docs(sdk): render skill siblings in the API-docs generator#1903
Draft
anthonyiscoding wants to merge 21 commits into
Draft
docs(sdk): render skill siblings in the API-docs generator#1903anthonyiscoding wants to merge 21 commits into
anthonyiscoding wants to merge 21 commits into
Conversation
…nd build helpers before SDK type check Engine integration tests and benches still imported iii_sdk::UpdateOp, which moved to iii_helpers::stream. The Node CI type-check step now also builds @iii-dev/helpers so its subpath type declarations resolve.
…e clean break) Add 0.20.0 entries for the @iii-dev/helpers extraction, the cross-language parity renames/removals, and the engine/protocol/internal/utils submodule moves, and correct the 0.20.0 index summary to state that the release is breaking (only the first errors/channel/trigger/runtime slice keeps deprecated root aliases).
… retention Add the StreamChangeEventDetail, Stream*Result, and MergePath name alignment to the 0.20.0 parity entry, and correct the removals list: only FieldPath was removed; MergePath is retained as a named export across the three SDKs.
bump_manifests bumped the iii-helpers crate/package to the prerelease but left the dependency requirements pointing at the stable version, so cargo (^0.19.4 vs 0.19.4-alpha.1) and pip (==0.19.4) could not resolve. Bump the iii-helpers workspace-dep pin (root Cargo.toml) and the python iii ==pin, mirroring iii-observability.
…i-observability (#1874)
- iii.rs: use Error (IIIError alias dropped from module scope) - engine: import UpdateOp/UpdateOpError/MergePath from iii_helpers::stream - regenerate Cargo.lock
main bumped the workspace to 0.19.5-next.1 but the helpers packages (added on this branch) were still on their pre-rebase versions. Align @iii-dev/helpers (node) and iii-helpers (python) to 0.19.5-next.1 / 0.19.5.dev1 to match the rest of the workspace and regenerate the uv locks. (The rust iii-helpers crate was already bumped during conflict resolution.)
Co-authored-by: Guilherme de S. Vieira Beira <guilherme.vieira.beira@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
generate-api-docs.mts wrote the api-reference *.mdx but not their <source>.skill.md siblings, so every regeneration left them stale and tripped iii-skill-check verify-rendered. After writing each page, invoke the canonical iii-skill-render binary on just that file (scoped to the generated pages, not the whole docs root, to avoid touching unrelated out-of-scope pages). Output is byte-identical to the committed siblings. If iii-skill-render is not on PATH the step warns and continues so local generation still works without it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
Contributor
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Repository UI Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
guibeira
force-pushed
the
sdk-refactor
branch
2 times, most recently
from
880c8b4 to
fa54475
Compare
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
2 participants
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.
What
docs/next/scripts/generate-api-docs.mtswrote theapi-reference/*.mdxpages but not their<source>.skill.mdsiblings, so every regeneration left the siblings stale and trippediii-skill-check verify-rendered.This makes the generator render each page's skill sibling right after writing it, by invoking the canonical
iii-skill-renderbinary on just that file (scoped to the generated pages, never the whole docs root, so unrelated out-of-scope pages aren't touched). Output is byte-identical to the committed siblings, so there is no drift.If
iii-skill-renderisn't onPATH, the step warns and continues so local generation still works without it.Why in the script (not the workflow)
Keeping it in the generator means any regeneration (local or CI) produces matching siblings, with no separate render step to forget. No workflow files are touched.
Follow-ups for the CI autoregen path (intentionally not done here)
For the scheduled
generate-api-docs.ymlpush-to-main job to render and commit siblings, two workflow tweaks are still needed (flagging, not changing):iii-skill-renderavailable in that job (it only sets up node/python/rust today; without the binary the script hits the warn-and-skip path).git add docs/next/api-reference/*.mdxto also stage*.skill.md.Test
verify-renderedreports noapi-referencedrift, and the working tree is clean after generation (output byte-identical to what's committed onsdk-refactor).🤖 Generated with Claude Code