import torch
import torch.nn as nn

class Model(nn.Module):
    """
    Performs a causal 1D convolution: left-padded so each output only depends on current and past inputs.
    Used in autoregressive sequence models (Mamba, Hyena, WaveNet, TCN) for causal sequence processing.

    Args:
        in_channels (int): Number of input channels.
        out_channels (int): Number of output channels.
        kernel_size (int): Size of the convolution kernel.
        dilation (int, optional): Dilation rate. Defaults to 1.
        bias (bool, optional): If True, adds a learnable bias. Defaults to False.
    """
    def __init__(self, in_channels: int, out_channels: int, kernel_size: int, dilation: int = 1, bias: bool = False):
        super(Model, self).__init__()
        self.causal_padding = (kernel_size - 1) * dilation
        self.conv1d = nn.Conv1d(in_channels, out_channels, kernel_size, dilation=dilation, bias=bias)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        x = torch.nn.functional.pad(x, (self.causal_padding, 0))
        return self.conv1d(x)

batch_size = 32
in_channels = 256
out_channels = 256
kernel_size = 4
dilation = 1
seq_len = 8192

def get_inputs():
    return [torch.randn(popcorn_pri.jitter_int(batch_size), in_channels, popcorn_pri.jitter_int(seq_len))]

def get_init_inputs():
    return [in_channels, out_channels, kernel_size, dilation]

import torch
import torch.nn as nn
from torch.utils.cpp_extension import load_inline

cuda_source = '''
#include <torch/extension.h>
#include <cuda.h>
#include <cuda_runtime.h>

__global__ void causal_conv1d_kernel(
    const float* __restrict__ input,
    const float* __restrict__ weight,
    float* __restrict__ output,
    int batch_size,
    int in_channels,
    int out_channels,
    int seq_len,
    int kernel_size,
    int dilation
) {
    int idx = blockIdx.x * blockDim.x + threadIdx.x;
    int total = batch_size * out_channels * seq_len;
    if (idx >= total) return;

    int s = idx % seq_len;
    int oc = (idx / seq_len) % out_channels;
    int b = idx / (seq_len * out_channels);

    double sum = 0.0;
    const int pad = (kernel_size - 1) * dilation;
    for (int k = 0; k < kernel_size; ++k) {
        int in_pos = s + k * dilation - pad;
        if (in_pos >= 0 && in_pos < seq_len) {
            for (int ic = 0; ic < in_channels; ++ic) {
                double in_val = input[(b * in_channels + ic) * seq_len + in_pos];
                double w_val = weight[(oc * in_channels + ic) * kernel_size + k];
                sum += in_val * w_val;
            }
        }
    }
    output[(b * out_channels + oc) * seq_len + s] = (float)sum;
}

torch::Tensor causal_conv1d_cuda(torch::Tensor input, torch::Tensor weight, int kernel_size, int dilation) {
    auto batch_size = input.size(0);
    auto in_channels = input.size(1);
    auto seq_len = input.size(2);
    auto out_channels = weight.size(0);

    auto output = torch::zeros({batch_size, out_channels, seq_len}, input.options());

    int total = batch_size * out_channels * seq_len;
    int threads = 512;
    int blocks = (total + threads - 1) / threads;

    causal_conv1d_kernel<<<blocks, threads>>>(
        input.data_ptr<float>(),
        weight.data_ptr<float>(),
        output.data_ptr<float>(),
        batch_size,
        in_channels,
        out_channels,
        seq_len,
        kernel_size,
        dilation
    );

    return output;
}
'''

cpp_source = '''
torch::Tensor causal_conv1d_cuda(torch::Tensor input, torch::Tensor weight, int kernel_size, int dilation);
'''

module = load_inline(
    name='causal_conv1d',
    cpp_sources=[cpp_source],
    cuda_sources=[cuda_source],
    functions=['causal_conv1d_cuda'],
    verbose=False
)

class ModelNew(nn.Module):
    def __init__(self, in_channels, out_channels, kernel_size, dilation=1, bias=False):
        super(ModelNew, self).__init__()
        self.conv1d = nn.Conv1d(in_channels, out_channels, kernel_size, dilation=dilation, bias=bias)
        self.kernel_size = kernel_size;
        self.dilation = dilation;

    def forward(self, x):
        weight = self.conv1d.weight
        return module.causal_conv1d_cuda(x, weight, self.kernel_size, self.dilation)