Upload And Snapshots

Stratum ingests uncompressed STDF into a server-side analytical dataset. Ingest and loading that dataset into an Analyze browser tab are separate operations.

Upload and review

1. Open Ingest and select or drop your STDF file. Decompress archives first.
2. Review the file’s inspected lot identity and metadata. Check the lot ID carefully, especially if that lot already exists in your account.
3. Review the parser profile and any optional tags or metadata before starting.
4. Choose Start Ingest, or Start Ingests for multiple files.
5. Follow the ingest job to completion and review its warnings before opening the lot in Analyze.

The initial inspection is a preview, not validation of every record in the file. A successful preview does not guarantee that the full ingest will succeed or that it will complete without warnings.

Uploads do not merge partial lots

Each ingest produces a separate snapshot from that file. Uploading a second file for the same lot ID adds another snapshot; it does not append that file’s wafers or measurements to the earlier one. Selecting several files does not combine their partial contents. Use a complete lot input for the snapshot you intend to analyze.

Upload through the API

To automate uploads, send the STDF file to POST /api/upload as a multipart file field with bearer authentication. The response includes a job_id; use GET /api/ingest-jobs/{id} to follow processing. A successful upload response does not mean ingest has finished. The same snapshot rules apply to API uploads.

See API Surface for authentication and key scopes, and API Reference for upload options and response fields.

Snapshot, analysis, and version

Item	What it preserves
Lot snapshot	The analytical dataset produced by an ingest of that lot.
Analysis	Your investigation of a lot, including its selected snapshot and view state.
Analysis version	A saved checkpoint of that investigation, including its selection and what-if limits.

An analysis save does not create a new ingest snapshot. A new ingest does not silently move an older saved analysis onto different data.

The lot defaults to its newest successfully ingested upload. Earlier snapshots stay selectable; their data is not merged into or replaced by the new snapshot. Reprocessing the same source file marks only that source’s previous snapshot as superseded, not snapshots from other files. Source identity follows the file’s contents, not just its filename or lot ID.

Choose or change the snapshot

Start a New analysis to choose a lot and snapshot freely. Use the latest snapshot for current work; choose an earlier one when checking a historical finding.

A saved analysis stays linked to its lot and pins its saved snapshot. To move editable saved work to the lot’s current data, use Rebase latest, review the changed results, and Save a new analysis version. Rebasing is deliberate; earlier saved versions still refer to their earlier snapshot. To investigate another lot, start a new analysis instead.

See Saved Work And Sharing for version and recovery behavior.

Readiness and warnings

Complete on an ingest job means server-side ingestion finished. Analyze still has to load its datasets. Its separate progress status ends at Lot data ready; test summaries may be available before exact die identity and other context. See Analyze Workflow.

Inspect warnings and the resulting population before drawing conclusions. Examples include high rates of not executed test records and unit or res_scal drift across lots of the same product. Warnings do not always block ingest, but an unexpected test population or inconsistent units can invalidate a comparison.

Once the lot is ready, follow the Quickstart.