Product Overview¶

This is the shortest repository-level explanation of one bounded proteomics product. It exists so the reader can understand the system shape, the current credible workflow family set, and the best next question before package ownership starts to dominate the route.

What matters now is that bijux-proteomics is no longer only a governance or release-shaping repository. The current product line has real scientific substance in the core package, reviewable execution in runtime, explicit scientific memory in knowledge, bounded recommendation posture in intelligence, and honest follow-up pressure in lab.

The important correction is that this product should now be read as a scientific workflow system with explicit review pressure, not as a docs-first governance shell around disconnected packages. The product promise is concrete: ingest benchmark-backed proteomics material, preserve reproducible shared meaning, execute a reviewable runtime path, ground the result scientifically, decide what recommendation posture is still justified, and only then ask whether downstream assay burden is worth accepting.

What Is Real Today¶

owner	what is materially real now	why it changes how readers should judge the repository
`bijux-proteomics-foundation`	shared document, identifier, compatibility, outcome, and canonical-serialization kernel	persisted scientific state can survive cross-package review and release change
`bijux-proteomics-core`	sequence parsing and digestion, chemistry and modifications, mzML and spectra ingestion, identification, quantification, PTM review, DIA support, benchmark assets, and workflow contracts	the repository now has real biological and analytical depth at its scientific center
`bijux-proteomics-runtime`	live execution, replay, import-backed reruns, raw-executable lanes, preflight checks, artifact integrity, and operator-facing rerun bundles	public workflow evidence can be reopened without maintainer storytelling
`bijux-proteomics-knowledge`	cited references, claim grounding, contradiction-aware evidence memory, and biological context such as pathways, complexes, kinases, drugs, disease, and orthologs	repository language can be challenged by explicit scientific memory instead of tone
`bijux-proteomics-intelligence`	ranking, falsifier-aware review, interpretation summaries, confidence, regret, challenge routes, and recommendation posture	analytical judgment is now inspectable rather than hidden inside summaries
`bijux-proteomics-lab`	assay planning, readiness, handoff honesty, refusal boundaries, and requested-versus-observed outcome reconciliation	downstream follow-up burden is visible enough to narrow the public story honestly

What The Repository Can Now Defend¶

benchmark-backed workflow-family language is no longer carried by release prose alone; it is carried by paired public benchmark packages, runtime rerun lanes, grounding routes, recommendation packets, and consequence ledgers
biology and chemistry depth no longer sit only in aspirational structure; the current core package now carries sequence, spectra, mzML, PTM, DIA, quantification, and benchmark-surface responsibility in one visible owner
the knowledge and intelligence layers now make disagreement, downgrade, and refusal legible enough that public language can be challenged instead of merely asserted

What Changed Since v0.3.7¶

the core package moved from "workflow contract owner" to a visibly broader scientific owner with deep biology, chemistry, spectra, and review surfaces
runtime moved from "can run things" to "can prove and bound reruns publicly"
knowledge and intelligence moved from abstract support roles to concrete grounding and policy owners
lab moved from a soft handoff idea to a visible consequence, refusal, and outcome owner

What A Serious Reader Can Verify Quickly¶

that the current workflow-family language is carried by open benchmark packages and not only by release prose
that runtime proof now extends beyond one launch path into rerun kits, replay challenges, and refusal surfaces
that knowledge, intelligence, and lab now narrow language explicitly instead of leaving disagreement and downstream burden implied
that the product story crosses biology, chemistry, execution, grounding, judgment, and consequence in one visible route

What This Product Still Refuses To Pretend¶

that broader module count alone creates broader scientific authority
that a raw-executable lane settles the grounding or recommendation question
that a grounded recommendation automatically deserves downstream assay spend
that one strong workflow family upgrades the whole repository

Start Here¶

Open Product Architecture when the question is how benchmark intake, runtime execution, scientific review, recommendation posture, and lab consequence fit together.
Open Workflow Families when the question is which family is credible today and what still blocks the rest.
Open Execution when the question is how an operator or outsider reruns the current flagship evidence.
Open Decision Support when the question is whether grounding, contradiction, recommendation, or public language is actually aligned.

Strongest Reader Proof¶

Open Flagship Release Candidate when the question is which workflow families currently survive public scrutiny and why that does not yet justify broader repository language.
Open Benchmark Assets when the question is whether the evidence roots are broad enough and honest enough to support the current workflow-family sentences.
Open Execution when the question is whether those benchmark assets can be rerun, imported, replayed, and inspected without private maintainer glue.

Why This Page Must Stay Product-First¶

package handbooks matter, but most first-time readers are trying to decide whether the repository has real scientific substance before they care which owner package carries it
the root story now needs to explain biological depth, runtime proof, grounding pressure, and assay consequence in one route or it will still understate what changed since v0.3.7
if this page ever sounds stronger than the claim-limit, readiness, and consequence pages, the repository-level wording is wrong

Current Credible Families¶

outsider-auditable today: dda, dia, ptm, targeted
review-grade-bounded today: lfq
internal-support-only today: multiplex
live repository ceiling: Release Readiness Matrix

Reader Routes¶

Public Proof Chain¶

benchmark assets prove which public evidence roots are actually shipped
execution proves which of those roots can be reopened, replayed, and bounded honestly
decision support proves whether grounded evidence, recommendation posture, and lab burden still agree
current capability limits prove where stronger language still stops

Category Routes¶

Benchmark Assets owns the route from public source identity to flagship benchmark packages, lineage pages, freshness review, and acceptance bars.
Execution owns run modes, rerun kits, replay challenges, environment contracts, and runtime refusal surfaces.
Workflow Families owns the family comparison surface and the shortest scientist route from one family to benchmark, runtime, knowledge, intelligence, and lab evidence.
Decision Support owns the route through knowledge grounding, literature audits, recommendation confidence, public artifact roles, and release-language limits.
Lab Consequence owns the route from recommendation posture to assay consequence, operator burden, and refusal limits.
Maintenance owns repository-health routes such as boundary checks, release gates, docs integrity, and final preflight.

Package Handbooks¶

The package handbooks still exist because package-local behavior needs a real owner. They should now be the second hop, not the first hop, for most readers.

Boundary¶

This page owns repository-wide routing. It should shrink confusion before the reader enters a package handbook, not duplicate package-local explanations once they get there.