Open
Conversation
Add LerobotV3MetadataChecker with 7 pre-ingestion validation checks: 1. tasks.parquet exists (flags tasks.jsonl-only datasets) 2. Episodes parquet has required columns (chunk_index, file_index, tasks) 3. Feature shapes are non-empty ([] should be [1]) 4. File path templates use standard placeholders 5. Video files referenced by episodes exist 6. Timestamp consistency (all relative or all absolute) 7. Episode row contiguity in data parquet Integrated into the main validation orchestrator. Includes 17 new tests and updated existing test fixtures to include v3 metadata. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Null/missing start_timestamp values are now flagged as errors instead of being silently skipped. Downstream converters require a valid collection timestamp for every episode — catching this at validation time prevents runtime failures. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add two new validation checks to catch dataset issues that cause
runtime failures in the LeRobot MCAP converter:
8. Data parquet must not contain video struct columns — these cause
CastError when the dataset loader tries to match features schema.
9. Episode parquet must include videos/{key}/chunk_index and
videos/{key}/from_timestamp for each video feature — the dataset
loader needs these to resolve video files and timestamps.
Both issues were discovered during fpvlabs dataset ingestion and
required workarounds in the converter.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add 6 P0 validators as lerobot_validator/v3_checks.py to catch the most
common data quality issues before partner upload:
- V1 validate_tasks_format: error if no tasks file, warn if only jsonl
- V2 validate_codebase_version: require codebase_version starts with v3.
- V5 validate_feature_shapes: reject shape=[], require 3-element image shapes
- V7 validate_timestamps: reject absolute Unix epoch in data parquets
- V11 validate_custom_metadata_csv: require episode_index/episode_id, reject
null/duplicate episode_ids
- V12 validate_start_timestamp: require plausible Unix epoch floats
Wire validate_v3_dataset() into the LerobotDatasetValidator orchestrator
so errors surface automatically, and add get_warnings() support. Update
existing test fixtures to include codebase_version so integration tests
pass with the new checks.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
3 tasks
…es, share CSV reads - Import REQUIRED_METADATA_COLUMNS from schemas.py instead of hardcoding - Add Issue.error() / Issue.warning() factory methods to reduce boilerplate - Share CSV DataFrame between V11 and V12 via _df_cache to avoid redundant reads - validate_timestamps checks only first episode for monotonicity (not all) - Remove unused `thorough` parameter from validate_v3_dataset - Make timestamp constants public (UNIX_EPOCH_THRESHOLD, UNIX_EPOCH_MAX) Co-Authored-By: Claude Opus 4.6 (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
Add
LerobotV3MetadataChecker— 7 pre-ingestion validation checks for LeRobot v3 datasets:data/chunk_index,data/file_index,tasks[]should be[1])data_pathandvideo_pathmust use standard placeholders ({episode_chunk},{episode_index},{video_key})episode_indexmust be contiguousMotivation
These checks catch data issues at sharing time rather than at conversion time, preventing runtime fixups in downstream consumers. Partners can validate their datasets before uploading.
Test plan
🤖 Generated with Claude Code