Module image_processing 

Image processing utilities — resize, convert, and transform images.

Mirrors packages/coding-agent/src/utils/image-*.ts. Provides cross-platform image processing for the agent’s vision capabilities.

Features

• Resize: Scale images while maintaining aspect ratio
• Convert: Convert between image formats (PNG, JPEG, WebP)
• EXIF handling: Strip EXIF data for privacy
• Optimization: Reduce file size for API transmission

Example

use rpi_tools::image_processing::{process_image, ImageProcessingOptions, OutputFormat};

let options = ImageProcessingOptions {
    max_width: Some(1024),
    max_height: Some(1024),
    format: Some(OutputFormat::Jpeg { quality: 85 }),
    strip_exif: true,
};

let processed = process_image(input_bytes, options)?;

Structs

ImageProcessingOptions

Options for image processing.

InlineImageProcessor

Shared processor for the built-in read tool.

Enums

OutputFormat

Output format for processed images.

Constants

INLINE_IMAGE_MAX_BASE64_BYTES

INLINE_IMAGE_MAX_SIDE

Pi’s default inline limits: 2000 pixels per side and 4.5 MiB of base64.

Functions

detect_image_mime

Detect the MIME type of an image from its bytes.

get_image_dimensions

Get image dimensions without fully decoding the image.

normalize_image_blocks

Normalize image blocks while retaining order and surfacing omissions as text, rather than sending an invalid image to a provider.

prepare_inline_image

Decode, orient, resize and encode an image for model input. Small supported images retain their original bytes; oversized images fall back to JPEG.

process_image

Process an image according to the given options.