burn-onnx 0.21.0-pre.3

use super::prelude::*;
use burn_store::TensorSnapshot;

impl NodeCodegen for onnx_ir::conv2d::Conv2dNode {
    fn inputs(&self) -> &[Argument] {
        &self.inputs
    }

    fn outputs(&self) -> &[Argument] {
        &self.outputs
    }

    fn field(&self) -> Option<Field> {
        let name = Ident::new(&self.name, Span::call_site());
        let weight_shape = self.inputs[1]
            .ty
            .static_shape_known()
            .expect("Conv2d: weight tensor shape must be known at codegen time");
        let groups = self.config.groups;
        let channels = [weight_shape[1] * groups, weight_shape[0]].to_tokens();
        let kernel_size = self.config.kernel_size.to_tokens();
        let stride = self.config.stride.to_tokens();
        let dilation = self.config.dilation.to_tokens();
        let groups = groups.to_tokens();
        let bias = self.inputs.len() == 3;

        let shape = self.inputs[0].ty.static_shape_known();
        let input_spatial = shape.as_deref().map(|s| &s[2..]);
        let padding = crate::burn::codegen::resolve_auto_pad_2d(
            &self.config.auto_pad,
            &self.config.padding,
            input_spatial,
            &self.config.kernel_size,
            &self.config.stride,
            &self.config.dilation,
        )
        .to_tokens();

        Some(Field::new(
            self.name.clone(),
            quote! {
                Conv2d<B>
            },
            quote! {
                let #name = Conv2dConfig::new(#channels, #kernel_size)
                    .with_stride(#stride)
                    .with_padding(#padding)
                    .with_dilation(#dilation)
                    .with_groups(#groups)
                    .with_bias(#bias)
                    .init(device);
            },
        ))
    }

    fn collect_snapshots(&self, field_name: &str) -> Vec<TensorSnapshot> {
        use crate::burn::node_traits::create_lazy_snapshot;

        let mut snapshots = vec![];

        // Weight tensor (input index 1)
        if let Some(weight_input) = self.inputs.get(1) {
            let weight_path = format!("{}.weight", field_name);
            if let Some(snapshot) = create_lazy_snapshot(weight_input, &weight_path, "Conv2d") {
                snapshots.push(snapshot);
            }
        }

        // Bias tensor (input index 2, optional)
        if self.inputs.len() > 2
            && let Some(bias_input) = self.inputs.get(2)
        {
            let bias_path = format!("{}.bias", field_name);
            if let Some(snapshot) = create_lazy_snapshot(bias_input, &bias_path, "Conv2d") {
                snapshots.push(snapshot);
            }
        }

        snapshots
    }

    fn forward(&self, scope: &mut ScopeAtPosition<'_>) -> TokenStream {
        let input = scope.arg(self.inputs.first().unwrap());
        let output = arg_to_ident(self.outputs.first().unwrap());
        let field = Ident::new(&self.name, Span::call_site());

        quote! {
            let #output = self.#field.forward(#input);
        }
    }

    fn register_imports(&self, imports: &mut BurnImports) {
        imports.register("burn::nn::PaddingConfig2d");
        imports.register("burn::nn::conv::Conv2d");
        imports.register("burn::nn::conv::Conv2dConfig");
    }
}

#[cfg(test)]
mod tests {
    use super::super::test_helpers::*;
    use burn::tensor::DType;
    use insta::assert_snapshot;
    use onnx_ir::conv2d::{Conv2dConfig, Conv2dNode, Conv2dNodeBuilder};
    use onnx_ir::padding::{AutoPad, PaddingConfig2d};

    fn create_conv2d_node(name: &str) -> Conv2dNode {
        let config = Conv2dConfig::new(
            [3, 3],
            [1, 1],
            PaddingConfig2d::Explicit(1, 1, 1, 1),
            [1, 1],
            1,
            AutoPad::NotSet,
        );

        Conv2dNodeBuilder::new(name)
            .input_tensor("input", 4, DType::F32)
            .input_static_tensor_shape("weight", vec![64, 3, 3, 3], DType::F32)
            .input_static_tensor_shape("bias", vec![64], DType::F32)
            .output_tensor("output", 4, DType::F32)
            .config(config)
            .build()
    }

    fn create_conv2d_node_asymmetric(name: &str) -> Conv2dNode {
        // Asymmetric padding: top=1, left=2, bottom=3, right=4
        let config = Conv2dConfig::new(
            [3, 3],
            [1, 1],
            PaddingConfig2d::Explicit(1, 2, 3, 4),
            [1, 1],
            1,
            AutoPad::NotSet,
        );

        Conv2dNodeBuilder::new(name)
            .input_tensor("input", 4, DType::F32)
            .input_static_tensor_shape("weight", vec![64, 3, 3, 3], DType::F32)
            .input_static_tensor_shape("bias", vec![64], DType::F32)
            .output_tensor("output", 4, DType::F32)
            .config(config)
            .build()
    }

    #[test]
    fn test_conv2d_forward() {
        let node = create_conv2d_node("conv1");
        let code = codegen_forward_default(&node);
        assert_snapshot!(code, @r"
        pub fn forward(&self, input: Tensor<B, 4>) -> Tensor<B, 4> {
            let output = self.conv1.forward(input);
            output
        }
        ");
    }

    #[test]
    fn test_conv2d_forward_with_clone() {
        let node = create_conv2d_node("conv1");
        let code = codegen_forward_with_clone(&node);
        assert_snapshot!(code, @r"
        pub fn forward(&self, input: Tensor<B, 4>) -> Tensor<B, 4> {
            let output = self.conv1.forward(input.clone());
            output
        }
        ");
    }

    fn create_conv2d_node_auto_pad(name: &str, auto_pad: AutoPad) -> Conv2dNode {
        let config = Conv2dConfig::new(
            [3, 3],
            [1, 1],
            PaddingConfig2d::Valid, // ignored when auto_pad is set
            [1, 1],
            1,
            auto_pad,
        );

        Conv2dNodeBuilder::new(name)
            .input_tensor_shape("input", vec![1, 3, 7, 7], DType::F32)
            .input_static_tensor_shape("weight", vec![64, 3, 3, 3], DType::F32)
            .input_static_tensor_shape("bias", vec![64], DType::F32)
            .output_tensor("output", 4, DType::F32)
            .config(config)
            .build()
    }

    #[test]
    fn test_conv2d_field_init_auto_pad_same_upper() {
        let node = create_conv2d_node_auto_pad("conv1", AutoPad::SameUpper);
        let code = codegen_field_init(&node);
        // auto_pad SAME_UPPER with input=7x7, kernel=3x3, stride=1 -> pad 1 each side
        assert_snapshot!(code, @r"
        let conv1 = Conv2dConfig::new([3, 64], [3, 3])
            .with_stride([1, 1])
            .with_padding(PaddingConfig2d::Explicit(1, 1, 1, 1))
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        ");
    }

    #[test]
    fn test_conv2d_field_init_auto_pad_valid() {
        let node = create_conv2d_node_auto_pad("conv1", AutoPad::Valid);
        let code = codegen_field_init(&node);
        assert_snapshot!(code, @r"
        let conv1 = Conv2dConfig::new([3, 64], [3, 3])
            .with_stride([1, 1])
            .with_padding(PaddingConfig2d::Valid)
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        ");
    }

    #[test]
    fn test_conv2d_field_init_asymmetric_padding() {
        let node = create_conv2d_node_asymmetric("conv1");
        let code = codegen_field_init(&node);
        // Asymmetric padding is passed directly to the module
        assert_snapshot!(code, @r"
        let conv1 = Conv2dConfig::new([3, 64], [3, 3])
            .with_stride([1, 1])
            .with_padding(PaddingConfig2d::Explicit(1, 2, 3, 4))
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        ");
    }
}