Kingfisher/Sources/Image/ImageProcessor.swift

//
//  ImageProcessor.swift
//  Kingfisher
//
//  Created by Wei Wang on 2016/08/26.
//
//  Copyright (c) 2019 Wei Wang <onevcat@gmail.com>
//
//  Permission is hereby granted, free of charge, to any person obtaining a copy
//  of this software and associated documentation files (the "Software"), to deal
//  in the Software without restriction, including without limitation the rights
//  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
//  copies of the Software, and to permit persons to whom the Software is
//  furnished to do so, subject to the following conditions:
//
//  The above copyright notice and this permission notice shall be included in
//  all copies or substantial portions of the Software.
//
//  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
//  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
//  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
//  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
//  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
//  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
//  THE SOFTWARE.

import Foundation
import CoreGraphics

#if canImport(AppKit) && !targetEnvironment(macCatalyst)
import AppKit
#endif

/// Represents an item which could be processed by an `ImageProcessor`.
///
/// - image: Input image. The processor should provide a way to apply
///          processing on this `image` and return the result image.
/// - data:  Input data. The processor should provide a way to apply
///          processing on this `image` and return the result image.
public enum ImageProcessItem {
    
    /// Input image. The processor should provide a way to apply
    /// processing on this `image` and return the result image.
    case image(KFCrossPlatformImage)
    
    /// Input data. The processor should provide a way to apply
    /// processing on this `image` and return the result image.
    case data(Data)
}

/// An `ImageProcessor` would be used to convert some downloaded data to an image.
public protocol ImageProcessor {
    /// Identifier of the processor. It will be used to identify the processor when 
    /// caching and retrieving an image. You might want to make sure that processors with
    /// same properties/functionality have the same identifiers, so correct processed images
    /// could be retrieved with proper key.
    /// 
    /// - Note: Do not supply an empty string for a customized processor, which is already reserved by
    /// the `DefaultImageProcessor`. It is recommended to use a reverse domain name notation string of
    /// your own for the identifier.
    var identifier: String { get }

    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: The parsed options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: The return value should be `nil` if processing failed while converting an input item to image.
    ///         If `nil` received by the processing caller, an error will be reported and the process flow stops.
    ///         If the processing flow is not critical for your flow, then when the input item is already an image
    ///         (`.image` case) and there is any errors in the processing, you could return the input image itself
    ///         to keep the processing pipeline continuing.
    /// - Note: Most processor only supports CG-based images. watchOS is not supported for processors containing
    ///         a filter, the input image will be returned directly on watchOS.
    func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage?
}

extension ImageProcessor {
    
    /// Appends an `ImageProcessor` to another. The identifier of the new `ImageProcessor`
    /// will be "\(self.identifier)|>\(another.identifier)".
    ///
    /// - Parameter another: An `ImageProcessor` you want to append to `self`.
    /// - Returns: The new `ImageProcessor` will process the image in the order
    ///            of the two processors concatenated.
    public func append(another: ImageProcessor) -> ImageProcessor {
        let newIdentifier = identifier.appending("|>\(another.identifier)")
        return GeneralProcessor(identifier: newIdentifier) {
            item, options in
            if let image = self.process(item: item, options: options) {
                return another.process(item: .image(image), options: options)
            } else {
                return nil
            }
        }
    }
}

func ==(left: ImageProcessor, right: ImageProcessor) -> Bool {
    return left.identifier == right.identifier
}

func !=(left: ImageProcessor, right: ImageProcessor) -> Bool {
    return !(left == right)
}

typealias ProcessorImp = ((ImageProcessItem, KingfisherParsedOptionsInfo) -> KFCrossPlatformImage?)
struct GeneralProcessor: ImageProcessor {
    let identifier: String
    let p: ProcessorImp
    func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        return p(item, options)
    }
}

/// The default processor. It converts the input data to a valid image.
/// Images of .PNG, .JPEG and .GIF format are supported.
/// If an image item is given as `.image` case, `DefaultImageProcessor` will
/// do nothing on it and return the associated image.
public struct DefaultImageProcessor: ImageProcessor {
    
    /// A default `DefaultImageProcessor` could be used across.
    public static let `default` = DefaultImageProcessor()
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier = ""
    
    /// Creates a `DefaultImageProcessor`. Use `DefaultImageProcessor.default` to get an instance,
    /// if you do not have a good reason to create your own `DefaultImageProcessor`.
    public init() {}
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
        case .data(let data):
            return KingfisherWrapper.image(data: data, options: options.imageCreatingOptions)
        }
    }
}

/// Represents the rect corner setting when processing a round corner image.
public struct RectCorner: OptionSet {
    
    /// Raw value of the rect corner.
    public let rawValue: Int
    
    /// Represents the top left corner.
    public static let topLeft = RectCorner(rawValue: 1 << 0)
    
    /// Represents the top right corner.
    public static let topRight = RectCorner(rawValue: 1 << 1)
    
    /// Represents the bottom left corner.
    public static let bottomLeft = RectCorner(rawValue: 1 << 2)
    
    /// Represents the bottom right corner.
    public static let bottomRight = RectCorner(rawValue: 1 << 3)
    
    /// Represents all corners.
    public static let all: RectCorner = [.topLeft, .topRight, .bottomLeft, .bottomRight]
    
    /// Creates a `RectCorner` option set with a given value.
    ///
    /// - Parameter rawValue: The value represents a certain corner option.
    public init(rawValue: Int) {
        self.rawValue = rawValue
    }
    
    var cornerIdentifier: String {
        if self == .all {
            return ""
        }
        return "_corner(\(rawValue))"
    }
}

#if !os(macOS)
/// Processor for adding an blend mode to images. Only CG-based images are supported.
public struct BlendImageProcessor: ImageProcessor {

    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String

    /// Blend Mode will be used to blend the input image.
    public let blendMode: CGBlendMode

    /// Alpha will be used when blend image.
    public let alpha: CGFloat

    /// Background color of the output image. If `nil`, it will stay transparent.
    public let backgroundColor: KFCrossPlatformColor?

    /// Creates a `BlendImageProcessor`.
    ///
    /// - Parameters:
    ///   - blendMode: Blend Mode will be used to blend the input image.
    ///   - alpha: Alpha will be used when blend image. From 0.0 to 1.0. 1.0 means solid image,
    ///            0.0 means transparent image (not visible at all). Default is 1.0.
    ///   - backgroundColor: Background color to apply for the output image. Default is `nil`.
    public init(blendMode: CGBlendMode, alpha: CGFloat = 1.0, backgroundColor: KFCrossPlatformColor? = nil) {
        self.blendMode = blendMode
        self.alpha = alpha
        self.backgroundColor = backgroundColor
        var identifier = "com.onevcat.Kingfisher.BlendImageProcessor(\(blendMode.rawValue),\(alpha))"
        if let color = backgroundColor {
            identifier.append("_\(color.hex)")
        }
        self.identifier = identifier
    }

    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.image(withBlendMode: blendMode, alpha: alpha, backgroundColor: backgroundColor)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}
#endif

#if os(macOS)
/// Processor for adding an compositing operation to images. Only CG-based images are supported in macOS.
public struct CompositingImageProcessor: ImageProcessor {

    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String

    /// Compositing operation will be used to the input image.
    public let compositingOperation: NSCompositingOperation

    /// Alpha will be used when compositing image.
    public let alpha: CGFloat

    /// Background color of the output image. If `nil`, it will stay transparent.
    public let backgroundColor: KFCrossPlatformColor?

    /// Creates a `CompositingImageProcessor`
    ///
    /// - Parameters:
    ///   - compositingOperation: Compositing operation will be used to the input image.
    ///   - alpha: Alpha will be used when compositing image.
    ///            From 0.0 to 1.0. 1.0 means solid image, 0.0 means transparent image.
    ///            Default is 1.0.
    ///   - backgroundColor: Background color to apply for the output image. Default is `nil`.
    public init(compositingOperation: NSCompositingOperation,
                alpha: CGFloat = 1.0,
                backgroundColor: KFCrossPlatformColor? = nil)
    {
        self.compositingOperation = compositingOperation
        self.alpha = alpha
        self.backgroundColor = backgroundColor
        var identifier = "com.onevcat.Kingfisher.CompositingImageProcessor(\(compositingOperation.rawValue),\(alpha))"
        if let color = backgroundColor {
            identifier.append("_\(color.hex)")
        }
        self.identifier = identifier
    }

    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.image(
                            withCompositingOperation: compositingOperation,
                            alpha: alpha,
                            backgroundColor: backgroundColor)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}
#endif

/// Represents a radius specified in a `RoundCornerImageProcessor`.
public enum Radius {
    /// The radius should be calculated as a fraction of the image width. Typically the associated value should be
    /// between 0 and 0.5, where 0 represents no radius and 0.5 represents using half of the image width.
    case widthFraction(CGFloat)
    /// The radius should be calculated as a fraction of the image height. Typically the associated value should be
    /// between 0 and 0.5, where 0 represents no radius and 0.5 represents using half of the image height.
    case heightFraction(CGFloat)
    /// Use a fixed point value as the round corner radius.
    case point(CGFloat)

    var radiusIdentifier: String {
        switch self {
        case .widthFraction(let f):
            return "w_frac_\(f)"
        case .heightFraction(let f):
            return "h_frac_\(f)"
        case .point(let p):
            return p.description
        }
    }
}

/// Processor for making round corner images. Only CG-based images are supported in macOS, 
/// if a non-CG image passed in, the processor will do nothing.
///
/// - Note: The input image will be rendered with round corner pixels removed. If the image itself does not contain
/// alpha channel (for example, a JPEG image), the processed image will contain an alpha channel in memory in order
/// to show correctly. However, when cached to disk, Kingfisher respects the original image format by default. That
/// means the alpha channel will be removed for these images. When you load the processed image from cache again, you
/// will lose transparent corner.
///
/// You could use `FormatIndicatedCacheSerializer.png` to force Kingfisher to serialize the image to PNG format in this
/// case.
///
public struct RoundCornerImageProcessor: ImageProcessor {

    /// Represents a radius specified in a `RoundCornerImageProcessor`.
    public typealias Radius = Kingfisher.Radius

    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String

    /// The radius will be applied in processing. Specify a certain point value with `.point`, or a fraction of the
    /// target image with `.widthFraction`. or `.heightFraction`. For example, given a square image with width and
    /// height equals,  `.widthFraction(0.5)` means use half of the length of size and makes the final image a round one.
    public let radius: Radius
    
    /// The target corners which will be applied rounding.
    public let roundingCorners: RectCorner
    
    /// Target size of output image should be. If `nil`, the image will keep its original size after processing.
    public let targetSize: CGSize?

    /// Background color of the output image. If `nil`, it will use a transparent background.
    public let backgroundColor: KFCrossPlatformColor?

    /// Creates a `RoundCornerImageProcessor`.
    ///
    /// - Parameters:
    ///   - cornerRadius: Corner radius in point will be applied in processing.
    ///   - targetSize: Target size of output image should be. If `nil`,
    ///                 the image will keep its original size after processing.
    ///                 Default is `nil`.
    ///   - corners: The target corners which will be applied rounding. Default is `.all`.
    ///   - backgroundColor: Background color to apply for the output image. Default is `nil`.
    ///
    /// - Note:
    ///
    /// This initializer accepts a concrete point value for `cornerRadius`. If you do not know the image size, but still
    /// want to apply a full round-corner (making the final image a round one), or specify the corner radius as a
    /// fraction of one dimension of the target image, use the `Radius` version instead.
    ///
    public init(
        cornerRadius: CGFloat,
        targetSize: CGSize? = nil,
        roundingCorners corners: RectCorner = .all,
        backgroundColor: KFCrossPlatformColor? = nil
    )
    {
        let radius = Radius.point(cornerRadius)
        self.init(radius: radius, targetSize: targetSize, roundingCorners: corners, backgroundColor: backgroundColor)
    }

    /// Creates a `RoundCornerImageProcessor`.
    ///
    /// - Parameters:
    ///   - radius: The radius will be applied in processing.
    ///   - targetSize: Target size of output image should be. If `nil`,
    ///                 the image will keep its original size after processing.
    ///                 Default is `nil`.
    ///   - corners: The target corners which will be applied rounding. Default is `.all`.
    ///   - backgroundColor: Background color to apply for the output image. Default is `nil`.
    public init(
        radius: Radius,
        targetSize: CGSize? = nil,
        roundingCorners corners: RectCorner = .all,
        backgroundColor: KFCrossPlatformColor? = nil
    )
    {
        self.radius = radius
        self.targetSize = targetSize
        self.roundingCorners = corners
        self.backgroundColor = backgroundColor

        self.identifier = {
            var identifier = ""

            if let size = targetSize {
                identifier = "com.onevcat.Kingfisher.RoundCornerImageProcessor" +
                             "(\(radius.radiusIdentifier)_\(size)\(corners.cornerIdentifier))"
            } else {
                identifier = "com.onevcat.Kingfisher.RoundCornerImageProcessor" +
                             "(\(radius.radiusIdentifier)\(corners.cornerIdentifier))"
            }
            if let backgroundColor = backgroundColor {
                identifier += "_\(backgroundColor)"
            }

            return identifier
        }()
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            let size = targetSize ?? image.kf.size

            let cornerRadius: CGFloat
            switch radius {
            case .point(let point):
                cornerRadius = point
            case .widthFraction(let widthFraction):
                cornerRadius = size.width * widthFraction
            case .heightFraction(let heightFraction):
                cornerRadius = size.height * heightFraction
            }

            return image.kf.scaled(to: options.scaleFactor)
                        .kf.image(
                            withRoundRadius: cornerRadius,
                            fit: size,
                            roundingCorners: roundingCorners,
                            backgroundColor: backgroundColor)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

public struct Border {
    public var color: KFCrossPlatformColor
    public var lineWidth: CGFloat
    
    /// The radius will be applied in processing. Specify a certain point value with `.point`, or a fraction of the
    /// target image with `.widthFraction`. or `.heightFraction`. For example, given a square image with width and
    /// height equals,  `.widthFraction(0.5)` means use half of the length of size and makes the final image a round one.
    public var radius: Radius
    
    /// The target corners which will be applied rounding.
    public var roundingCorners: RectCorner
    
    public init(
        color: KFCrossPlatformColor = .black,
        lineWidth: CGFloat = 4,
        radius: Radius = .point(0),
        roundingCorners: RectCorner = .all
    ) {
        self.color = color
        self.lineWidth = lineWidth
        self.radius = radius
        self.roundingCorners = roundingCorners
    }
    
    var identifier: String {
        "\(color.hex)_\(lineWidth)_\(radius.radiusIdentifier)_\(roundingCorners.cornerIdentifier)"
    }
}

public struct BorderImageProcessor: ImageProcessor {
    public var identifier: String { "com.onevcat.Kingfisher.RoundCornerImageProcessor(\(border)" }
    public let border: Border
    
    public init(border: Border) {
        self.border = border
    }
    
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.addingBorder(border)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Represents how a size adjusts itself to fit a target size.
///
/// - none: Not scale the content.
/// - aspectFit: Scales the content to fit the size of the view by maintaining the aspect ratio.
/// - aspectFill: Scales the content to fill the size of the view.
public enum ContentMode {
    /// Not scale the content.
    case none
    /// Scales the content to fit the size of the view by maintaining the aspect ratio.
    case aspectFit
    /// Scales the content to fill the size of the view.
    case aspectFill
}

/// Processor for resizing images.
/// If you need to resize a data represented image to a smaller size, use `DownsamplingImageProcessor`
/// instead, which is more efficient and uses less memory.
public struct ResizingImageProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// The reference size for resizing operation in point.
    public let referenceSize: CGSize
    
    /// Target content mode of output image should be.
    /// Default is `.none`.
    public let targetContentMode: ContentMode
    
    /// Creates a `ResizingImageProcessor`.
    ///
    /// - Parameters:
    ///   - referenceSize: The reference size for resizing operation in point.
    ///   - mode: Target content mode of output image should be.
    ///
    /// - Note:
    ///   The instance of `ResizingImageProcessor` will follow its `mode` property
    ///   and try to resizing the input images to fit or fill the `referenceSize`.
    ///   That means if you are using a `mode` besides of `.none`, you may get an
    ///   image with its size not be the same as the `referenceSize`.
    ///
    ///   **Example**: With input image size: {100, 200}, 
    ///   `referenceSize`: {100, 100}, `mode`: `.aspectFit`,
    ///   you will get an output image with size of {50, 100}, which "fit"s
    ///   the `referenceSize`.
    ///
    ///   If you need an output image exactly to be a specified size, append or use
    ///   a `CroppingImageProcessor`.
    public init(referenceSize: CGSize, mode: ContentMode = .none) {
        self.referenceSize = referenceSize
        self.targetContentMode = mode
        
        if mode == .none {
            self.identifier = "com.onevcat.Kingfisher.ResizingImageProcessor(\(referenceSize))"
        } else {
            self.identifier = "com.onevcat.Kingfisher.ResizingImageProcessor(\(referenceSize), \(mode))"
        }
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.resize(to: referenceSize, for: targetContentMode)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for adding blur effect to images. `Accelerate.framework` is used underhood for 
/// a better performance. A simulated Gaussian blur with specified blur radius will be applied.
public struct BlurImageProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Blur radius for the simulated Gaussian blur.
    public let blurRadius: CGFloat

    /// Creates a `BlurImageProcessor`
    ///
    /// - parameter blurRadius: Blur radius for the simulated Gaussian blur.
    public init(blurRadius: CGFloat) {
        self.blurRadius = blurRadius
        self.identifier = "com.onevcat.Kingfisher.BlurImageProcessor(\(blurRadius))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            let radius = blurRadius * options.scaleFactor
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.blurred(withRadius: radius)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for adding an overlay to images. Only CG-based images are supported in macOS.
public struct OverlayImageProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Overlay color will be used to overlay the input image.
    public let overlay: KFCrossPlatformColor
    
    /// Fraction will be used when overlay the color to image.
    public let fraction: CGFloat
    
    /// Creates an `OverlayImageProcessor`
    ///
    /// - parameter overlay:  Overlay color will be used to overlay the input image.
    /// - parameter fraction: Fraction will be used when overlay the color to image. 
    ///                       From 0.0 to 1.0. 0.0 means solid color, 1.0 means transparent overlay.
    public init(overlay: KFCrossPlatformColor, fraction: CGFloat = 0.5) {
        self.overlay = overlay
        self.fraction = fraction
        self.identifier = "com.onevcat.Kingfisher.OverlayImageProcessor(\(overlay.hex)_\(fraction))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.overlaying(with: overlay, fraction: fraction)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for tint images with color. Only CG-based images are supported.
public struct TintImageProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Tint color will be used to tint the input image.
    public let tint: KFCrossPlatformColor
    
    /// Creates a `TintImageProcessor`
    ///
    /// - parameter tint: Tint color will be used to tint the input image.
    public init(tint: KFCrossPlatformColor) {
        self.tint = tint
        self.identifier = "com.onevcat.Kingfisher.TintImageProcessor(\(tint.hex))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.tinted(with: tint)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for applying some color control to images. Only CG-based images are supported.
/// watchOS is not supported.
public struct ColorControlsProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Brightness changing to image.
    public let brightness: CGFloat
    
    /// Contrast changing to image.
    public let contrast: CGFloat
    
    /// Saturation changing to image.
    public let saturation: CGFloat
    
    /// InputEV changing to image.
    public let inputEV: CGFloat
    
    /// Creates a `ColorControlsProcessor`
    ///
    /// - Parameters:
    ///   - brightness: Brightness changing to image.
    ///   - contrast: Contrast changing to image.
    ///   - saturation: Saturation changing to image.
    ///   - inputEV: InputEV changing to image.
    public init(brightness: CGFloat, contrast: CGFloat, saturation: CGFloat, inputEV: CGFloat) {
        self.brightness = brightness
        self.contrast = contrast
        self.saturation = saturation
        self.inputEV = inputEV
        self.identifier = "com.onevcat.Kingfisher.ColorControlsProcessor(\(brightness)_\(contrast)_\(saturation)_\(inputEV))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.adjusted(brightness: brightness, contrast: contrast, saturation: saturation, inputEV: inputEV)
        case .data:
            return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for applying black and white effect to images. Only CG-based images are supported.
/// watchOS is not supported.
public struct BlackWhiteProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier = "com.onevcat.Kingfisher.BlackWhiteProcessor"
    
    /// Creates a `BlackWhiteProcessor`
    public init() {}
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        return ColorControlsProcessor(brightness: 0.0, contrast: 1.0, saturation: 0.0, inputEV: 0.7)
            .process(item: item, options: options)
    }
}

/// Processor for cropping an image. Only CG-based images are supported.
/// watchOS is not supported.
public struct CroppingImageProcessor: ImageProcessor {
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Target size of output image should be.
    public let size: CGSize
    
    /// Anchor point from which the output size should be calculate.
    /// The anchor point is consisted by two values between 0.0 and 1.0.
    /// It indicates a related point in current image. 
    /// See `CroppingImageProcessor.init(size:anchor:)` for more.
    public let anchor: CGPoint
    
    /// Creates a `CroppingImageProcessor`.
    ///
    /// - Parameters:
    ///   - size: Target size of output image should be.
    ///   - anchor: The anchor point from which the size should be calculated.
    ///             Default is `CGPoint(x: 0.5, y: 0.5)`, which means the center of input image.
    /// - Note:
    ///   The anchor point is consisted by two values between 0.0 and 1.0.
    ///   It indicates a related point in current image, eg: (0.0, 0.0) for top-left
    ///   corner, (0.5, 0.5) for center and (1.0, 1.0) for bottom-right corner.
    ///   The `size` property of `CroppingImageProcessor` will be used along with
    ///   `anchor` to calculate a target rectangle in the size of image.
    ///    
    ///   The target size will be automatically calculated with a reasonable behavior.
    ///   For example, when you have an image size of `CGSize(width: 100, height: 100)`,
    ///   and a target size of `CGSize(width: 20, height: 20)`: 
    ///   - with a (0.0, 0.0) anchor (top-left), the crop rect will be `{0, 0, 20, 20}`; 
    ///   - with a (0.5, 0.5) anchor (center), it will be `{40, 40, 20, 20}`
    ///   - while with a (1.0, 1.0) anchor (bottom-right), it will be `{80, 80, 20, 20}`
    public init(size: CGSize, anchor: CGPoint = CGPoint(x: 0.5, y: 0.5)) {
        self.size = size
        self.anchor = anchor
        self.identifier = "com.onevcat.Kingfisher.CroppingImageProcessor(\(size)_\(anchor))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            return image.kf.scaled(to: options.scaleFactor)
                        .kf.crop(to: size, anchorOn: anchor)
        case .data: return (DefaultImageProcessor.default |> self).process(item: item, options: options)
        }
    }
}

/// Processor for downsampling an image. Compared to `ResizingImageProcessor`, this processor
/// does not render the images to resize. Instead, it downsamples the input data directly to an
/// image. It is a more efficient than `ResizingImageProcessor`. Prefer to use `DownsamplingImageProcessor` as possible
/// as you can than the `ResizingImageProcessor`.
///
/// Only CG-based images are supported. Animated images (like GIF) is not supported.
public struct DownsamplingImageProcessor: ImageProcessor {
    
    /// Target size of output image should be. It should be smaller than the size of
    /// input image. If it is larger, the result image will be the same size of input
    /// data without downsampling.
    public let size: CGSize
    
    /// Identifier of the processor.
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public let identifier: String
    
    /// Creates a `DownsamplingImageProcessor`.
    ///
    /// - Parameter size: The target size of the downsample operation.
    public init(size: CGSize) {
        self.size = size
        self.identifier = "com.onevcat.Kingfisher.DownsamplingImageProcessor(\(size))"
    }
    
    /// Processes the input `ImageProcessItem` with this processor.
    ///
    /// - Parameters:
    ///   - item: Input item which will be processed by `self`.
    ///   - options: Options when processing the item.
    /// - Returns: The processed image.
    ///
    /// - Note: See documentation of `ImageProcessor` protocol for more.
    public func process(item: ImageProcessItem, options: KingfisherParsedOptionsInfo) -> KFCrossPlatformImage? {
        switch item {
        case .image(let image):
            guard let data = image.kf.data(format: .unknown) else {
                return nil
            }
            return KingfisherWrapper.downsampledImage(data: data, to: size, scale: options.scaleFactor)
        case .data(let data):
            return KingfisherWrapper.downsampledImage(data: data, to: size, scale: options.scaleFactor)
        }
    }
}

infix operator |>: AdditionPrecedence
public func |>(left: ImageProcessor, right: ImageProcessor) -> ImageProcessor {
    return left.append(another: right)
}

extension KFCrossPlatformColor {
    var hex: String {
        var r: CGFloat = 0
        var g: CGFloat = 0
        var b: CGFloat = 0
        var a: CGFloat = 0

        #if os(macOS)
        (usingColorSpace(.sRGB) ?? self).getRed(&r, green: &g, blue: &b, alpha: &a)
        #else
        getRed(&r, green: &g, blue: &b, alpha: &a)
        #endif

        let rInt = Int(r * 255) << 24
        let gInt = Int(g * 255) << 16
        let bInt = Int(b * 255) << 8
        let aInt = Int(a * 255)
        
        let rgba = rInt | gInt | bInt | aInt
        
        return String(format:"#%08x", rgba)
    }
}