docs: add Claude Code skill for the LitData codebase#847
Merged
Conversation
Adds a `.claude/skills/litdata` skill that helps contributors and agents work on the codebase: understand the architecture, contribute, write/run tests, and debug/profile. A lean SKILL.md acts as a router into on-demand reference files so it loads fast: - reference/streaming.md — read pipeline (dataset/dataloader/reader, chunk format, shuffling, resume, item loaders) - reference/cache-and-chunk-lifecycle.md — Cache <-> Writer/Reader, PrepareChunksThread, distributed sampling, shared-chunk deletion & refcounting - reference/processing.md — write pipeline (optimize/map), worker model, raw indexer - reference/contributing.md — dev setup, style, PR flow, lint/type/CI gates - reference/testing.md — pytest commands, fixtures, mocking cloud, gating - reference/debugging.md — tracing (enable_tracer/Litracer), env knobs, failure modes Docs only; no source or runtime changes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
for more information, see https://pre-commit.ci
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #847 +/- ##
===================================
- Coverage 81% 81% -0%
===================================
Files 54 54
Lines 7683 7683
===================================
- Hits 6210 6209 -1
- Misses 1473 1474 +1 🚀 New features to boost your workflow:
|
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Adds a Claude Code skill for the LitData codebase under
.claude/skills/litdata, to help contributors (and coding agents) get up to speed and work effectively: understand the architecture, contribute, write/run tests, and debug/profile.Docs only — no source or runtime changes.
Layout
A lean
SKILL.mdacts as a router (a task → file table + public-API map + quick commands) so only a small file loads by default; the depth lives inreference/files read on demand:SKILL.mdreference/streaming.mdStreamingDataset/StreamingDataLoader/BinaryReader, chunk format, shuffling, resume, item loadersreference/cache-and-chunk-lifecycle.mdCache↔ Writer/Reader,PrepareChunksThread, distributed sampling, and the shared-chunk deletion / reference-counting machineryreference/processing.mdoptimize/map,DataProcessorworker model, raw indexerreference/contributing.mdreference/testing.mdreference/debugging.mdenable_tracer/ Litracer, env-var knobs, common failure modes + triageAll line/file references were derived by reading the current tree; the content is grounded in
src/litdata/**,tests/**,pyproject.toml,.pre-commit-config.yaml, and the CI workflows.🤖 Generated with Claude Code