find

Fuzzy component discovery over the shipped manifest — ``python -m chirp_ui find``. Sprint 6 of the agent-grounding-depth epic. The manifest is the registry projection; this module turns it into a one-liner for humans and agents who know *what* they want but not the exact BEM block name. Search model ------------ A component matches a query when the query is a case-insensitive substring of any of: * ``name`` — the manifest key, e.g. ``metric-card`` * ``block`` — the BEM block, e.g. ``metric-card`` (usually same as name) * ``category`` — e.g. ``data-display``, ``feedback`` * ``description`` — the ``{#- chirp-ui: ... -#}`` doc-block text Query filters (``--category`` etc.) are applied AFTER substring match, so ``find --category=feedback`` lists every component in that category even when no term is given. Use ``--authoring=preferred`` to list the registry-blessed primitives agents should reach for first, or ``--authoring=compatibility`` to audit legacy helpers retained for existing templates. Output ------ One line per hit, stable/sorted:: {name:<30} {category:<16} {first-line-of-description} Exit code ``0`` always (including zero hits) — this is discovery, not validation. See ``docs/plans/PLAN-agent-grounding-depth.md § Sprint 6``.

module
Fuzzy component discovery over the shipped manifest — ``python -m chirp_ui find``. Sprint 6 of the agent-grounding-depth epic. The manifest is the registry projection; this module turns it into a one-liner for humans and agents who know *what* they want but not the exact BEM block name. Search model ------------ A component matches a query when the query is a case-insensitive substring of any of: * ``name`` — the manifest key, e.g. ``metric-card`` * ``block`` — the BEM block, e.g. ``metric-card`` (usually same as name) * ``category`` — e.g. ``data-display``, ``feedback`` * ``description`` — the ``{#- chirp-ui: ... -#}`` doc-block text Query filters (``--category`` etc.) are applied AFTER substring match, so ``find --category=feedback`` lists every component in that category even when no term is given. Use ``--authoring=preferred`` to list the registry-blessed primitives agents should reach for first, or ``--authoring=compatibility`` to audit legacy helpers retained for existing templates. Output ------ One line per hit, stable/sorted:: {name:<30} {category:<16} {first-line-of-description} Exit code ``0`` always (including zero hits) — this is discovery, not validation. See ``docs/plans/PLAN-agent-grounding-depth.md § Sprint 6``.
components_by_authoring
function

Return manifest component entries with the given authoring hint. This is the Python API twin of ``python -m chirp_ui find --authoring=...``. It returns a name-keyed dict in sorted order so downstream agents/tools can inspect full manifest entries without hand-filtering JSON.

preferred_components
function

Return components marked as preferred authoring vocabulary.

compatibility_components
function

Return compatibility components retained for existing or narrow uses.

search
function

Return ``(name, category, summary)`` rows sorted by name.

format_rows
function

Format rows as an aligned three-column table (no header).

main
function