Stats¶

datasetlint stats computes dataset distributions in addition to pass/fail validation.

datasetlint stats examples/minimal_dataset
datasetlint stats examples/minimal_dataset --format json
datasetlint stats examples/minimal_dataset --format markdown

Python:

from datasetlint.stats import compute_dataset_stats

stats = compute_dataset_stats("examples/minimal_dataset")
print(stats.frame_counts)
print(stats.label_class_counts)

The returned DatasetStats model includes:

• dataset_path
• duration_sec
• sensors
• frame_counts
• inferred_rates_hz
• label_class_counts
• confidence_summary
• track_length_summary
• speed_summary
• acceleration_summary
• missing_frame_counts
• issue_summary

Stats are printed to stdout. Redirect JSON or Markdown output when you want an artifact:

datasetlint stats examples/minimal_dataset --format json > datasetlint-stats.json