Function c_encode_vector_block 

pub fn c_encode_vector_block(
    vec: &[f32],
    codebook: &[f32],
    codebook_size: usize,
    block_dim: usize,
) -> Vec<u16>

Encode a vector block: find nearest codeword for each block of block_dim floats. Returns the codeword indices.

Safety

The caller must ensure:

• vec.len() == dim and dim % block_dim == 0
• codebook.len() == codebook_size * block_dim
• out_indices has capacity dim / block_dim