rustyml 0.13.0

//! 2D max pooling layer that selects the maximum value within each window across height and width

use crate::error::Error;
use crate::neural_network::Tensor;
use crate::neural_network::layers::TrainingParameters;
use crate::neural_network::layers::convolution::PaddingType;
use crate::neural_network::layers::layer_weight::LayerWeight;
use crate::neural_network::layers::pooling::layer_functions_2d_pooling;
use crate::neural_network::layers::pooling::pooling_engine::{
    PoolKind, windowed_pool_backward, windowed_pool_forward,
};
use crate::neural_network::layers::pooling::validation::{
    validate_all_dims_positive, validate_input_shape_dims, validate_pool_size_2d,
    validate_strides_2d,
};
use crate::neural_network::layers::shape_helpers::calculate_output_shape_2d_pooling;
use crate::neural_network::traits::Layer;

/// 2D max pooling layer
///
/// Selects the maximum value within each pooling window across height and width
/// Input tensor shape: `[batch_size, channels, height, width]`. Output tensor shape:
/// `[batch_size, channels, pooled_height, pooled_width]` where
/// `pooled_height = (height - pool_size_h) / stride_h + 1` and
/// `pooled_width = (width - pool_size_w) / stride_w + 1`
///
/// # Examples
///
/// ```rust
/// use rustyml::neural_network::sequential::Sequential;
/// use rustyml::neural_network::layers::*;
/// use rustyml::neural_network::optimizers::*;
/// use rustyml::neural_network::losses::*;
/// use ndarray::Array4;
///
/// // 4D input tensor: [batch_size, channels, height, width]
/// // batch size 2, 3 input channels, 6x6 pixels
/// let mut input_data = Array4::zeros((2, 3, 6, 6));
///
/// // Set specific values so the max pooling result is predictable
/// for b in 0..2 {
///     for c in 0..3 {
///         for i in 0..6 {
///             for j in 0..6 {
///                 // Easily observable pattern
///                 input_data[[b, c, i, j]] = (i * j) as f32 + b as f32 * 0.1 + c as f32 * 0.01;
///             }
///         }
///     }
/// }
///
/// let x = input_data.clone().into_dyn();
///
/// let mut model = Sequential::new();
/// model
///     // strides default to pool_size (2, 2) and padding defaults to Valid
///     .add(MaxPooling2D::new((2, 2), vec![2, 3, 6, 6]).unwrap())
///     .compile(RMSprop::new(0.001, 0.9, 1e-8, 0.0).unwrap(), MeanSquaredError::new());
///
/// // Target tensor matching the pooled shape
/// let y = Array4::ones((2, 3, 3, 3)).into_dyn();
///
/// model.summary();
///
/// model.fit(&x, &y, 3).unwrap();
///
/// let prediction = model.predict(&x).unwrap();
/// println!("MaxPooling2D prediction results: {:?}", prediction);
///
/// assert_eq!(prediction.shape(), &[2, 3, 3, 3]);
/// ```
///
/// # Performance
///
/// Parallel execution is used when `batch_size * channels >= 32`
#[derive(Debug)]
pub struct MaxPooling2D {
    /// Size of the pooling window as (height, width)
    pool_size: (usize, usize),
    /// Step size of the pooling operation as (vertical stride, horizontal stride)
    strides: (usize, usize),
    /// Shape of the input tensor declared at construction time
    input_shape: Vec<usize>,
    /// Padding mode applied around the input before pooling
    padding: PaddingType,
    /// Shape of the most recent forward input, cached for backpropagation
    forward_input_shape: Option<Vec<usize>>,
    /// Cached flat per-output arg-max indices used for backpropagation
    argmax: Option<Vec<usize>>,
}

impl MaxPooling2D {
    /// Creates a new 2D max pooling layer
    ///
    /// # Parameters
    ///
    /// - `pool_size` - Size of the pooling window as (height, width)
    /// - `input_shape` - Input tensor shape `[batch_size, channels, height, width]`
    ///
    /// # Notes
    ///
    /// Strides default to `pool_size` and padding defaults to [`PaddingType::Valid`]. Override them
    /// with [`MaxPooling2D::with_strides`] and [`MaxPooling2D::with_padding`]
    ///
    /// # Returns
    ///
    /// - `Result<MaxPooling2D, Error>` - New layer instance on success
    ///
    /// # Errors
    ///
    /// - `Error::InvalidInput` - If `input_shape` is not 4D
    /// - `Error::InvalidParameter` - If `pool_size` has a zero dimension
    pub fn new(pool_size: (usize, usize), input_shape: Vec<usize>) -> Result<Self, Error> {
        validate_input_shape_dims(&input_shape, 4, "MaxPooling2D")?;
        validate_all_dims_positive(&input_shape)?;
        validate_pool_size_2d(pool_size, input_shape[2], input_shape[3])?;

        Ok(MaxPooling2D {
            pool_size,
            strides: pool_size,
            input_shape,
            padding: PaddingType::Valid,
            forward_input_shape: None,
            argmax: None,
        })
    }

    /// Sets the pooling strides (defaults to `pool_size`)
    ///
    /// # Parameters
    ///
    /// - `strides` - Strides of the pooling operation as (vertical stride, horizontal stride)
    ///
    /// # Returns
    ///
    /// - `Result<Self, Error>` - The updated layer, or an error if any stride is zero
    pub fn with_strides(mut self, strides: (usize, usize)) -> Result<Self, Error> {
        validate_strides_2d(strides)?;
        self.strides = strides;
        Ok(self)
    }

    /// Sets the padding mode (defaults to [`PaddingType::Valid`])
    ///
    /// # Parameters
    ///
    /// - `padding` - `Valid` (no padding) or `Same` (pad so the output covers the input; padded
    ///   cells are excluded from each window)
    ///
    /// # Returns
    ///
    /// - `Self` - The updated layer
    pub fn with_padding(mut self, padding: PaddingType) -> Self {
        self.padding = padding;
        self
    }
}

impl Layer for MaxPooling2D {
    fn forward(&mut self, input: &Tensor) -> Result<Tensor, Error> {
        if input.ndim() != 4 {
            return Err(Error::invalid_input("input tensor is not 4D"));
        }

        // Cache the input shape and arg-max positions for the backward pass
        self.forward_input_shape = Some(input.shape().to_vec());

        let (output, argmax) = windowed_pool_forward(
            input,
            &[self.pool_size.0, self.pool_size.1],
            &[self.strides.0, self.strides.1],
            PoolKind::Max,
            self.padding,
        );
        self.argmax = argmax;
        Ok(output)
    }

    /// Inference forward (eval mode, writes no caches). See [`Layer::predict`]
    fn predict(&self, input: &Tensor) -> Result<Tensor, Error> {
        if input.ndim() != 4 {
            return Err(Error::invalid_input("input tensor is not 4D"));
        }

        let (output, _) = windowed_pool_forward(
            input,
            &[self.pool_size.0, self.pool_size.1],
            &[self.strides.0, self.strides.1],
            PoolKind::Max,
            self.padding,
        );
        Ok(output)
    }

    fn backward(&mut self, grad_output: &Tensor) -> Result<Tensor, Error> {
        let input_shape = self
            .forward_input_shape
            .as_ref()
            .ok_or_else(|| Error::forward_pass_not_run("MaxPooling2D"))?;
        let argmax = self
            .argmax
            .as_ref()
            .ok_or_else(|| Error::forward_pass_not_run("MaxPooling2D"))?;

        Ok(windowed_pool_backward(
            grad_output,
            input_shape,
            &[self.pool_size.0, self.pool_size.1],
            &[self.strides.0, self.strides.1],
            PoolKind::Max,
            Some(argmax),
            self.padding,
        ))
    }

    fn layer_type(&self) -> &str {
        "MaxPooling2D"
    }

    layer_functions_2d_pooling!();
}