar report compare \
  --before docs/examples/report-compare/before.json \
  --after docs/examples/report-compare/after.json \
  --same-scope \
  --format html \
  --output comparison.html

The checked-in output files are the actual CLI results for those fixture inputs. Use a new destination for another run: the command intentionally refuses to replace an existing output. The before/after SHA-256 values in the comparison identify supplied bytes; they do not authenticate a source or prove equivalent assessment coverage. In particular, only-in-before is not verified remediation.