Operations

bijux-proteomics-knowledge operations is about maintaining trustworthy evidence state. The package does not earn trust by always producing one neat answer. It earns trust by preserving provenance, keeping contradictions visible, and making resolution changes reviewable over time.

flowchart LR
    source["new source or resolution change"]
    lineage["check lineage and graph integrity"]
    confidence["check confidence and contradiction behavior"]
    review["rebuild review outputs"]
    compatibility["check schema and serialization continuity"]
    release["publish updated knowledge behavior"]

    source --> lineage --> confidence --> review --> compatibility --> release

What Operations Means Here

• the operational problem is epistemic drift as much as software drift
• release confidence depends on preserving reviewability of evidence state
• changes to trust scoring or contradiction resolution deserve the same rigor as interface changes

Start With

• open Common Workflows when you need the standard route from evidence change to released package
• open Observability and Diagnostics when contradiction behavior, confidence outputs, or review packets look wrong
• open Failure Recovery when a knowledge state must be repaired without erasing lineage
• open Release and Versioning before publishing changes that alter how evidence is scored or resolved

Route From Failure Mode

• Local Development and Installation and Setup for reproducible knowledge-state work
• Deployment Boundaries and Security and Safety for the limits around sensitive evidence and durable records
• Performance and Scaling when evidence volume, graph size, or review packet generation becomes the real operator pain

First Proof Check

• src/bijux_proteomics_knowledge/claims.py and evidence.py
• src/bijux_proteomics_knowledge/confidence/segments.py and review.py
• packages/bijux-proteomics-knowledge/tests