📝 docs(CHANGELOG.md): update changelog with v0.4.2 release details

🐛 fix(auth.rs): prevent redundant retrieval of pushPageId and ssid on verification code error
♻️ refactor(classifier.rs): simplify grayscale conversion and mean_std calculation for better readability and performance

CHANGELOG.md

@@ -1,3 +1,7 @@
+## [0.4.2] - 2024-05-16
+
+- 验证码错误时不再重复获取 pushPageId 和 ssid
+
 ## [0.4.1] - 2024-05-12
 
 - 将 `image` 替换成 `resize` 和 `zune-jpeg` 以减小包体积
@@ -26,6 +30,7 @@
 - Initial release
 - 在 Windows 和 Linux 上通过测试
 
+[0.4.2]: https://github.com/ShanghaitechGeekPie/net-loginer/releases/tag/v0.4.2
 [0.4.1]: https://github.com/ShanghaitechGeekPie/net-loginer/releases/tag/v0.4.1
 [0.4.0]: https://github.com/ShanghaitechGeekPie/net-loginer/releases/tag/v0.4.0
 [0.3.1]: https://github.com/ShanghaitechGeekPie/net-loginer/releases/tag/v0.3.1

src/auth.rs

@@ -175,8 +175,9 @@ impl Authenticator {
     }
 
     fn login_for_ip(&self, ip_address: &Ipv4Addr) -> Result<()> {
+        let (push_page_id, ssid) = self.get_page_params(*ip_address)?;
+
         loop {
-            let (push_page_id, ssid) = self.get_page_params(*ip_address)?;
             let verify_code = self.get_verify_code(*ip_address)?;
 
             let json_value = self

src/classifier.rs

@@ -78,33 +78,31 @@ impl Classifier {
         let (image, width, height) = self.resize_image(image)?;
 
         let image_bytes = match self.channels {
-            ModelChannels::Gray => {
-                let mut gray_image = vec![0; image.len() / 3];
-                for (i, pixels) in image.chunks(3).enumerate() {
-                    let gray = 0.2989 * pixels[0] as f32
+            ModelChannels::Gray => image
+                .chunks(3)
+                .map(|pixels| {
+                    (0.2989 * pixels[0] as f32
                         + 0.5870 * pixels[1] as f32
-                        + 0.1140 * pixels[2] as f32;
-                    gray_image[i] = gray as u8;
-                }
-                gray_image
-            }
+                        + 0.1140 * pixels[2] as f32) as u8
+                })
+                .collect::<Vec<_>>(),
             ModelChannels::RGB => image,
         };
 
+        let mean_std = match self.channels {
+            ModelChannels::Gray => vec![(0.456f32, 0.224f32)],
+            ModelChannels::RGB => vec![
+                (0.485f32, 0.229f32),
+                (0.456f32, 0.224f32),
+                (0.406f32, 0.225f32),
+            ],
+        };
+
         let tensor = Array::from_shape_fn(
             (1, self.channels as usize, height, width),
             |(_, c, i, j)| {
                 let now = image_bytes[(i * width + j) * self.channels as usize + c] as f32;
-                let (mean, std) = if self.channels == ModelChannels::Gray {
-                    (0.456f32, 0.224f32)
-                } else {
-                    match c {
-                        0 => (0.485f32, 0.229f32),
-                        1 => (0.456f32, 0.224f32),
-                        2 => (0.406f32, 0.225f32),
-                        _ => unreachable!(),
-                    }
-                };
+                let (mean, std) = mean_std[c];
                 ((now / 255f32) - mean) / std
             },
         );