[DamianReeves]

Summary

Reads Morphir IR formats 1 and 2 and upgrades them to format 3.0.0. Before this change, both were refused with unsupported_format_version_major. Bead morphir-ukvn.8.12. The OpenSpec change add-ir-formats-1-2 is archived in this branch, and the IR specs are updated.

• Format 1: morphir.ir.v3.codec_v1 ports the decoders of morphir-elm */CodecV1.elm at bc99af69. Tags are snake_case, access control is ["public", x], documentation is [doc, x] or a bare x, record fields are [name, type], package modules are {"name", "spec"} and {"name", "def"}, and modules have no doc.
• Format 2: morphir.ir.v3.codec_v2. The morphir-elm history (8e0e729d, e742a620, 02aa36af, 7cb7b300) shows that format 2 is format 3.0.0 with snake_case value and pattern tags (Value/CodecV2.elm at 02aa36af). morphir-elm HEAD refuses format 2, so reading it here is a deliberate divergence. The finos v2 schema docs say that format 2 literals are snake_case, but morphir-elm wrote them in TitleCase, and morphir-elm wins.
• Both formats decode into the existing format 3.0.0 model, as elm does. The upgrade is to encode the result with morphir.ir.v3.codec. The legacy modules have no encoders.
• Format version: LegacyFormat, recognize_legacy_format, read_document_with_legacy and read_legacy_document. These recognize only the integers 1 and 2. The support table and the diagnostic codes do not change. read_format_version still refuses 1 and 2, with a message that names the legacy readers. write_format_version refuses majors 1 and 2.
• CLI: ir inspect reports format: 1 (legacy). ir convert upgrades a legacy file to format 3.0.0 or to draft v4.
• Fixtures: the morphir-elm rentals example in format 1 and in format 2, copied byte for byte into tests/fixtures/ir/legacy/ with provenance. Both files decode to the same Library.
• The 3.0.0 codec, the v4 package and migrate.py do not change.
• Rebased on feat(ir): realign the v4 model with upstream and drop per-input attributes on request #35 and feat(ir): read and write IR 3.1.0 with a v3 Specs distribution #34. The support table is now [3.0.0,3.2.0),[4.0.0,4.1.0), and the v3 distribution is Library | Specs. The legacy formats always decode to a format 3.0.0 Library and never to a Specs. The spec deltas were rewritten against the merged specs, and the change was archived again.

Test plan

• mise run check: lint, format, mypy, pyright and 2058 tests pass, with 98.5% coverage. The new modules have 100% coverage.
• mise run test-bdd: 29 scenarios pass, including the new legacy upgrade scenarios.
• openspec validate --specs --strict passes.
• Per-tag tests with hand-written JSON cover every format 1 and format 2 type, value, pattern, literal and wrapper tag. The spellings come from the elm source.
• The format 1 and format 2 rentals fixtures decode to equal Libraries, and ir convert --to v3 writes identical files for both.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-27T07:29:05.086645Z	515e97d	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[DamianReeves]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Keep it up!

Reviewed commit: 515e97d531

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".