Module Map¶

bijux-proteomics-foundation is the contract floor of the repository. It gives every higher package the same vocabulary for identity, deterministic documents, compatibility, and explicit outcomes without importing proteomics workflows into that vocabulary.

flowchart LR
    I[identity] --> S[serialization]
    S --> C[compatibility]
    I --> P[support and provenance]
    S --> O[outcomes]
    C --> O
    T[testing policy] -. verifies .-> I
    T -. verifies .-> S
    T -. verifies .-> C

Public contract¶

The package root exposes a deliberately narrow set of stable primitives. Identifier types distinguish runs, samples, spectra, peptides, proteins, claims, reviews, and artifacts without reducing every key to an interchangeable string. DocumentSchema and JsonModel define document shape. Canonical JSON, SHA-256 helpers, and fingerprint functions make the same logical value produce the same bytes and digest.

The root API is lazy: importing the package resolves public names only when they are used. This keeps the dependency floor small while preserving one discoverable import surface.

Owned families¶

Family	Responsibility	Representative modules
`identity`	Typed, stable identifiers shared across package boundaries	`identifiers.py`
`serialization`	Canonical JSON, stable values, document schemas, fingerprints, and scientific-value encoding	`canonical_json.py`, `document_schema.py`, `fingerprints.py`
`compatibility`	Schema versions, migration plans, import migrations, and compatibility assessments	`schema_versions.py`, `schema_migrations.py`, `schema_assessments.py`
`outcomes`	Structured failures, refusals, results, exceptions, and optional-dependency errors	`failures.py`, `refusals.py`, `results.py`
`support`	Provenance records, lifecycle states, charter metadata, and API declaration support	`provenance.py`, `states.py`, `public_api.py`
`testing`	Reusable repository checks for public boundaries, generated files, skips, and source-tree limits	`public_function_type_boundaries.py`, `pytest_artifacts.py`, `source_tree_limits.py`

Boundary test¶

A concept belongs here only when several packages need exactly the same meaning and can use it without depending on a scientific workflow. Peptide scoring, experimental design, evidence review, and run execution remain with their owning packages. Foundation can represent their identifiers, documents, provenance, and failures; it does not decide their domain policy.