Struct CocoWriter 

pub struct CocoWriter { /* private fields */ }

COCO writer for generating JSON and ZIP files.

Example

use edgefirst_client::coco::{CocoDataset, CocoWriter};

let writer = CocoWriter::new();
let dataset = CocoDataset::default();
writer.write_json(&dataset, "annotations.json")?;

Implementations

impl CocoWriter

pub fn new() -> Self

Create a new COCO writer with default options.

pub fn with_options(options: CocoWriteOptions) -> Self

Create a new COCO writer with custom options.

pub fn write_json<P: AsRef<Path>>( &self, dataset: &CocoDataset, path: P, ) -> Result<(), Error>

Write COCO dataset to a JSON file.

Arguments

• dataset - The COCO dataset to write
• path - Output file path

pub fn write_zip<P: AsRef<Path>>( &self, dataset: &CocoDataset, images: impl Iterator<Item = (String, Vec<u8>)>, path: P, ) -> Result<(), Error>

Write COCO dataset to a ZIP file with images.

Creates a ZIP archive with:

• annotations/instances.json - The COCO annotations
• Images at their original relative paths

Arguments

• dataset - The COCO dataset to write
• images - Iterator of (filename, image_data) pairs
• path - Output ZIP file path

pub fn write_zip_from_dir<P: AsRef<Path>>( &self, dataset: &CocoDataset, images_dir: P, path: P, ) -> Result<(), Error>

Write COCO dataset to a ZIP file with images from a source directory.

Arguments

• dataset - The COCO dataset to write
• images_dir - Directory containing source images
• path - Output ZIP file path

pub fn write_split_by_group<P: AsRef<Path>>( &self, dataset: &CocoDataset, group_assignments: &[String], images_source: Option<&Path>, output_dir: P, ) -> Result<HashMap<String, usize>, Error>

Split a dataset by group and write each group to its own directory.

Creates a directory structure like:

output_dir/
├── train/
│   ├── annotations/instances_train.json
│   └── images/
│       └── *.jpg
└── val/
    ├── annotations/instances_val.json
    └── images/
        └── *.jpg

Arguments

• dataset - The COCO dataset to split
• group_assignments - Parallel array of group names for each image
• images_source - Optional source directory containing images to copy
• output_dir - Output root directory

Returns

HashMap of group name → number of images written

pub fn write_split_by_group_zip<P: AsRef<Path>>( &self, dataset: &CocoDataset, group_assignments: &[String], images_source: Option<&Path>, output_dir: P, ) -> Result<HashMap<String, usize>, Error>

Split a dataset by group and write each group to its own ZIP archive.

Creates ZIP archives like:

• output_dir/train.zip containing train split
• output_dir/val.zip containing val split

Arguments

• dataset - The COCO dataset to split
• group_assignments - Parallel array of group names for each image
• images_source - Optional source directory containing images to include
• output_dir - Output directory for ZIP files

Returns

HashMap of group name → number of images written

Trait Implementations

impl Default for CocoWriter

fn default() -> Self

Returns the “default value” for a type.