Unstructured zip ingest
Time to complete: ~15–30 min (depends on zip size)
What you'll accomplish: Upload a labeled zip through Lake and land tabular results in bronze.
Before you start
- Zip organized per folder layout (folder names = class labels)
- Sidebar → Lake (
/data) - This is bulk ingest — for single-file APIs use Read/Vision instead
Steps
- Sidebar → Data Warehouse → Lake.
- Click Add data → choose Unstructured ("Labeled zip of images or documents").
- Upload your zip; name the job if prompted.
- Wait for completion — progress banner shows running jobs; click Resume job if you navigate away.
- Bronze tab → new table with
filenameandlabelcolumns (plus extracted features). - Optional: Lab profile → silver → gold publish → ML.
Legacy URL
/forge redirects to /data?ingest=unstructured — same flow.
What happens next
- Zip folder layout guide
- Lake overview
- Unstructured APIs for one-off text/image calls without Lake
Common questions
Can I skip gold and train directly from bronze?
No. Publish gold from Lake for ML or Forecast training. Bronze is raw ingest; gold passes validation checks.
Vision / Read APIs
For single-request text or image analysis without the lake, see the Unstructured suite.