fix: Clean up Rust warnings and add Python vital signs detection

Rust changes:
- Fix unused variable warnings in wifi-densepose-nn (densepose.rs, inference.rs, tensor.rs, translator.rs)
- Remove unused imports in wifi-densepose-mat (breathing.rs, pipeline.rs, csi_receiver.rs, debris_model.rs, vital_signs_classifier.rs)
- All tests continue to pass

Python changes:
- Add vital_signs.py module with BreathingDetector and HeartbeatDetector classes
- Mirror Rust wifi-densepose-mat detection functionality
- Update v1 package version to 1.2.0
- Export new vital signs classes from core module

rust-port/wifi-densepose-rs/crates/wifi-densepose-nn/src/densepose.rs

@@ -252,7 +252,7 @@ impl DensePoseHead {
         })?;
 
         let input_arr = input.as_array4()?;
-        let (batch, _channels, height, width) = input_arr.dim();
+        let (_batch, _channels, _height, _width) = input_arr.dim();
 
         // Apply shared convolutions
         let mut current = input_arr.clone();

rust-port/wifi-densepose-rs/crates/wifi-densepose-nn/src/inference.rs

@@ -206,7 +206,7 @@ impl Backend for MockBackend {
         self.output_shapes.get(name).cloned()
     }
 
-    fn run(&self, inputs: HashMap<String, Tensor>) -> NnResult<HashMap<String, Tensor>> {
+    fn run(&self, _inputs: HashMap<String, Tensor>) -> NnResult<HashMap<String, Tensor>> {
         let mut outputs = HashMap::new();
 
         for (name, shape) in &self.output_shapes {

rust-port/wifi-densepose-rs/crates/wifi-densepose-nn/src/tensor.rs

@@ -266,7 +266,7 @@ impl Tensor {
     }
 
     /// Apply softmax along axis
-    pub fn softmax(&self, axis: usize) -> NnResult<Tensor> {
+    pub fn softmax(&self, _axis: usize) -> NnResult<Tensor> {
         match self {
             Tensor::Float4D(a) => {
                 let max = a.fold(f32::NEG_INFINITY, |acc, &x| acc.max(x));

rust-port/wifi-densepose-rs/crates/wifi-densepose-nn/src/translator.rs

@@ -342,7 +342,7 @@ impl ModalityTranslator {
         })?;
 
         let input_arr = input.as_array4()?;
-        let (batch, _channels, height, width) = input_arr.dim();
+        let (_batch, _channels, _height, _width) = input_arr.dim();
 
         // Encode
         let mut encoder_outputs = Vec::new();
@@ -461,7 +461,7 @@ impl ModalityTranslator {
         weights: &ConvBlockWeights,
     ) -> NnResult<Array4<f32>> {
         let (batch, in_channels, in_height, in_width) = input.dim();
-        let (out_channels, _, kernel_h, kernel_w) = weights.conv_weight.dim();
+        let (out_channels, _, _kernel_h, _kernel_w) = weights.conv_weight.dim();
 
         // Upsample 2x
         let out_height = in_height * 2;
@@ -536,7 +536,7 @@ impl ModalityTranslator {
     fn apply_attention(
         &self,
         input: &Array4<f32>,
-        weights: &AttentionWeights,
+        _weights: &AttentionWeights,
     ) -> NnResult<(Array4<f32>, Array4<f32>)> {
         let (batch, channels, height, width) = input.dim();
         let seq_len = height * width;