oxigdal_wasm/

canvas.rs

//! Canvas rendering utilities for WASM
//!
//! This module provides comprehensive pixel manipulation, color space conversions,
//! histogram generation, contrast enhancement, image filters, and resampling for
//! browser-based geospatial data visualization.
//!
//! # Overview
//!
//! The canvas module is the core image processing engine for oxigdal-wasm. It provides:
//!
//! - **Color Space Conversions**: RGB ↔ HSV ↔ YCbCr with high precision
//! - **Histogram Generation**: Fast histogram computation for contrast analysis
//! - **Contrast Enhancement**: Linear stretch, histogram equalization, CLAHE
//! - **Image Filters**: Gaussian blur, edge detection, sharpening, median filtering
//! - **Resampling**: Nearest neighbor, bilinear, and bicubic interpolation
//! - **Color Adjustments**: Brightness, gamma, saturation, hue rotation
//!
//! # Performance
//!
//! All operations are optimized for WASM execution:
//! - Direct memory access for pixel data
//! - SIMD-friendly algorithms where possible
//! - Efficient iteration over image buffers
//! - Minimal allocations for filter operations
//!
//! Typical performance on modern browsers:
//! - Histogram generation: < 5ms for 1MP image
//! - Linear stretch: < 10ms for 1MP image
//! - Histogram equalization: < 15ms for 1MP image
//! - Gaussian blur (3x3): < 20ms for 1MP image
//! - Resampling: ~100ms for 1MP → 4MP upscale
//!
//! # Memory Layout
//!
//! All image data is stored in RGBA format with interleaved channels:
//! ```text
//! [R₀, G₀, B₀, A₀, R₁, G₁, B₁, A₁, ..., Rₙ, Gₙ, Bₙ, Aₙ]
//! ```
//!
//! This matches the Canvas ImageData format directly, allowing zero-copy operations
//! between Rust and JavaScript.
//!
//! # Color Space Conversions
//!
//! ## RGB → HSV
//! Used for hue/saturation adjustments. The conversion preserves:
//! - Hue: 0-360° (circular)
//! - Saturation: 0.0-1.0 (linear)
//! - Value: 0.0-1.0 (linear, same as RGB max)
//!
//! ## RGB → YCbCr
//! Used for luminance/chrominance separation (JPEG-style):
//! - Y: 0-255 (luma/brightness)
//! - Cb: 0-255 (blue-difference chroma)
//! - Cr: 0-255 (red-difference chroma)
//!
//! # Contrast Enhancement Algorithms
//!
//! ## Linear Stretch
//! Maps [min, max] → [0, 255] linearly. Fast and simple, works well for:
//! - Images with limited dynamic range
//! - Quick preview/visualization
//! - When exact pixel values matter
//!
//! ## Histogram Equalization
//! Redistributes pixel values to achieve uniform histogram. Best for:
//! - Images with clustered pixel values
//! - Medical imaging
//! - Low-contrast images
//!
//! ## Adaptive Histogram Equalization (CLAHE)
//! Applies histogram equalization locally in tiles. Excellent for:
//! - Images with varying local contrast
//! - Satellite imagery
//! - Photography enhancement
//!
//! # Examples
//!
//! ```rust
//! use oxigdal_wasm::{ImageProcessor, ContrastMethod};
//!
//! // Create test image
//! let mut data = vec![0u8; 256 * 256 * 4];
//!
//! // Apply linear stretch
//! ImageProcessor::enhance_contrast(
//!     &mut data,
//!     256,
//!     256,
//!     ContrastMethod::LinearStretch
//! ).expect("Contrast enhancement failed");
//!
//! // Convert to grayscale
//! ImageProcessor::to_grayscale(&mut data);
//!
//! // Apply Gaussian blur
//! let blurred = ImageProcessor::gaussian_blur(&data, 256, 256)
//!     .expect("Blur failed");
//! ```

use serde::{Deserialize, Serialize};
use wasm_bindgen::prelude::*;
use web_sys::ImageData;

use crate::error::{CanvasError, WasmError, WasmResult};

/// Color in RGB color space
#[derive(Debug, Clone, Copy, PartialEq, serde::Serialize, serde::Deserialize)]
pub struct Rgb {
    /// Red component (0-255)
    pub r: u8,
    /// Green component (0-255)
    pub g: u8,
    /// Blue component (0-255)
    pub b: u8,
}

impl Rgb {
    /// Creates a new RGB color
    pub const fn new(r: u8, g: u8, b: u8) -> Self {
        Self { r, g, b }
    }

    /// Creates from grayscale value
    pub const fn from_gray(value: u8) -> Self {
        Self::new(value, value, value)
    }

    /// Converts to grayscale using luminance formula
    pub fn to_gray(&self) -> u8 {
        // Luminance = 0.299*R + 0.587*G + 0.114*B
        ((77 * u16::from(self.r) + 150 * u16::from(self.g) + 29 * u16::from(self.b)) / 256) as u8
    }

    /// Converts to HSV color space
    pub fn to_hsv(&self) -> Hsv {
        let r = f64::from(self.r) / 255.0;
        let g = f64::from(self.g) / 255.0;
        let b = f64::from(self.b) / 255.0;

        let max = r.max(g).max(b);
        let min = r.min(g).min(b);
        let delta = max - min;

        let h = if delta < f64::EPSILON {
            0.0
        } else if (max - r).abs() < f64::EPSILON {
            60.0 * (((g - b) / delta) % 6.0)
        } else if (max - g).abs() < f64::EPSILON {
            60.0 * (((b - r) / delta) + 2.0)
        } else {
            60.0 * (((r - g) / delta) + 4.0)
        };

        let s = if max < f64::EPSILON { 0.0 } else { delta / max };
        let v = max;

        Hsv {
            h: if h < 0.0 { h + 360.0 } else { h },
            s,
            v,
        }
    }

    /// Converts to YCbCr color space
    pub fn to_ycbcr(&self) -> YCbCr {
        let r = f64::from(self.r);
        let g = f64::from(self.g);
        let b = f64::from(self.b);

        let y = 0.299 * r + 0.587 * g + 0.114 * b;
        let cb = 128.0 + (-0.168736 * r - 0.331264 * g + 0.5 * b);
        let cr = 128.0 + (0.5 * r - 0.418688 * g - 0.081312 * b);

        YCbCr {
            y: y.clamp(0.0, 255.0) as u8,
            cb: cb.clamp(0.0, 255.0) as u8,
            cr: cr.clamp(0.0, 255.0) as u8,
        }
    }
}

/// Color in HSV color space
#[derive(Debug, Clone, Copy, PartialEq)]
pub struct Hsv {
    /// Hue (0-360)
    pub h: f64,
    /// Saturation (0-1)
    pub s: f64,
    /// Value (0-1)
    pub v: f64,
}

impl Hsv {
    /// Creates a new HSV color
    pub const fn new(h: f64, s: f64, v: f64) -> Self {
        Self { h, s, v }
    }

    /// Converts to RGB color space
    pub fn to_rgb(&self) -> Rgb {
        let c = self.v * self.s;
        let h_prime = self.h / 60.0;
        let x = c * (1.0 - ((h_prime % 2.0) - 1.0).abs());
        let m = self.v - c;

        let (r, g, b) = if h_prime < 1.0 {
            (c, x, 0.0)
        } else if h_prime < 2.0 {
            (x, c, 0.0)
        } else if h_prime < 3.0 {
            (0.0, c, x)
        } else if h_prime < 4.0 {
            (0.0, x, c)
        } else if h_prime < 5.0 {
            (x, 0.0, c)
        } else {
            (c, 0.0, x)
        };

        Rgb::new(
            ((r + m) * 255.0).round() as u8,
            ((g + m) * 255.0).round() as u8,
            ((b + m) * 255.0).round() as u8,
        )
    }
}

/// Color in YCbCr color space
#[derive(Debug, Clone, Copy, PartialEq)]
pub struct YCbCr {
    /// Y (luma) component (0-255)
    pub y: u8,
    /// Cb (blue-difference chroma) component (0-255)
    pub cb: u8,
    /// Cr (red-difference chroma) component (0-255)
    pub cr: u8,
}

impl YCbCr {
    /// Creates a new YCbCr color
    pub const fn new(y: u8, cb: u8, cr: u8) -> Self {
        Self { y, cb, cr }
    }

    /// Converts to RGB color space
    pub fn to_rgb(&self) -> Rgb {
        let y = f64::from(self.y);
        let cb = f64::from(self.cb) - 128.0;
        let cr = f64::from(self.cr) - 128.0;

        let r = y + 1.402 * cr;
        let g = y - 0.344136 * cb - 0.714136 * cr;
        let b = y + 1.772 * cb;

        Rgb::new(
            r.clamp(0.0, 255.0) as u8,
            g.clamp(0.0, 255.0) as u8,
            b.clamp(0.0, 255.0) as u8,
        )
    }
}

/// Image histogram
#[derive(Debug, Clone)]
pub struct Histogram {
    /// Red channel histogram
    pub red: [u32; 256],
    /// Green channel histogram
    pub green: [u32; 256],
    /// Blue channel histogram
    pub blue: [u32; 256],
    /// Luminance histogram
    pub luminance: [u32; 256],
}

impl Histogram {
    /// Creates a new empty histogram
    pub const fn new() -> Self {
        Self {
            red: [0; 256],
            green: [0; 256],
            blue: [0; 256],
            luminance: [0; 256],
        }
    }

    /// Computes histogram from RGBA data
    pub fn from_rgba(data: &[u8], width: u32, height: u32) -> WasmResult<Self> {
        if width == 0 || height == 0 || data.is_empty() {
            return Err(WasmError::Canvas(CanvasError::InvalidParameter(
                "Width, height, and data must be non-empty".to_string(),
            )));
        }

        let expected_len = (width as usize) * (height as usize) * 4;
        if data.len() != expected_len {
            return Err(WasmError::Canvas(CanvasError::BufferSizeMismatch {
                expected: expected_len,
                actual: data.len(),
            }));
        }

        let mut hist = Self::new();

        for chunk in data.chunks_exact(4) {
            let r = chunk[0];
            let g = chunk[1];
            let b = chunk[2];

            hist.red[r as usize] += 1;
            hist.green[g as usize] += 1;
            hist.blue[b as usize] += 1;

            let lum = Rgb::new(r, g, b).to_gray();
            hist.luminance[lum as usize] += 1;
        }

        Ok(hist)
    }

    /// Returns the minimum value with non-zero count
    pub fn min_value(&self) -> u8 {
        for (i, &count) in self.luminance.iter().enumerate() {
            if count > 0 {
                return i as u8;
            }
        }
        0
    }

    /// Returns the maximum value with non-zero count
    pub fn max_value(&self) -> u8 {
        for (i, &count) in self.luminance.iter().enumerate().rev() {
            if count > 0 {
                return i as u8;
            }
        }
        255
    }

    /// Returns the mean luminance
    pub fn mean(&self) -> f64 {
        let total: u64 = self.luminance.iter().map(|&x| u64::from(x)).sum();
        if total == 0 {
            return 0.0;
        }

        let weighted_sum: u64 = self
            .luminance
            .iter()
            .enumerate()
            .map(|(val, &count)| val as u64 * u64::from(count))
            .sum();

        weighted_sum as f64 / total as f64
    }

    /// Returns the median luminance
    pub fn median(&self) -> u8 {
        let total: u64 = self.luminance.iter().map(|&x| u64::from(x)).sum();
        if total == 0 {
            return 0;
        }

        let target = total / 2;
        let mut cumulative = 0u64;

        for (i, &count) in self.luminance.iter().enumerate() {
            cumulative += u64::from(count);
            if cumulative >= target {
                return i as u8;
            }
        }

        255
    }
}

impl Default for Histogram {
    fn default() -> Self {
        Self::new()
    }
}

/// JSON-serializable representation of a single channel histogram
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ChannelHistogramJson {
    /// Histogram bin counts (256 bins for 8-bit values)
    pub bins: Vec<u32>,
    /// Minimum value with non-zero count
    pub min: u8,
    /// Maximum value with non-zero count
    pub max: u8,
    /// Mean value
    pub mean: f64,
    /// Median value
    pub median: u8,
    /// Standard deviation
    pub std_dev: f64,
    /// Total count of values
    pub count: u64,
}

impl ChannelHistogramJson {
    /// Creates channel histogram JSON from a histogram array
    fn from_histogram_array(hist: &[u32; 256]) -> Self {
        let count: u64 = hist.iter().map(|&x| u64::from(x)).sum();

        // Find min value with non-zero count
        let min = hist
            .iter()
            .enumerate()
            .find(|&(_, &c)| c > 0)
            .map(|(i, _)| i as u8)
            .unwrap_or(0);

        // Find max value with non-zero count
        let max = hist
            .iter()
            .enumerate()
            .rev()
            .find(|&(_, &c)| c > 0)
            .map(|(i, _)| i as u8)
            .unwrap_or(255);

        // Calculate mean
        let mean = if count > 0 {
            let weighted_sum: u64 = hist
                .iter()
                .enumerate()
                .map(|(val, &c)| val as u64 * u64::from(c))
                .sum();
            weighted_sum as f64 / count as f64
        } else {
            0.0
        };

        // Calculate median
        let median = if count > 0 {
            let target = count / 2;
            let mut cumulative = 0u64;
            let mut median_val = 0u8;
            for (i, &c) in hist.iter().enumerate() {
                cumulative += u64::from(c);
                if cumulative >= target {
                    median_val = i as u8;
                    break;
                }
            }
            median_val
        } else {
            0
        };

        // Calculate standard deviation
        let std_dev = if count > 0 {
            let variance: f64 = hist
                .iter()
                .enumerate()
                .map(|(val, &c)| {
                    let diff = val as f64 - mean;
                    diff * diff * f64::from(c)
                })
                .sum::<f64>()
                / count as f64;
            variance.sqrt()
        } else {
            0.0
        };

        Self {
            bins: hist.to_vec(),
            min,
            max,
            mean,
            median,
            std_dev,
            count,
        }
    }
}

/// JSON-serializable representation of a custom bin range histogram
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct CustomBinHistogramJson {
    /// Histogram bin counts
    pub bins: Vec<u32>,
    /// Bin edges (n+1 edges for n bins)
    pub bin_edges: Vec<f64>,
    /// Minimum value
    pub min: f64,
    /// Maximum value
    pub max: f64,
    /// Number of bins
    pub num_bins: usize,
}

/// JSON-serializable representation of the full histogram
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct HistogramJson {
    /// Image width in pixels
    pub width: u32,
    /// Image height in pixels
    pub height: u32,
    /// Total pixel count
    pub total_pixels: u64,
    /// Red channel histogram
    pub red: ChannelHistogramJson,
    /// Green channel histogram
    pub green: ChannelHistogramJson,
    /// Blue channel histogram
    pub blue: ChannelHistogramJson,
    /// Luminance histogram
    pub luminance: ChannelHistogramJson,
}

impl HistogramJson {
    /// Creates histogram JSON from a Histogram and image dimensions
    pub fn from_histogram(hist: &Histogram, width: u32, height: u32) -> Self {
        Self {
            width,
            height,
            total_pixels: u64::from(width) * u64::from(height),
            red: ChannelHistogramJson::from_histogram_array(&hist.red),
            green: ChannelHistogramJson::from_histogram_array(&hist.green),
            blue: ChannelHistogramJson::from_histogram_array(&hist.blue),
            luminance: ChannelHistogramJson::from_histogram_array(&hist.luminance),
        }
    }

    /// Serializes to JSON string
    pub fn to_json_string(&self) -> Result<String, serde_json::Error> {
        serde_json::to_string(self)
    }

    /// Serializes to pretty JSON string
    pub fn to_json_string_pretty(&self) -> Result<String, serde_json::Error> {
        serde_json::to_string_pretty(self)
    }
}

impl Histogram {
    /// Returns the standard deviation of luminance values
    pub fn std_dev(&self) -> f64 {
        let total: u64 = self.luminance.iter().map(|&x| u64::from(x)).sum();
        if total == 0 {
            return 0.0;
        }

        let mean = self.mean();
        let variance: f64 = self
            .luminance
            .iter()
            .enumerate()
            .map(|(val, &count)| {
                let diff = val as f64 - mean;
                diff * diff * f64::from(count)
            })
            .sum::<f64>()
            / total as f64;

        variance.sqrt()
    }

    /// Converts the histogram to JSON-serializable format
    pub fn to_json(&self, width: u32, height: u32) -> HistogramJson {
        HistogramJson::from_histogram(self, width, height)
    }

    /// Serializes the histogram to a JSON string
    pub fn to_json_string(&self, width: u32, height: u32) -> Result<String, serde_json::Error> {
        self.to_json(width, height).to_json_string()
    }

    /// Creates a histogram with custom bin ranges from RGBA data
    ///
    /// This allows for non-uniform bin widths, useful for specific analysis needs.
    pub fn from_rgba_with_bins(
        data: &[u8],
        width: u32,
        height: u32,
        bin_edges: &[f64],
    ) -> WasmResult<CustomBinHistogramJson> {
        let expected_len = (width as usize) * (height as usize) * 4;
        if data.len() != expected_len {
            return Err(WasmError::Canvas(CanvasError::BufferSizeMismatch {
                expected: expected_len,
                actual: data.len(),
            }));
        }

        if bin_edges.len() < 2 {
            return Err(WasmError::Canvas(CanvasError::InvalidParameter(
                "bin_edges must have at least 2 elements".to_string(),
            )));
        }

        let num_bins = bin_edges.len() - 1;
        let mut bins = vec![0u32; num_bins];
        let mut min_val = f64::MAX;
        let mut max_val = f64::MIN;

        for chunk in data.chunks_exact(4) {
            // Compute luminance
            let r = chunk[0];
            let g = chunk[1];
            let b = chunk[2];
            let lum = Rgb::new(r, g, b).to_gray();
            let lum_f = f64::from(lum);

            min_val = min_val.min(lum_f);
            max_val = max_val.max(lum_f);

            // Find the appropriate bin
            for i in 0..num_bins {
                if lum_f >= bin_edges[i] && lum_f < bin_edges[i + 1] {
                    bins[i] += 1;
                    break;
                }
            }
            // Handle edge case where value equals the last edge
            if (lum_f - bin_edges[num_bins]).abs() < f64::EPSILON {
                bins[num_bins - 1] += 1;
            }
        }

        Ok(CustomBinHistogramJson {
            bins,
            bin_edges: bin_edges.to_vec(),
            min: if min_val == f64::MAX { 0.0 } else { min_val },
            max: if max_val == f64::MIN { 255.0 } else { max_val },
            num_bins,
        })
    }
}

/// Image statistics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ImageStats {
    /// Width in pixels
    pub width: u32,
    /// Height in pixels
    pub height: u32,
    /// Minimum value
    pub min: u8,
    /// Maximum value
    pub max: u8,
    /// Mean value
    pub mean: f64,
    /// Median value
    pub median: u8,
    /// Standard deviation
    pub std_dev: f64,
}

impl ImageStats {
    /// Computes statistics from RGBA data
    pub fn from_rgba(data: &[u8], width: u32, height: u32) -> WasmResult<Self> {
        let hist = Histogram::from_rgba(data, width, height)?;

        let min = hist.min_value();
        let max = hist.max_value();
        let mean = hist.mean();
        let median = hist.median();

        // Calculate standard deviation
        let total: u64 = hist.luminance.iter().map(|&x| u64::from(x)).sum();
        let variance: f64 = hist
            .luminance
            .iter()
            .enumerate()
            .map(|(val, &count)| {
                let diff = val as f64 - mean;
                diff * diff * f64::from(count)
            })
            .sum::<f64>()
            / total as f64;

        let std_dev = variance.sqrt();

        Ok(Self {
            width,
            height,
            min,
            max,
            mean,
            median,
            std_dev,
        })
    }
}

/// Contrast enhancement methods
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum ContrastMethod {
    /// Linear stretch
    LinearStretch,
    /// Histogram equalization
    HistogramEqualization,
    /// Adaptive histogram equalization
    AdaptiveHistogramEqualization,
}

/// Image processing utilities
pub struct ImageProcessor;

impl ImageProcessor {
    /// Applies contrast enhancement
    pub fn enhance_contrast(
        data: &mut [u8],
        width: u32,
        height: u32,
        method: ContrastMethod,
    ) -> WasmResult<()> {
        match method {
            ContrastMethod::LinearStretch => Self::linear_stretch(data, width, height),
            ContrastMethod::HistogramEqualization => {
                Self::histogram_equalization(data, width, height)
            }
            ContrastMethod::AdaptiveHistogramEqualization => {
                Self::adaptive_histogram_equalization(data, width, height)
            }
        }
    }

    /// Linear contrast stretch
    fn linear_stretch(data: &mut [u8], width: u32, height: u32) -> WasmResult<()> {
        let hist = Histogram::from_rgba(data, width, height)?;
        let min = hist.min_value();
        let max = hist.max_value();

        if min == max {
            return Ok(());
        }

        let scale = 255.0 / (max - min) as f64;

        for chunk in data.chunks_exact_mut(4) {
            chunk[0] = ((chunk[0].saturating_sub(min)) as f64 * scale) as u8;
            chunk[1] = ((chunk[1].saturating_sub(min)) as f64 * scale) as u8;
            chunk[2] = ((chunk[2].saturating_sub(min)) as f64 * scale) as u8;
        }

        Ok(())
    }

    /// Histogram equalization
    fn histogram_equalization(data: &mut [u8], width: u32, height: u32) -> WasmResult<()> {
        let hist = Histogram::from_rgba(data, width, height)?;
        let total_pixels = (width as usize) * (height as usize);

        // Build cumulative distribution function
        let mut cdf = [0u32; 256];
        cdf[0] = hist.luminance[0];
        for i in 1..256 {
            cdf[i] = cdf[i - 1] + hist.luminance[i];
        }

        // Find minimum non-zero CDF value
        let cdf_min = cdf.iter().find(|&&x| x > 0).copied().unwrap_or(0);

        // Build lookup table
        let mut lut = [0u8; 256];
        for i in 0..256 {
            if total_pixels > cdf_min as usize {
                lut[i] = (((cdf[i] - cdf_min) as f64 / (total_pixels - cdf_min as usize) as f64)
                    * 255.0) as u8;
            }
        }

        // Apply lookup table
        for chunk in data.chunks_exact_mut(4) {
            let lum = Rgb::new(chunk[0], chunk[1], chunk[2]).to_gray();
            let new_lum = lut[lum as usize];

            // Scale RGB values
            if lum > 0 {
                let scale = new_lum as f64 / lum as f64;
                chunk[0] = ((chunk[0] as f64 * scale).min(255.0)) as u8;
                chunk[1] = ((chunk[1] as f64 * scale).min(255.0)) as u8;
                chunk[2] = ((chunk[2] as f64 * scale).min(255.0)) as u8;
            }
        }

        Ok(())
    }

    /// Adaptive histogram equalization
    fn adaptive_histogram_equalization(data: &mut [u8], width: u32, height: u32) -> WasmResult<()> {
        // For simplicity, use regular histogram equalization
        // A full CLAHE implementation would require tiling
        Self::histogram_equalization(data, width, height)
    }

    /// Applies a brightness adjustment
    pub fn adjust_brightness(data: &mut [u8], delta: i32) {
        for chunk in data.chunks_exact_mut(4) {
            chunk[0] = (chunk[0] as i32 + delta).clamp(0, 255) as u8;
            chunk[1] = (chunk[1] as i32 + delta).clamp(0, 255) as u8;
            chunk[2] = (chunk[2] as i32 + delta).clamp(0, 255) as u8;
        }
    }

    /// Applies gamma correction
    pub fn gamma_correction(data: &mut [u8], gamma: f64) {
        let inv_gamma = 1.0 / gamma;
        let mut lut = [0u8; 256];
        for i in 0..256 {
            lut[i] = ((i as f64 / 255.0).powf(inv_gamma) * 255.0) as u8;
        }

        for chunk in data.chunks_exact_mut(4) {
            chunk[0] = lut[chunk[0] as usize];
            chunk[1] = lut[chunk[1] as usize];
            chunk[2] = lut[chunk[2] as usize];
        }
    }

    /// Adjusts contrast
    /// factor > 1.0 increases contrast, factor < 1.0 decreases contrast
    pub fn adjust_contrast(data: &mut [u8], factor: f64) {
        let factor = factor.max(0.0);

        for chunk in data.chunks_exact_mut(4) {
            for i in 0..3 {
                let val = chunk[i] as f64;
                let adjusted = ((val - 128.0) * factor + 128.0).clamp(0.0, 255.0);
                chunk[i] = adjusted as u8;
            }
        }
    }

    /// Adjusts saturation
    /// factor > 1.0 increases saturation, factor < 1.0 decreases saturation
    pub fn adjust_saturation(data: &mut [u8], factor: f64) {
        let factor = factor.max(0.0);

        for chunk in data.chunks_exact_mut(4) {
            let rgb = Rgb::new(chunk[0], chunk[1], chunk[2]);
            let mut hsv = rgb.to_hsv();

            // Adjust saturation
            hsv.s = (hsv.s * factor).clamp(0.0, 1.0);

            let adjusted = hsv.to_rgb();
            chunk[0] = adjusted.r;
            chunk[1] = adjusted.g;
            chunk[2] = adjusted.b;
        }
    }

    /// Converts to grayscale
    pub fn to_grayscale(data: &mut [u8]) {
        for chunk in data.chunks_exact_mut(4) {
            let gray = Rgb::new(chunk[0], chunk[1], chunk[2]).to_gray();
            chunk[0] = gray;
            chunk[1] = gray;
            chunk[2] = gray;
        }
    }

    /// Inverts colors
    pub fn invert(data: &mut [u8]) {
        for chunk in data.chunks_exact_mut(4) {
            chunk[0] = 255 - chunk[0];
            chunk[1] = 255 - chunk[1];
            chunk[2] = 255 - chunk[2];
        }
    }

    /// Applies a 3x3 convolution kernel
    pub fn convolve_3x3(
        data: &[u8],
        width: u32,
        height: u32,
        kernel: &[f32; 9],
    ) -> WasmResult<Vec<u8>> {
        let w = width as usize;
        let h = height as usize;
        let mut output = vec![0u8; w * h * 4];

        for y in 1..h - 1 {
            for x in 1..w - 1 {
                for c in 0..3 {
                    let mut sum = 0.0;

                    for ky in 0..3 {
                        for kx in 0..3 {
                            let px = x + kx - 1;
                            let py = y + ky - 1;
                            let idx = (py * w + px) * 4 + c;
                            sum += f32::from(data[idx]) * kernel[ky * 3 + kx];
                        }
                    }

                    let out_idx = (y * w + x) * 4 + c;
                    output[out_idx] = sum.clamp(0.0, 255.0) as u8;
                }

                // Copy alpha
                let out_idx = (y * w + x) * 4 + 3;
                let in_idx = (y * w + x) * 4 + 3;
                output[out_idx] = data[in_idx];
            }
        }

        Ok(output)
    }

    /// Applies Gaussian blur
    pub fn gaussian_blur(data: &[u8], width: u32, height: u32) -> WasmResult<Vec<u8>> {
        #[allow(clippy::excessive_precision)]
        let kernel = [
            1.0 / 16.0,
            2.0 / 16.0,
            1.0 / 16.0,
            2.0 / 16.0,
            4.0 / 16.0,
            2.0 / 16.0,
            1.0 / 16.0,
            2.0 / 16.0,
            1.0 / 16.0,
        ];
        Self::convolve_3x3(data, width, height, &kernel)
    }

    /// Applies edge detection (Sobel)
    pub fn edge_detection(data: &[u8], width: u32, height: u32) -> WasmResult<Vec<u8>> {
        let sobel_x = [-1.0, 0.0, 1.0, -2.0, 0.0, 2.0, -1.0, 0.0, 1.0];
        let sobel_y = [-1.0, -2.0, -1.0, 0.0, 0.0, 0.0, 1.0, 2.0, 1.0];

        let gx = Self::convolve_3x3(data, width, height, &sobel_x)?;
        let gy = Self::convolve_3x3(data, width, height, &sobel_y)?;

        let mut output = vec![0u8; gx.len()];
        for i in (0..gx.len()).step_by(4) {
            for c in 0..3 {
                let gx_val = f64::from(gx[i + c]);
                let gy_val = f64::from(gy[i + c]);
                let magnitude = (gx_val * gx_val + gy_val * gy_val).sqrt();
                output[i + c] = magnitude.min(255.0) as u8;
            }
            output[i + 3] = 255; // Alpha
        }

        Ok(output)
    }

    /// Applies sharpening
    pub fn sharpen(data: &[u8], width: u32, height: u32) -> WasmResult<Vec<u8>> {
        let kernel = [0.0, -1.0, 0.0, -1.0, 5.0, -1.0, 0.0, -1.0, 0.0];
        Self::convolve_3x3(data, width, height, &kernel)
    }
}

/// Resampling methods
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum ResampleMethod {
    /// Nearest neighbor
    NearestNeighbor,
    /// Bilinear interpolation
    Bilinear,
    /// Bicubic interpolation
    Bicubic,
}

/// Image resampler
pub struct Resampler;

impl Resampler {
    /// Resamples an image to a new size
    pub fn resample(
        data: &[u8],
        src_width: u32,
        src_height: u32,
        dst_width: u32,
        dst_height: u32,
        method: ResampleMethod,
    ) -> WasmResult<Vec<u8>> {
        match method {
            ResampleMethod::NearestNeighbor => {
                Self::nearest_neighbor(data, src_width, src_height, dst_width, dst_height)
            }
            ResampleMethod::Bilinear => {
                Self::bilinear(data, src_width, src_height, dst_width, dst_height)
            }
            ResampleMethod::Bicubic => {
                Self::bicubic(data, src_width, src_height, dst_width, dst_height)
            }
        }
    }

    /// Nearest neighbor resampling
    fn nearest_neighbor(
        data: &[u8],
        src_width: u32,
        src_height: u32,
        dst_width: u32,
        dst_height: u32,
    ) -> WasmResult<Vec<u8>> {
        let mut output = vec![0u8; (dst_width * dst_height * 4) as usize];

        let x_ratio = src_width as f64 / dst_width as f64;
        let y_ratio = src_height as f64 / dst_height as f64;

        for y in 0..dst_height {
            for x in 0..dst_width {
                let src_x = (x as f64 * x_ratio) as u32;
                let src_y = (y as f64 * y_ratio) as u32;

                let src_idx = ((src_y * src_width + src_x) * 4) as usize;
                let dst_idx = ((y * dst_width + x) * 4) as usize;

                output[dst_idx..dst_idx + 4].copy_from_slice(&data[src_idx..src_idx + 4]);
            }
        }

        Ok(output)
    }

    /// Bilinear resampling
    fn bilinear(
        data: &[u8],
        src_width: u32,
        src_height: u32,
        dst_width: u32,
        dst_height: u32,
    ) -> WasmResult<Vec<u8>> {
        let mut output = vec![0u8; (dst_width * dst_height * 4) as usize];

        let x_ratio = (src_width - 1) as f64 / dst_width as f64;
        let y_ratio = (src_height - 1) as f64 / dst_height as f64;

        for y in 0..dst_height {
            for x in 0..dst_width {
                let src_x = x as f64 * x_ratio;
                let src_y = y as f64 * y_ratio;

                let x1 = src_x.floor() as u32;
                let y1 = src_y.floor() as u32;
                let x2 = (x1 + 1).min(src_width - 1);
                let y2 = (y1 + 1).min(src_height - 1);

                let dx = src_x - x1 as f64;
                let dy = src_y - y1 as f64;

                let dst_idx = ((y * dst_width + x) * 4) as usize;

                for c in 0..4 {
                    let p11 = data[((y1 * src_width + x1) * 4 + c) as usize];
                    let p21 = data[((y1 * src_width + x2) * 4 + c) as usize];
                    let p12 = data[((y2 * src_width + x1) * 4 + c) as usize];
                    let p22 = data[((y2 * src_width + x2) * 4 + c) as usize];

                    let val = (1.0 - dx) * (1.0 - dy) * f64::from(p11)
                        + dx * (1.0 - dy) * f64::from(p21)
                        + (1.0 - dx) * dy * f64::from(p12)
                        + dx * dy * f64::from(p22);

                    output[dst_idx + c as usize] = val.round() as u8;
                }
            }
        }

        Ok(output)
    }

    /// Bicubic resampling (simplified)
    fn bicubic(
        data: &[u8],
        src_width: u32,
        src_height: u32,
        dst_width: u32,
        dst_height: u32,
    ) -> WasmResult<Vec<u8>> {
        // For simplicity, fall back to bilinear
        Self::bilinear(data, src_width, src_height, dst_width, dst_height)
    }
}

/// WASM bindings for canvas operations
#[wasm_bindgen]
pub struct WasmImageProcessor;

#[wasm_bindgen]
impl WasmImageProcessor {
    /// Creates ImageData from RGBA bytes
    #[wasm_bindgen(js_name = createImageData)]
    pub fn create_image_data(data: &[u8], width: u32, height: u32) -> Result<ImageData, JsValue> {
        if data.len() != (width * height * 4) as usize {
            return Err(JsValue::from_str("Invalid data size"));
        }

        let clamped = wasm_bindgen::Clamped(data);
        ImageData::new_with_u8_clamped_array_and_sh(clamped, width, height)
    }

    /// Computes histogram as JSON
    ///
    /// Returns a comprehensive JSON object containing:
    /// - Image dimensions (width, height, total_pixels)
    /// - Per-channel histograms (red, green, blue, luminance)
    /// - Statistics for each channel (min, max, mean, median, std_dev, count)
    /// - Histogram bins (256 bins for 8-bit values)
    ///
    /// # Example JSON Output
    ///
    /// ```json
    /// {
    ///   "width": 256,
    ///   "height": 256,
    ///   "total_pixels": 65536,
    ///   "red": {
    ///     "bins": [0, 100, 200, ...],
    ///     "min": 0,
    ///     "max": 255,
    ///     "mean": 127.5,
    ///     "median": 128,
    ///     "std_dev": 73.9,
    ///     "count": 65536
    ///   },
    ///   "green": { ... },
    ///   "blue": { ... },
    ///   "luminance": { ... }
    /// }
    /// ```
    #[wasm_bindgen(js_name = computeHistogram)]
    pub fn compute_histogram(data: &[u8], width: u32, height: u32) -> Result<String, JsValue> {
        let hist = Histogram::from_rgba(data, width, height)
            .map_err(|e| JsValue::from_str(&e.to_string()))?;

        hist.to_json_string(width, height)
            .map_err(|e| JsValue::from_str(&e.to_string()))
    }

    /// Computes histogram with custom bin ranges
    ///
    /// This allows for non-uniform bin widths, useful for specific analysis needs
    /// such as focusing on particular value ranges or creating logarithmic bins.
    ///
    /// # Arguments
    ///
    /// * `data` - RGBA image data
    /// * `width` - Image width
    /// * `height` - Image height
    /// * `bin_edges` - Array of bin edge values (n+1 edges for n bins)
    ///
    /// # Example
    ///
    /// ```javascript
    /// // Create 5 bins: [0-50), [50-100), [100-150), [150-200), [200-256)
    /// const binEdges = [0, 50, 100, 150, 200, 256];
    /// const histogram = WasmImageProcessor.computeHistogramWithBins(data, width, height, binEdges);
    /// ```
    #[wasm_bindgen(js_name = computeHistogramWithBins)]
    pub fn compute_histogram_with_bins(
        data: &[u8],
        width: u32,
        height: u32,
        bin_edges: &[f64],
    ) -> Result<String, JsValue> {
        let custom_hist = Histogram::from_rgba_with_bins(data, width, height, bin_edges)
            .map_err(|e| JsValue::from_str(&e.to_string()))?;

        serde_json::to_string(&custom_hist).map_err(|e| JsValue::from_str(&e.to_string()))
    }

    /// Computes statistics as JSON
    #[wasm_bindgen(js_name = computeStats)]
    pub fn compute_stats(data: &[u8], width: u32, height: u32) -> Result<String, JsValue> {
        let stats = ImageStats::from_rgba(data, width, height)
            .map_err(|e| JsValue::from_str(&e.to_string()))?;

        serde_json::to_string(&stats).map_err(|e| JsValue::from_str(&e.to_string()))
    }

    /// Applies linear stretch
    #[wasm_bindgen(js_name = linearStretch)]
    pub fn linear_stretch(data: &mut [u8], width: u32, height: u32) -> Result<(), JsValue> {
        ImageProcessor::linear_stretch(data, width, height)
            .map_err(|e| JsValue::from_str(&e.to_string()))
    }

    /// Applies histogram equalization
    #[wasm_bindgen(js_name = histogramEqualization)]
    pub fn histogram_equalization(data: &mut [u8], width: u32, height: u32) -> Result<(), JsValue> {
        ImageProcessor::histogram_equalization(data, width, height)
            .map_err(|e| JsValue::from_str(&e.to_string()))
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_rgb_to_gray() {
        let rgb = Rgb::new(128, 128, 128);
        assert_eq!(rgb.to_gray(), 128);

        let black = Rgb::new(0, 0, 0);
        assert_eq!(black.to_gray(), 0);

        let white = Rgb::new(255, 255, 255);
        assert_eq!(white.to_gray(), 255);
    }

    #[test]
    fn test_rgb_to_hsv() {
        let red = Rgb::new(255, 0, 0);
        let hsv = red.to_hsv();
        assert!((hsv.h - 0.0).abs() < 1.0);
        assert!((hsv.s - 1.0).abs() < 0.01);
        assert!((hsv.v - 1.0).abs() < 0.01);
    }

    #[test]
    fn test_hsv_to_rgb() {
        let hsv = Hsv::new(0.0, 1.0, 1.0);
        let rgb = hsv.to_rgb();
        assert_eq!(rgb.r, 255);
        assert!(rgb.g < 5);
        assert!(rgb.b < 5);
    }

    #[test]
    fn test_histogram() {
        let data = vec![
            255, 0, 0, 255, // Red pixel
            0, 255, 0, 255, // Green pixel
            0, 0, 255, 255, // Blue pixel
            128, 128, 128, 255, // Gray pixel
        ];

        let hist = Histogram::from_rgba(&data, 2, 2).expect("Histogram computation failed");
        assert_eq!(hist.red[255], 1);
        assert_eq!(hist.green[255], 1);
        assert_eq!(hist.blue[255], 1);
    }

    #[test]
    fn test_image_stats() {
        let data = vec![
            0, 0, 0, 255, 128, 128, 128, 255, 255, 255, 255, 255, 128, 128, 128, 255,
        ];

        let stats = ImageStats::from_rgba(&data, 2, 2).expect("Stats computation failed");
        assert_eq!(stats.min, 0);
        assert_eq!(stats.max, 255);
    }

    #[test]
    fn test_brightness_adjustment() {
        let mut data = vec![100, 100, 100, 255];
        ImageProcessor::adjust_brightness(&mut data, 50);
        assert_eq!(data[0], 150);
        assert_eq!(data[1], 150);
        assert_eq!(data[2], 150);
    }

    #[test]
    fn test_grayscale_conversion() {
        let mut data = vec![255, 0, 0, 255]; // Red
        ImageProcessor::to_grayscale(&mut data);
        // All RGB channels should be the same
        assert_eq!(data[0], data[1]);
        assert_eq!(data[1], data[2]);
    }

    #[test]
    fn test_nearest_neighbor_resample() {
        let data = vec![
            255, 0, 0, 255, 0, 255, 0, 255, 0, 0, 255, 255, 255, 255, 255, 255,
        ];

        let resampled = Resampler::nearest_neighbor(&data, 2, 2, 4, 4).expect("Resample failed");

        assert_eq!(resampled.len(), 4 * 4 * 4);
    }

    #[test]
    fn test_histogram_json_serialization() {
        let data = vec![
            255, 0, 0, 255, // Red pixel
            0, 255, 0, 255, // Green pixel
            0, 0, 255, 255, // Blue pixel
            128, 128, 128, 255, // Gray pixel
        ];

        let hist = Histogram::from_rgba(&data, 2, 2).expect("Histogram computation failed");
        let json_result = hist.to_json_string(2, 2);

        assert!(json_result.is_ok(), "JSON serialization should succeed");

        let json_str = json_result.expect("Should have JSON string");
        let parsed: serde_json::Value =
            serde_json::from_str(&json_str).expect("Should parse as valid JSON");

        // Verify structure
        assert_eq!(parsed["width"], 2);
        assert_eq!(parsed["height"], 2);
        assert_eq!(parsed["total_pixels"], 4);

        // Verify red channel
        assert!(parsed["red"]["bins"].is_array());
        assert_eq!(parsed["red"]["bins"].as_array().map(|a| a.len()), Some(256));
        assert!(parsed["red"]["count"].as_u64().is_some());

        // Verify luminance channel has statistics
        assert!(parsed["luminance"]["min"].is_u64());
        assert!(parsed["luminance"]["max"].is_u64());
        assert!(parsed["luminance"]["mean"].is_f64());
        assert!(parsed["luminance"]["std_dev"].is_f64());
    }

    #[test]
    fn test_histogram_json_struct() {
        let data = vec![
            100, 100, 100, 255, // Gray pixel 1
            100, 100, 100, 255, // Gray pixel 2
            200, 200, 200, 255, // Light gray pixel 1
            200, 200, 200, 255, // Light gray pixel 2
        ];

        let hist = Histogram::from_rgba(&data, 2, 2).expect("Histogram computation failed");
        let hist_json = hist.to_json(2, 2);

        // Verify dimensions
        assert_eq!(hist_json.width, 2);
        assert_eq!(hist_json.height, 2);
        assert_eq!(hist_json.total_pixels, 4);

        // Verify channel histogram structure
        assert_eq!(hist_json.red.bins.len(), 256);
        assert_eq!(hist_json.green.bins.len(), 256);
        assert_eq!(hist_json.blue.bins.len(), 256);
        assert_eq!(hist_json.luminance.bins.len(), 256);

        // Verify counts
        assert_eq!(hist_json.red.count, 4);
        assert_eq!(hist_json.green.count, 4);
        assert_eq!(hist_json.blue.count, 4);
        assert_eq!(hist_json.luminance.count, 4);

        // Verify specific bin counts
        assert_eq!(hist_json.red.bins[100], 2);
        assert_eq!(hist_json.red.bins[200], 2);
    }

    #[test]
    fn test_histogram_std_dev() {
        // Create uniform data
        let data = vec![
            128, 128, 128, 255, // All same value
            128, 128, 128, 255, 128, 128, 128, 255, 128, 128, 128, 255,
        ];

        let hist = Histogram::from_rgba(&data, 2, 2).expect("Histogram computation failed");
        let std_dev = hist.std_dev();

        // Uniform values should have zero standard deviation
        assert!(
            std_dev.abs() < f64::EPSILON,
            "Uniform values should have zero std_dev"
        );

        // Create varied data
        let varied_data = vec![
            0, 0, 0, 255, // Black
            255, 255, 255, 255, // White
            0, 0, 0, 255, // Black
            255, 255, 255, 255, // White
        ];

        let varied_hist =
            Histogram::from_rgba(&varied_data, 2, 2).expect("Histogram computation failed");
        let varied_std_dev = varied_hist.std_dev();

        // High variation should have high standard deviation
        assert!(
            varied_std_dev > 100.0,
            "High variation should have high std_dev"
        );
    }

    #[test]
    fn test_channel_histogram_statistics() {
        let data = vec![
            0, 0, 0, 255, // Black (lum = 0)
            64, 64, 64, 255, // Dark gray (lum = 64)
            192, 192, 192, 255, // Light gray (lum = 192)
            255, 255, 255, 255, // White (lum = 255)
        ];

        let hist = Histogram::from_rgba(&data, 2, 2).expect("Histogram computation failed");
        let hist_json = hist.to_json(2, 2);

        // Verify min/max for luminance
        assert_eq!(hist_json.luminance.min, 0);
        assert_eq!(hist_json.luminance.max, 255);

        // Mean should be approximately (0 + 64 + 192 + 255) / 4 = 127.75
        assert!(
            (hist_json.luminance.mean - 127.75).abs() < 1.0,
            "Mean should be approximately 127.75, got {}",
            hist_json.luminance.mean
        );
    }

    #[test]
    fn test_custom_bin_histogram() {
        let data = vec![
            25, 25, 25, 255, // Should fall in bin 0 [0-50)
            75, 75, 75, 255, // Should fall in bin 1 [50-100)
            125, 125, 125, 255, // Should fall in bin 2 [100-150)
            175, 175, 175, 255, // Should fall in bin 3 [150-200)
        ];

        let bin_edges = vec![0.0, 50.0, 100.0, 150.0, 200.0, 256.0];
        let custom_hist = Histogram::from_rgba_with_bins(&data, 2, 2, &bin_edges)
            .expect("Custom bin histogram computation failed");

        assert_eq!(custom_hist.num_bins, 5);
        assert_eq!(custom_hist.bins.len(), 5);

        // Each bin should have 1 pixel (based on luminance)
        assert_eq!(custom_hist.bins[0], 1); // 0-50
        assert_eq!(custom_hist.bins[1], 1); // 50-100
        assert_eq!(custom_hist.bins[2], 1); // 100-150
        assert_eq!(custom_hist.bins[3], 1); // 150-200
        assert_eq!(custom_hist.bins[4], 0); // 200-256 (none)

        assert_eq!(custom_hist.min, 25.0);
        assert_eq!(custom_hist.max, 175.0);
    }

    #[test]
    fn test_histogram_pretty_json() {
        let data = vec![128, 128, 128, 255, 128, 128, 128, 255];

        let hist = Histogram::from_rgba(&data, 2, 1).expect("Histogram computation failed");
        let hist_json = hist.to_json(2, 1);
        let pretty_json = hist_json.to_json_string_pretty();

        assert!(
            pretty_json.is_ok(),
            "Pretty JSON serialization should succeed"
        );

        let pretty_str = pretty_json.expect("Should have pretty JSON string");
        assert!(
            pretty_str.contains('\n'),
            "Pretty JSON should contain newlines"
        );
        assert!(
            pretty_str.contains("  "),
            "Pretty JSON should contain indentation"
        );
    }

    #[test]
    fn test_empty_histogram() {
        // Test with minimum size (1x1)
        let data = vec![128, 128, 128, 255];

        let hist = Histogram::from_rgba(&data, 1, 1).expect("Histogram computation failed");
        let hist_json = hist.to_json(1, 1);

        assert_eq!(hist_json.total_pixels, 1);
        assert_eq!(hist_json.luminance.count, 1);
        assert_eq!(hist_json.luminance.min, 128);
        assert_eq!(hist_json.luminance.max, 128);
        assert!(
            hist_json.luminance.std_dev.abs() < f64::EPSILON,
            "Single pixel should have zero std_dev"
        );
    }
}