Skip to content

data, help: render structured dataset paths and slot formats - #3

Merged
karalabe merged 2 commits into
dark-bio:mainfrom
karalabe:dataset-paths
Sep 15, 2026
Merged

karalabe merged 2 commits into
dark-bio:mainfrom
karalabe:dataset-paths

Conversation

@karalabe

Copy link
Copy Markdown
Member

ark data paths printed the Ark's generated readme as raw markdown, and --json wrapped it in a single string an agent could not query. With darkbio-wire 0.10.0 the Ark replies with structured path entries instead, and each slot carries a format text for whoever fills it. This PR renders both.

  • ark data paths prints the path patterns as a compact tree. It marks directories, the ones a manifest may grant, and data this Ark lacks, the last only on the topmost entry of a missing subtree. --json carries every field, with desc named description as in the slot output. When anything is unavailable, a hint points at ark data list.
  • ark data show prints each slot's description and format, and ark data list --json carries both.
  • The help pages cover the new output. apps now explains the manifest, data grants and the sandbox, datasets gains a section on the data paths, and agents points at ark data paths --json for writing an app.
  • The CLI moves to darkbio-wire 0.10.0.

Scenario tests run data paths, data list and data show in child processes against invented entries, checking the exact JSON, the reading view within 80 columns, and the hints on stderr.

karalabe and others added 2 commits September 15, 2026 11:26
Co-authored-by: Astra <noreply@openai.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
@karalabe
karalabe merged commit 8433df8 into dark-bio:main Sep 15, 2026
7 of 8 checks passed
karalabe added a commit that referenced this pull request Sep 15, 2026
Develop images stamp their publish time when their build starts, not
when they are uploaded. The first develop image speaking the new
protocol reports 2026-09-15 08:33:55 UTC, four minutes before the cutoff
set in #3, so the CLI refuses the very image the cutoff was meant to
admit. This lowers the cutoff to that publish time.

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant