Output Formats¶
DOCTA produces two JSON report files.
Delta Report (delta_report.json)¶
Generated by diff compare or diff full. Contains categorized document changes based on file hash comparison.
| Field | Description |
|---|---|
unchanged |
Files with identical content hashes |
modified |
Files at the same path with different content |
renamed_candidates |
Potential renames detected via similarity matching (configurable threshold) |
added |
New files present only in the newer version |
removed |
Files no longer present in the newer version |
Semantic Diff Report (semantic_diff_report.json)¶
Generated by diff scan or diff full. Contains detailed block-level changes for each modified or renamed document.
Each document entry includes:
| Field | Description |
|---|---|
| Section additions/removals/modifications | Changes at the heading/section level |
| Text content changes | Diff of paragraph and text blocks with similarity scores |
| Code block changes | Added, removed, or modified code blocks |
| Table changes | Structure and data changes within tables |
| List modifications | Changes to ordered and unordered lists |
| Metadata changes | Document-level metadata differences |
Changes are reported semantically (e.g., "Installation section: added 3 paragraphs, modified 1 code block") rather than as raw HTML diffs. This means cosmetic changes (CSS class names, whitespace, tag attributes) are filtered out.