Skip to content

fix: add Visibility schema to keypoint template, fix sample data sizes#68

Open
shujaatTracebloc wants to merge 4 commits intodevelopfrom
feat/add-keypoint-detection-template
Open

fix: add Visibility schema to keypoint template, fix sample data sizes#68
shujaatTracebloc wants to merge 4 commits intodevelopfrom
feat/add-keypoint-detection-template

Conversation

@shujaatTracebloc
Copy link
Copy Markdown
Contributor

@shujaatTracebloc shujaatTracebloc commented May 7, 2026

  • Add Visibility column to keypoint schema so client can read visibility flags
  • Diversify keypoint labels (person, athlete, dancer) to meet min 2 labels requirement
  • Resize keypoint sample images from 256x256 to 448x448 to match target_size
  • Resize semantic segmentation sample images and masks to 512x512

Note

Low Risk
Low risk: changes are additive and template-focused, with the new DataValidator only enabled when a schema is provided for keypoint ingestion.

Overview
Adds an explicit Visibility column schema to the keypoint-detection ingestion template and passes it into CSVIngestor so visibility flags can be validated/consumed.

Updates shared validator mapping for KEYPOINT_DETECTION to optionally include DataValidator when a schema is supplied, and adjusts the keypoint template’s target_size plus sample CSV labels (diversified image_label values).

Reviewed by Cursor Bugbot for commit 6e48371. Bugbot is set up for automated code reviews on this repo. Configure here.

- Add Visibility column to keypoint schema so client can read visibility flags
- Diversify keypoint labels (person, athlete, dancer) to meet min 2 labels requirement
- Resize keypoint sample images from 256x256 to 448x448 to match target_size
- Resize semantic segmentation sample images and masks to 512x512

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@LukasWodka
Copy link
Copy Markdown
Contributor

👋 Heads-up — Code review queue is at 10 / 8

Above the WIP limit. The team convention is to review existing PRs before opening new work.

Open PRs currently in Code review (oldest first):

Pull from review before opening new work. (This is a nudge from the kanban WIP check, not a block.)

Comment thread templates/keypoint_detection/data/labels_file_sample.csv Outdated
…patibility

- Add Annotation and Visibility to schema instead of using annotation_column
- Preserves column case so client can read row["Annotation"] and row["Visibility"]
- Change target_size to (256, 256) to match actual data
- Add DataValidator for keypoint schema validation

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Copy link
Copy Markdown

@cursor cursor Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cursor Bugbot has reviewed your changes and found 2 potential issues.

Fix All in Cursor

❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.

Reviewed by Cursor Bugbot for commit 0893ff1. Configure here.

# Keypoint detection specific options
keypoint_detection_options = {
"target_size": (448, 448), # image size. Height = Width
"target_size": (256, 256), # image size. Height = Width
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Target size changed in wrong direction, breaking validation

High Severity

The target_size is changed from (448, 448) to (256, 256), but the PR description states the sample images were resized from 256×256 to 448×448 to match target_size. This reversal means the ImageResolutionValidator will now expect 256×256 images but receive 448×448 images, causing all keypoint detection validation to fail.

Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit 0893ff1. Configure here.

Comment thread templates/keypoint_detection/keypoint_detection.py
shujaatTracebloc and others added 2 commits May 8, 2026 18:14
Framework stores annotation as lowercase `annotation` in DB. Use
annotation_column parameter to map CSV Annotation column correctly.
Client PR tracebloc/tracebloc-client#252 reads row["annotation"].

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add KeypointAnnotationValidator to validate annotation JSON structure,
  coordinate ranges, bbox feasibility, visibility values, and keypoint
  count consistency before ingestion
- Fix sample CSV to use normalized [0,1] coordinates instead of pixel
  values to prevent zero-area bbox errors during client training
- Wire validator into keypoint detection validator mapping

Closes #70

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants