Skip to main content
This reference documents the lifecycle of a sequence in Avala, including every possible status and the transitions between them.

Overview

A sequence in Avala represents a unit of data (a set of frames, a point cloud scan, or a group of images) that moves through a structured annotation workflow. Each sequence has a status that reflects its current position in the pipeline. The status system supports both 4D (3D point cloud + time) and 2D (image/video) annotation stages, as well as review and approval workflows.

Status Reference

State Transition Diagram

Status Details

unattempted

The initial status for all newly ingested sequences. Data has been uploaded and processed, but no annotation work has been planned or started.
  • Triggered by: Data ingestion or upload
  • Next status: pending (when added to a project or work batch)

pending

The sequence is recognized by the project and queued for assignment. It is waiting to be placed in a work batch or assigned to an annotator.
  • Triggered by: Adding the sequence to a project
  • Next status: ready_for_annotation (when assigned to a batch or annotator)

ready_for_annotation

The sequence is assigned and available for an annotator to pick up. It appears in the annotator’s work queue.
  • Triggered by: Batch assignment or direct assignment
  • Next status: labeling_4d (for datasets with 3D data) or labeling_2d (for 2D-only datasets)

labeling_4d

An annotator is actively creating 3D/4D annotations (e.g., 3D cuboids in point clouds). The sequence is locked to the assigned annotator.
  • Triggered by: Annotator opens the sequence and begins 3D annotation
  • Next status: review_4d (when the annotator submits the work)

review_4d

The 3D/4D annotations have been submitted and are awaiting review by a designated reviewer.
  • Triggered by: Annotator submits 3D annotations
  • Next status: ready_for_2d (if approved) or labeling_4d (if rejected and sent back for rework)

ready_for_2d

The 4D annotation stage is complete and approved. The sequence is now ready for 2D annotation work (e.g., annotating camera images using projected 3D cuboids as reference).
  • Triggered by: 4D review approval
  • Next status: labeling_2d (when an annotator begins 2D work)

labeling_2d

An annotator is actively creating 2D annotations (e.g., bounding boxes, polygons, segmentation masks on camera images). The sequence is locked to the assigned annotator.
  • Triggered by: Annotator opens the sequence and begins 2D annotation
  • Next status: review_2d (when the annotator submits the work)

review_2d

The 2D annotations have been submitted and are awaiting review.
  • Triggered by: Annotator submits 2D annotations
  • Next status: final_review (if approved) or labeling_2d (if rejected and sent back for rework)

final_review

All annotation stages (4D and/or 2D) are complete. A senior reviewer or project manager performs a final quality check on the full set of annotations.
  • Triggered by: All preceding review stages are approved
  • Next status: completed (if approved) or rework_required (if issues are found)

rework_required

The final reviewer identified quality issues that need correction. The sequence is routed back to the appropriate annotation stage for fixes.
  • Triggered by: Final reviewer rejects the sequence with feedback
  • Next status: labeling_4d, labeling_2d, or final_review (depending on which stage needs correction)

completed

All annotations are finalized and approved through every review stage. The sequence is ready for export or delivery.
  • Triggered by: Final review approval
  • Next status: customer_approved (when the customer confirms acceptance)

customer_approved

The customer or external stakeholder has reviewed the delivered annotations and confirmed they meet acceptance criteria. This is the terminal status.
  • Triggered by: Customer approval action
  • Next status: None (terminal status)
Not all statuses apply to every dataset. For 2D-only datasets (images, video without point clouds), the labeling_4d, review_4d, and ready_for_2d stages are skipped. Sequences move directly from ready_for_annotation to labeling_2d.

Filtering by Status

You can filter sequences by status throughout Avala:
  • Sequence list: Use the status filter dropdown to show sequences in specific states
  • Smart views: Create saved views filtered by status for quick access
  • API: Filter sequences by status in the GET /sequences endpoint
  • Exports: Export only sequences in completed or customer_approved status

Status Counts

The project dashboard displays a summary of sequences by status, giving you an at-a-glance view of pipeline throughput:

Next Steps