article Table

Overview

Item	Details
Table name	article
Description	Table for managing the "10-second race result summary" articles tied to races (auto racing, keirin, horse racing). Handles both AI-generated articles (created with status: generating, which becomes draft once built) and manually created articles
Primary use	Creating, updating, publishing articles and tracking generation failures

Column Definitions

Column	Type	PK	FK	Unique	Nullable	Default	Notes
_id	ObjectId	◯		◯			Auto-generated by DocumentDB
race_id	string		auto_racing:_id / bicycle_racing:_id / horse_racing:_id				References the _id of one of the racing tables. Indexed
title	string
body_html	string						HTML. For AI-generated articles, produced by the article_builder Lambda rendering the RaceSummary React component with renderToStaticMarkup. A self-contained fragment with its styles inlined
status	enum					draft	generating / draft / published / failed
opening_at	number						Scheduled publish date/time. Indexed
closing_at	number				◯		Scheduled unpublish date/time
accuracy_test_data	object				◯		Accuracy-verification data, used by scripts/article_accuracy_test/audit.py
generation_error	string				◯		Failure reason when status=failed (max 2000 chars). Shown in the admin UI
custom_prompt	string				◯		Extra instructions supplied by an operator on regeneration. Carried over to later regenerations
created_at	number
updated_at	number						Lists are sorted by update date/time by default
deleted_at	number				◯

No category_id

article has no category. Only blog / old_blog carry categories.

Status transitions

stateDiagram-v2
    [*] --> generating: POST /v1/articles/generation
    [*] --> draft: POST /v1/articles (manual)
    generating --> draft: article_builder stores body_html
    generating --> failed: generation failed on the final attempt
    failed --> generating: POST /v1/articles/{id}/regeneration
    draft --> published: operator publishes
    published --> draft: operator unpublishes

generating and failed are set by the generation pipeline; they cannot be supplied as input to POST /v1/articles or PUT /v1/articles/{id} (doing so returns 400).

Shape of accuracy_test_data

{
  "generated_article": {
    "title": "string",
    "contents": ["string", "string", "string"]
  },
  "race_result_in_generated_article": {},
  "race_analysis": "string"
}

Key	Contents
generated_article	The generated article (headline + each summary line)
race_result_in_generated_article	The numeric facts (finishing order, payouts, …) the article used, declared by the generating AI in the same shape as race_result
race_analysis	Output of stage 1 (race video / course image analysis). Kept so the mix of race-result vs. video provenance can be measured afterwards