torsh-vision 0.1.2

//! Computer vision operations for ToRSh
//!
//! This crate provides PyTorch-compatible computer vision functionality including:
//! - Image transformations and augmentations
//! - Pre-trained models and model architectures
//! - Dataset loaders for common vision datasets
//! - Spatial operations and feature matching
//! - Video processing capabilities
//! - 3D visualization utilities
//!
//! Built on top of the SciRS2 ecosystem for high-performance image processing.
//!
//! # Examples
//!
//! ```rust,ignore
//! use torsh_vision::transforms::*;
//!
//! // Create a standard image transformation pipeline
//! let transform = Compose::new(vec![
//!     Box::new(Resize::new((224, 224))),
//!     Box::new(ToTensor::new()),
//!     Box::new(Normalize::new(vec![0.485, 0.456, 0.406], vec![0.229, 0.224, 0.225])),
//! ]);
//! ```

#![allow(clippy::all)]
#![allow(dead_code)]
#![allow(unused_imports)]
#![allow(ambiguous_glob_reexports)]

pub mod advanced_transforms;
pub mod benchmarks;
pub mod datasets;
pub mod datasets_impl;
pub mod error_handling;
pub mod examples;
pub mod explainability;
pub mod feature_detection_advanced;
pub mod hardware;
pub mod interactive;
pub mod io;
pub mod memory;
pub mod models;
pub mod ops;
pub mod optimized_impl;
pub mod prelude;
pub mod scirs2_integration;
pub mod segmentation_advanced;
pub mod self_supervised;
pub mod spatial;
pub mod streaming;
pub mod transforms;
pub mod unified_transforms;
pub mod utils;
pub mod video;
pub mod viz3d;

pub use advanced_transforms::*;
pub use datasets::{DatasetConfig, DatasetError, DatasetStats};
pub use datasets_impl::{CifarDataset, CocoDataset, ImageFolder, MnistDataset, VocDataset};
pub use error_handling::*;
pub use examples::*;
pub use explainability::{
    AttentionVisualizer, BaselineType, FeatureVisualizer, GradCAM, IntegratedGradients, SaliencyMap,
};
pub use hardware::*;
pub use interactive::*;
pub use io::*;
pub use memory::*;
pub use models::*;
pub use ops::{
    adjust_brightness, adjust_contrast, adjust_hue, adjust_saturation, calculate_iou, center_crop,
    edge_detection, gaussian_blur, generate_anchors, histogram_equalization, horizontal_flip,
    morphological_operation, nms, normalize, random_crop, resize, rgb_to_grayscale, rotate,
    sobel_edge_detection, vertical_flip,
};
pub use spatial::{
    distance::PatchMatcher,
    interpolation::{ImageWarper, OpticalFlowInterpolator, SpatialInterpolator},
    matching::{Feature, FeatureMatcher, Keypoint, TemplateMatcher},
    structures::{BoundingBox, PointCloudProcessor, SpatialObjectTracker},
    transforms::{GeometricProcessor, ImageRegistrar, PoseEstimator},
    FeatureMatch, SpatialConfig, SpatialPoint, SpatialProcessor, TransformResult,
};
pub use transforms::*;
pub use unified_transforms::*;
pub use utils::*;
pub use video::*;
pub use viz3d::*;

// Comprehensive scirs2-vision integration
pub use scirs2_integration::{
    ContrastMethod, CornerPoint, DenoiseMethod, DisparityMap, EdgeDetectionMethod,
    Keypoint as SciKeypoint, MemoryStrategy, OpticalFlow, OrbFeatures, QualityLevel,
    SciRS2VisionProcessor, SiftFeatures, SimdLevel, SurfFeatures, VisionConfig,
};

// Self-supervised learning augmentations
pub use self_supervised::{
    BYOLAugmentation, DINOAugmentation, GaussianBlur, MoCoAugmentation, RandomGrayscale,
    SimCLRAugmentation, Solarize, SwAVAugmentation,
};

// Advanced segmentation algorithms (NEW in 0.1.5 integration)
pub use segmentation_advanced::{
    graph_cuts, region_growing, watershed, Connectivity, GraphCutsConfig, RegionGrowingConfig,
    WatershedConfig, WatershedMarkers,
};

// Advanced feature detection and matching (NEW in 0.1.5 integration)
pub use feature_detection_advanced::{
    apply_ratio_test, AttentionMatcher, AttentionMatcherConfig, BruteForceMatcher, DistanceMetric,
    Feature as AdvancedFeature, FeatureMatch as AdvancedFeatureMatch, LearnedSiftConfig,
    LearnedSiftDetector, MultiScaleConfig, MultiScaleDetector, SuperPointConfig,
    SuperPointDetector,
};

// Real-time streaming and performance (NEW in 0.1.5 integration)
pub use streaming::{
    BatchProcessor, Frame, FrameMetadata, FramePreprocessor, QualityAdaptation, StreamConfig,
    StreamProcessor, StreamStats,
};

// Comprehensive benchmarking suite
pub use benchmarks::{
    run_full_benchmark_suite, run_quick_benchmark, AccuracyMetrics, BenchmarkConfig,
    BenchmarkResult, VisionBenchmarkSuite,
};

// Version information
pub const VERSION: &str = env!("CARGO_PKG_VERSION");
pub const VERSION_MAJOR: u32 = 0;
pub const VERSION_MINOR: u32 = 1;
pub const VERSION_PATCH: u32 = 0;

#[derive(Debug, thiserror::Error)]
pub enum VisionError {
    #[error("Image processing error: {0}")]
    ImageError(#[from] image::ImageError),

    #[error("Model error: {0}")]
    ModelError(String),

    #[error("Transform error: {0}")]
    TransformError(String),

    #[error("Invalid shape: {0}")]
    InvalidShape(String),

    #[error("Invalid argument: {0}")]
    InvalidArgument(String),

    #[error("Invalid input: {0}")]
    InvalidInput(String),

    #[error("Invalid parameter: {0}")]
    InvalidParameter(String),

    #[error("IO error: {0}")]
    IoError(#[from] std::io::Error),

    #[error("Tensor error: {0}")]
    TensorError(#[from] torsh_core::error::TorshError),

    #[cfg(feature = "pretrained")]
    #[error("Request error: {0}")]
    RequestError(#[from] reqwest::Error),

    #[error("Other error: {0}")]
    Other(#[from] anyhow::Error),
}

pub type Result<T> = std::result::Result<T, VisionError>;

// Conversion from VisionError to TorshError for Module trait compatibility
impl From<VisionError> for torsh_core::error::TorshError {
    fn from(err: VisionError) -> Self {
        torsh_core::error::TorshError::Other(err.to_string())
    }
}