Fix repo-wide component_overview.org drift
Table of Contents
This page is a capture in the inbox bucket of the product backlog — a pre-sprint idea, not yet pulled into a sprint as a story.
What
Run ./projects/ores.codegen/validate_docs.sh and work through every
reported component: add #+type: component to modeling/component_overview.org
where missing, fill in the required sections (Inputs, Outputs, Entry points,
Dependencies, See also), and add a modeling/*.puml class diagram where
absent. ores.marketdata and ores.synthetic are missing
component_overview.org entirely and need one authored from scratch (see
doc-add-component-model skill). Affected components at time of writing:
ores.analytics.quant, ores.cli, ores.codegen, ores.compass, ores.compute,
ores.connections, ores.controller, ores.database, ores.diff, ores.dq,
ores.eventing, ores.fpml, ores.geo, ores.history, ores.iam, ores.lisp,
ores.logging, ores.nats, ores.orgmode, ores.platform, ores.refdata,
ores.reporting, ores.scheduler, ores.security, ores.seeder, ores.service,
ores.shell, ores.sql, ores.storage, ores.testing, ores.trading,
ores.utility, ores.workflow, ores.workspace, ores.wt.service.
Why
This drift accumulated silently — no CI gate runs validate_docs.sh, so
components missing their frontmatter type, standard sections, or class
diagram go unnoticed until someone runs the validator by hand (as
happened while wiring day_count_fraction_type into ores.refdata). A
dedicated pass would bring every component doc up to the same standard
and let validate_docs.sh be added as a CI check with a clean baseline.
References
./projects/ores.codegen/validate_docs.sh— the validator to run and re-run.
See also
- doc-add-component-model — skill for scaffolding a component_overview.org.