Skip to content

Interfaces

bijux-proteomics-intelligence interfaces are how recommendation logic becomes usable to humans and tooling. This section should help a reader see which surfaces accept candidates and policy inputs, which ones emit rankings, briefings, and outcomes, and where explanation is part of the contract instead of optional decoration.

flowchart LR
    inputs["candidate and policy inputs"]
    scoring["scoring and evaluator surfaces"]
    briefs["briefs and report artifacts"]
    outcomes["recommendation outcomes"]
    review["review boards and downstream packages"]

    inputs --> scoring --> briefs --> outcomes --> review

What These Interfaces Need To Preserve

  • recommendation output must stay explainable enough for review, not just machine-readable enough for automation
  • scoring surfaces must reveal policy shape without pretending to be evidence truth
  • downstream consumers need stable outcome and report forms because real portfolio decisions depend on them

Start With

  • open Data Contracts when the question is what a score, shortlist, brief, or outcome is allowed to mean
  • open Artifact Contracts when the real concern is report payload shape and explanation output
  • open Operator Workflows when the reader wants to follow the decision surface in practice
  • open Compatibility Commitments before changing an explanation or outcome contract that review processes may depend on

Read By Consumer

What This Section Should Clarify

  • where recommendation interfaces end and governance interpretation begins
  • which outputs are part of the stable surface versus transient implementation detail
  • why explanation belongs in the interface contract for this package

First Proof Check

  • src/bijux_proteomics_intelligence/candidates.py, policies.py, and evaluators.py
  • src/bijux_proteomics_intelligence/report/, briefs.py, and outcomes.py
  • packages/bijux-proteomics-intelligence/tests