How Evidence Becomes Outputs¶

This section explains how source material becomes visible output. It belongs on the public surface because practical questions quickly turn into architecture questions:

how did this report, map point, or bundle get here
which part of the system governs it
where should I look when a visible output seems stronger or weaker than I expected

The goal here is not to tour internal names. It is to make the publication flow understandable enough that you can trace one public output back to the part of the runtime that governs it.

Flow¶

flowchart LR
    commands["command surface"]
    collection["source collection"]
    normalization["evidence normalization"]
    review["evidence review"]
    assembly["publication assembly"]
    writing["public artifact writing"]
    checks["unit and regression checks"]

    commands --> collection
    collection --> normalization
    normalization --> review
    normalization --> assembly
    review --> assembly
    assembly --> writing
    checks --> commands
    checks --> review
    checks --> writing

The Main Stages¶

commands declare what kind of rebuild, check, or inspection is being requested
collection brings governed source material into the repository
normalization turns mixed upstream inputs into comparable repository-owned evidence files
review surfaces expose strengths, blockers, caveats, and refusal reasons
publication writes country, regional, and world-facing outputs
checks fail when those layers drift apart or start implying too much

What This Section Helps Explain¶

why a visible output is never the whole story by itself
why source intake, evidence normalization, review, and publication must stay visibly separate
where to go next when your question is about system flow rather than evidence content

Durable Boundaries¶

command_line/ owns CLI parsing, dispatch, and command registration
data_downloader/ owns source-family collection, intake helpers, and tracked context normalization
adna/ owns animal aDNA intake, extraction, normalization, and validation
analysis/review/ owns ranking review surfaces rather than public rendering
reporting/ owns publication assembly, rendering, and governed report writing
foundation/ owns repository-truth, release posture, and architecture contracts

Use This Section When You Need To Know¶

how commands line up with tracked source material
where evidence becomes reviewable before it becomes public output
which parts of the repository own review versus rendering
where to look if an output changes unexpectedly
how to trace a public-facing surface back to its governing evidence and rules

Expanded Pages¶

runtime system model explains the end-to-end flow in the order it actually runs
module map explains which directories own which parts of the lifecycle
package split explains why runtime, maintainer, and alias distributions stay separate