agentic-proteins 0.3.0

Deterministic, artifact-first protein design runtime and CLI

agentic-proteins

agentic-proteins is the deterministic runtime package in the proteomics workspace. It owns execution flow, runtime artifacts, and replay-oriented guardrails for protein design workflows.

If you need execution behavior, runtime state transitions, or inspectable run artifacts, start with this package.

What this package owns

• deterministic design-loop execution and runtime lifecycle state
• run artifacts, report generation, and replay-safe execution records
• package-local CLI and API entry boundaries
• provider contracts and runtime capability checks

What this package does not own

• protein program-stage governance and review gate domain contracts
• candidate ranking policy and scenario recommendation behavior
• evidence trust, contradiction policies, and claim-lineage semantics
• lab planning, scheduling, and rerun strategy outputs

Source map

• src/agentic_proteins/runtime for runtime execution control and lifecycle behavior
• src/agentic_proteins/design_loop for design-loop orchestration
• src/agentic_proteins/interfaces for CLI surfaces
• src/agentic_proteins/api for HTTP-facing boundaries
• tests for executable package expectations

Read this next

• Architecture
• Boundaries
• Contracts
• PyPI maintainer notes

Primary entrypoint

• console script: agentic-proteins

Changelog

All notable changes for agentic-proteins are recorded here.

0.3.0 - 2026-04-06

Added

• Package-local release manifest and maintainer-facing package docs: README.md, docs/ARCHITECTURE.md, docs/BOUNDARIES.md, docs/CONTRACTS.md, and docs/maintainer/pypi.md.
• Package-local changelog publishing path wired in package and root metadata.

Changed

• Package URLs now consistently reference bijux.io/bijux-proteomics and github.com/bijux/bijux-proteomics.

Fixed

• Test path resolution now uses explicit monorepo-root detection so e2e, regression, and governance tests stay stable with nested package manifests.

0.2.3 - 2026-01-16

Added

• Expanded provider test coverage for ColabFold, OpenProtein, and local ESMFold utilities.
• Runtime capability validation tests and candidate filter unit coverage.
• Stability marking test for module annotations.

Changed

• Hardened local ESMFold utility tests to exercise error and success branches.

Fixed

• Reliability checks and helper tests to keep coverage and gating stable.

0.2.2 - 2026-01-16

Added

• Release alignment for docs, gates, and CI structure.

Changed

• Consistent documentation build and validation wiring.

Fixed

• Minor release hygiene issues discovered in CI.

0.2.1 - 2026-01-16

Added

• Expanded unit and integration coverage with new invariants, API, and docs gates.
• Additional tests for provider isolation, reproducibility, and abuse-case blocking.
• Fancy PyPI readme fragments for README + changelog publishing.

Changed

• Refactored tests/unit into a structured layout for clearer ownership.

Fixed

• Coverage floors and CI gates stabilized around new test layout.

0.2.0 - 2026-01-16

Added

• Architecture invariants, threat model skeleton, and design debt ledger.
• Reproducible runs via agentic-proteins reproduce <run_id> with hash checks.
• Determinism tests, artifact immutability tests, and invariant regression coverage.
• Provider isolation checks and chaos failure test for mid-run provider loss.
• Benchmark regression gate and per-module coverage floors in CI.
• Documentation system contracts, lint gates, and CLI surface audit coverage.
• API error taxonomy enforcement, correlation ID logging test, and OpenAPI drift guard.
• Dependency allowlist enforcement for SBOM changes.

0.1.0 - 2026-01-14

Added

• Deterministic, artifact-first execution engine with explicit run directories and state snapshots.
• Agent-based architecture covering planning, analysis, execution, verification, and reporting.
• End-to-end design loop with failure handling, stagnation detection, and human-in-the-loop gating.
• CLI for running, resuming, inspecting, comparing, and exporting protein design runs.
• Local and remote provider abstractions with explicit capability and requirement checks.
• Structured reporting system with machine-readable artifacts and human-readable summaries.
• Integrated evaluation pipeline supporting structure-based metrics and ground-truth comparison.
• Reproducibility controls, observability hooks, and execution telemetry.
• Example datasets and reference runs for local experimentation and validation.
• Comprehensive test suite covering unit, integration, regression, and execution boundaries.