WilData CLI Reference

Complete command-line interface reference for the wildata CLI.

WilData uses Typer for its CLI, providing dataset import, transformation, ROI extraction, visualization, GPS management, and a REST API server.

wildata [COMMAND] [OPTIONS]

---

version

Show version information.

wildata version

---

import-dataset

Import a dataset from various formats (COCO, YOLO, Label Studio) into the WilData pipeline.

wildata import-dataset [SOURCE_PATH] [OPTIONS]

You can use either a config file (--config) or provide arguments directly — not both.

Using a Config File (Recommended)

wildata import-dataset --config configs/import-config-example.yaml

Using Direct Arguments

wildata import-dataset annotations.json --format coco --name my_dataset

All Options

Argument / Option	Type	Default	Description
SOURCE_PATH	PATH	None	Path to source dataset (argument)
-c, --config	PATH	None	Path to YAML config file
-f, --format	str	None	Source format: coco, yolo, or ls
-n, --name	str	None	Dataset name
-r, --root	str	data	Root directory for data storage
-s, --split	str	train	Split name: train, val, or test
-m, --mode	str	batch	Processing mode: streaming or batch
--track-dvc	bool	false	Track dataset with DVC
--bbox-tolerance	int	5	Bounding box validation tolerance
--dotenv	str	None	Path to .env file
--ls-config	str	None	Label Studio XML config path
--parse-ls-config	bool	false	Parse Label Studio config
--disable-roi	bool	false	Disable ROI extraction
-v, --verbose	bool	false	Verbose output

Transformation Options

Option	Type	Default	Description
--enable-bbox-clipping	bool	None	Enable bounding box clipping
--bbox-clipping-tolerance	int	5	Bbox clipping tolerance
--skip-invalid-bbox	bool	false	Skip invalid bounding boxes
--enable-augmentation	bool	None	Enable data augmentation
--aug-prob	float	1.0	Augmentation probability
--num-augs	int	2	Number of augmentations per image
--enable-tiling	bool	None	Enable image tiling
--tile-size	int	512	Tile size in pixels
--tile-stride	int	416	Tile stride in pixels
--min-visibility	float	0.1	Minimum visibility ratio for bboxes in tiles

Examples:

# From config (recommended for production)
wildata import-dataset -c configs/import-config-example.yaml

# Direct with tiling
wildata import-dataset annotations.json \
    --format coco \
    --name wildlife_train \
    --enable-tiling \
    --tile-size 800 \
    --tile-stride 640

# With verbose output
wildata import-dataset -c my_config.yaml -v

See Import Config Reference for full YAML config documentation.

---

bulk-import-datasets

Bulk import multiple datasets from all files in a directory using multiprocessing.

wildata bulk-import-datasets [OPTIONS]

Option	Type	Default	Description
-c, --config	PATH	(required)	Path to YAML config file
-v, --verbose	bool	false	Verbose output
-n, --num-workers	int	2	Number of parallel workers

Each file in the source directory is imported as a separate dataset. Dataset names are derived from filenames.

Example:

wildata bulk-import-datasets -c configs/bulk-import-train.yaml -n 4

---

create-roi-dataset

Create an ROI (Region of Interest) classification dataset from a source detection dataset.

wildata create-roi-dataset [OPTIONS]

Option	Type	Default	Description
-c, --config	PATH	(required)	Path to YAML config file
-v, --verbose	bool	false	Verbose output

Example:

wildata create-roi-dataset -c configs/roi-create-config.yaml

See ROI Config Reference for config documentation.

---

bulk-create-roi-datasets

Bulk create ROI datasets from all files in a directory using multiprocessing.

wildata bulk-create-roi-datasets [OPTIONS]

Option	Type	Default	Description
-c, --config	PATH	(required)	Path to YAML config file
-v, --verbose	bool	false	Verbose output
-n, --num-workers	int	2	Number of parallel workers

Example:

wildata bulk-create-roi-datasets -c configs/bulk-roi-create-config.yaml -n 4

---

list-datasets

List all available datasets in the data root.

wildata list-datasets [OPTIONS]

Option	Type	Default	Description
-r, --root	str	data	Root directory for data storage
-v, --verbose	bool	false	Show detailed info per dataset

Example:

# Basic listing
wildata list-datasets

# With details
wildata list-datasets -v --root D:/data

---

visualize-classification

Visualize a classification dataset in FiftyOne.

wildata visualize-classification DATASET_NAME [OPTIONS]

Argument / Option	Type	Default	Description
DATASET_NAME	str	(required)	Name for the FiftyOne dataset
--root	str	None	Root data directory
--single-class	bool	false	Load as single class
--background-class	str	background	Background class name
--single-class-name	str	wildlife	Single class name
--keep-classes	str	None	Comma-separated list of classes to keep
--discard-classes	str	None	Comma-separated list of classes to discard
--split	str	train	Dataset split

Example:

wildata visualize-classification my_roi_dataset \
    --root D:/data/roi \
    --single-class \
    --split train

---

visualize-detection

Visualize a detection dataset in FiftyOne.

wildata visualize-detection DATASET_NAME [OPTIONS]

Argument / Option	Type	Default	Description
DATASET_NAME	str	(required)	Name for the FiftyOne dataset
--root	str	(required)	Root data directory
--split	str	train	Dataset split

Example:

wildata visualize-detection my_detection_dataset --root D:/data/detection --split val

---

update-gps-from-csv

Update EXIF GPS metadata for images using coordinates from a CSV file.

wildata update-gps-from-csv [OPTIONS]

You can use either a config file (--config) or provide arguments directly — not both.

All Options

Option	Type	Default	Description
-c, --config	PATH	None	Path to YAML config file
-i, --image-folder	PATH	None	Path to folder containing images
--csv	PATH	None	Path to CSV file with GPS coordinates
-o, --output	PATH	None	Output directory for updated images
--skip-rows	int	0	Number of rows to skip in CSV
--filename-col	str	filename	CSV column name for filenames
--lat-col	str	latitude	CSV column name for latitude
--lon-col	str	longitude	CSV column name for longitude
--alt-col	str	altitude	CSV column name for altitude
-v, --verbose	bool	false	Verbose output

Examples:

# Using config file
wildata update-gps-from-csv -c configs/gps-update-config-example.yaml

# Using direct arguments
wildata update-gps-from-csv \
    --image-folder images/ \
    --csv coordinates.csv \
    --output updated_images/ \
    --skip-rows 4

See GPS Update Config Reference for YAML config details.

---

api — API Server Commands

REST API server for programmatic access to WilData features.

api serve

Start the WilData API server.

wildata api serve [OPTIONS]

Option	Type	Default	Description
-h, --host	str	0.0.0.0	Host to bind to
-p, --port	int	8000	Port to bind to
-r, --reload	bool	false	Enable auto-reload
-w, --workers	int	1	Number of worker processes

Example:

wildata api serve --port 8080 --workers 4

API documentation is available at http://<host>:<port>/docs.

---

api check

Check current API configuration values.

wildata api check

---

Quick Reference

Command	Description
wildata import-dataset -c CONFIG	Import dataset from config
wildata bulk-import-datasets -c CONFIG	Bulk import datasets
wildata create-roi-dataset -c CONFIG	Create ROI dataset
wildata bulk-create-roi-datasets -c CONFIG	Bulk create ROI datasets
wildata list-datasets	List all datasets
wildata visualize-classification NAME	Visualize classification data
wildata visualize-detection NAME --root DIR	Visualize detection data
wildata update-gps-from-csv -c CONFIG	Update image GPS from CSV
wildata api serve	Start REST API server
wildata version	Show version

---

For configuration file details, see Configuration Reference.
For shell scripts, see WilData Scripts.