scaffold canonical-samples + ship v1 corpus (20 schemas, 340 rows) - #1
Merged
Conversation
The data backbone that folddb-dev new --from-prod will fetch sample rows from. Per-schema files at samples/<descriptive_name>.json; the generator fetches schema_service's /v1/snapshot, drafts plausible anonymized rows via Claude Haiku 4.5, validates each row's shape against field_types, and writes the file plus a manifest entry. CI runs validate.ts on every PR. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LLM-generated against schema_service-dev snapshot 2026-05-26T21:11Z via Claude Haiku 4.5. Curated set covers commonly-encountered canonical types (Message, Recipe, Movie, Book, Article, Review, NewsArticle, PostalAddress, ContactPoint, ImageObject, VideoObject, Episode, Game, SportsEvent, Course, Action, Brand, Dataset, GeoCoordinates, Comment). Every row passes tools/validate.ts shape validation against the schemas' declared field_types. Larger schemas (Person 69 fields, Place 48, CreativeWork 117) deferred to follow-up PRs — v1 picks tractable field counts (5-14) so generation stays cheap and reviewable. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
shiba4life
enabled auto-merge (squash)
May 26, 2026 21:21
The EdgeVector org default-branch-protection ruleset requires every PR to have a green status check whose context is exactly "ci-required". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Stands up the
EdgeVector/canonical-samplesrepo: layout, a Bun-basedgenerator + validator, CI, and the first 20-schema corpus of
LLM-generated sample rows.
This is the data backbone that
folddb-dev new --from-prod(separatefollow-up task) will fetch from.
What ships
Layout:
samples/<descriptive_name>.jsonper schema, plusmanifest.json,tools/for the generator/validator, and a GitHubActions workflow that runs shape validation on every PR.
tools/generate.ts— Bun script that fetches schema_service's/v1/snapshot, drives Claude Haiku 4.5 to author plausibleanonymized rows per schema, shape-validates each row, and writes the
file plus a manifest entry. CLI:
--schemas <names>or--all.tools/validate.ts— pure shape validator (reused by thegenerator and CI). Type-only — does not police semantic content.
v1 corpus — 20 schemas, 340 total rows, all passing strict
shape validation against the live snapshot:
What's out of scope
Person69 fields,Place48,CreativeWork117) —deferred to keep v1 cheap to generate and easy to review.
folddb-dev new --from-prodintegration — the next kanban task.Notes
EXEMEM_DEV_API_KEYsecret — until thatis wired, CI will pass with a warning rather than fail. Add the
secret in Settings → Secrets and variables → Actions to enable
full validation.
Test plan
bun tools/validate.ts samples/— all 20 files pass shape validationMessage.json,Recipe.json,PostalAddress.json— anonymized, plausible, variedEXEMEM_DEV_API_KEYsecret in repo settings so CI runs full validationCloses the v1 deliverable of the canonical-samples kanban task.
Co-Authored-By: Claude Opus 4.7 (1M context) noreply@anthropic.com