Frank14f/CelerisLab
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

test: versioned reader for DFG Re20 result.json (schema 1 and 2)

Browse Source 
read_result() returns cases/provenance for the old bare-list receipts and
the new schema-2 object, and rejects unknown schemas. Existing evidence
(e.g. 70f6617) and its digest scripts stay valid for schema 1; queue and
digest scripts must use read_result(...)['cases'] for new receipts.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Claude
2026-10-20 18:29:52 +00:00
co-authored by Claude Opus 5.5
parent 2220cdd157
commit c33af7a66d
3 changed files with 35 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/validation/dfg_re20.md
+3 -1
View File
@@ -48,5 +48,7 @@ The uniform D80 control separates method error from refinement error. `--plan` w
(per case: lattice parameters, every sample, window ranges, status, verdict).
Receipts: `plan.json` and `result.json` carry `schema` 2 and `provenance` (HEAD, src tree,
runner hash; clean checkout required, refused if it changes during the run). Receipts
written before this (schema 1, e.g. evidence 70f6617) are a bare list of cases.
written before this (schema 1, e.g. evidence 70f6617) are a bare list of cases. Read both
with `run_dfg_re20.read_result(path)` (`["cases"][0]` replaces the old `[0]`); existing
evidence scripts stay valid for schema-1 receipts and need this reader for new ones.
Host contract tests: `tests/unit/test_dfg_re20.py`.