Configuration¶

Dataset Configuration¶

Dataset paths are configured via YAML files. The DATASETS_YAML environment variable selects which config to use.

Local Development¶

# wire_detection/config/datasets.yaml
hand_drawn:
  path: /path/to/roboflow_test
  image_glob: "**/images/*.jpg"
  label_format: yolo_obb

synthetic:
  path: /path/to/dataset_pose
  image_glob: "train/images/*.jpg"
  label_format: yolov8_pose

hdc:
  path: /path/to/roboflow_test2
  image_glob: "**/images/*.jpg"
  label_format: yolo_obb
  component_labels: true

database:
  path: /path/to/Database
  image_glob: "*/*.jpg"
  label_format: null

Docker¶

# wire_detection/config/datasets.docker.yaml
hand_drawn:
  path: /data/hand_drawn
  image_glob: "**/images/*.jpg"

hdc:
  path: /data/hdc
  image_glob: "**/images/*.jpg"
  label_format: yolo_obb
  component_labels: true

Datasets¶

Key	Images	Labels	Purpose
`hand_drawn`	140	YOLO OBB (wire polygons)	Primary wire detection benchmark
`hdc`	1,993	YOLO OBB (57 component classes)	Component detection + masked wire detection
`synthetic`	2,000+	YOLOv8 pose (2 keypoints)	Large-scale parameter sweeps
`database`	662	None	Visual inspection / real-world testing

Environment Variables¶

Variable	Default	Description
`DATASETS_YAML`	`config/datasets.yaml`	Path to dataset config YAML
`NEXT_PUBLIC_API_URL`	`http://localhost:8000`	Backend URL for the frontend

Docker Compose¶

services:
  backend:
    build:
      context: .
      dockerfile: Dockerfile.python
    ports: ["8000:8000"]
    volumes:
      - /path/to/datasets:/data
    environment:
      - DATASETS_YAML=/app/wire_detection/config/datasets.docker.yaml

  frontend:
    build: ./ui
    ports: ["3000:3000"]
    environment:
      - NEXT_PUBLIC_API_URL=http://localhost:8000