Public Interfaces¶
This section describes the runtime contracts that readers are expected to use directly.
It has to show more than a list of commands. The runtime exposes several different entry surfaces because different readers need different depths of control and different artifact guarantees.
Interface Families¶
- CLI commands for end-to-end workflows and reviewer-facing files
- workflow Python surfaces under
bijux_phylogenetics.api - native Python contracts for runtime objects and inference results
- benchmark and evidence-adjacent runtime contracts
- emitted artifact and report families
What Each Interface Unlocks¶
| Interface family | Best for | What the reader gets back |
|---|---|---|
| CLI | governed command execution | files, reports, manifests, and reproducible output directories |
| Workflow Python | notebooks and pipelines | typed workflow results with stable writers |
| Native Python | lower-level scientific control | direct owned contracts for trees, likelihood, inference, and analysis |
| Artifact surfaces | review and downstream inspection | explicit schemas, output families, and consumption rules |
Interface Selection Heuristic¶
| If you need | Start here | Why |
|---|---|---|
| one governed workflow or report | CLI | easiest path to durable artifacts |
| notebook or pipeline composition | workflow Python | typed results and stable writers |
| lower-level runtime control | native Python | direct access to owned contracts |
| trust or schema review | artifact pages | explicit output families and boundaries |
flowchart LR
A[Need an interface] --> B[CLI]
A --> C[Workflow Python]
A --> D[Native Python]
A --> E[Artifact surfaces]
B --> F[Files and reports]
C --> G[Typed workflow results]
D --> H[Lower-level owned contracts]
E --> I[Schema and output review]
Why This Section Matters¶
The package looks smaller than it is when all interfaces are described as if they were one generic API. They are not:
- the CLI is optimized for governed runs and durable outputs
- the workflow API is optimized for typed orchestration in Python
- the native runtime surface is optimized for readers who need owned lower-level scientific contracts
- the artifact pages are optimized for downstream consumers and reviewers
Start Here¶
- If you need help choosing an entry point, use the surface selection guide.
- If you already know the entry point, continue to the matching reference page below.
Section Map¶
- Surface selection guide
- CLI surface
- API surface
- Python surface
- Native inference and benchmarks
- Artifact consumption guide
- Artifact contracts
The surface selection guide and artifact consumption guide are the two fastest ways to orient a new reader before they drop into the larger CLI and Python reference pages.