How This Helps
Export datasets directly using
curl without needing Python or SDKs. Ideal for automation scripts or CLI-based workflows.Initiate Export
Use aGET request to trigger export:
<DATASET_ID>: Your dataset IDexport_format: Supported values arejsonorparquetinclude_images:trueorfalsefile_name: Name for the output ZIP
Sample Response
Poll Export Status
Use the exportid to monitor progress:
"status": "COMPLETED".
Download ZIP
When ready, download your dataset:-L flag to follow redirects (required for S3 links).
Unzip and Explore
Extract the ZIP archive:Full Automation Script
Filter Exported Data by Uniqueness
Usepandas to find the most unique images:
Copy Filtered Images
Summary
- Initiated export
- Polled export status
- Downloaded and unzipped archive
- Parsed metadata using
pandas - Used
uniqueness_scorefor image filtering