feat(tms_service): 添加html_scraper模块替代openai解析链接

使用html_scraper模块通过CSS选择器直接解析HTML获取链接和token，替代之前依赖openai的方式

src-tauri/Cargo.lock

@@ -1593,6 +1593,15 @@ dependencies = [
  "match_token",
 ]
 
+[[package]]
+name = "html_scraper"
+version = "0.1.0"
+dependencies = [
+ "regex",
+ "scraper",
+ "thiserror 1.0.69",
+]
+
 [[package]]
 name = "http"
 version = "1.3.1"
@@ -4458,6 +4467,7 @@ checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20"
 name = "tms_service"
 version = "0.1.0"
 dependencies = [
+ "html_scraper",
  "openai",
  "reqwest",
  "serde",

src-tauri/html_scraper/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "html_scraper"
+version = "0.1.0"
+edition = "2024"
+
+[dependencies]
+scraper = "0.19.1"
+thiserror = "1.0"
+regex = "1.10.5"

src-tauri/html_scraper/src/lib.rs

@@ -0,0 +1,73 @@
+use scraper::{Html, Selector};
+use thiserror::Error;
+use regex::Regex;
+
+/// 定义库可能返回的错误类型
+#[derive(Debug, Error)]
+pub enum ScrapeError {
+    /// 当内部使用的CSS选择器无效时返回。
+    #[error("无效的CSS选择器: {0}")]
+    InvalidSelector(String),
+    /// 当正则表达式编译失败时返回。
+    #[error("无效的正则表达式: {0}")]
+    InvalidRegex(#[from] regex::Error),
+}
+
+/// 通用函数：根据CSS选择器和链接文本查找链接。
+fn find_link_by_text(html_body: &str, selector_str: &str, link_text: &str) -> Result<Option<String>, ScrapeError> {
+    let selector = Selector::parse(selector_str)
+        .map_err(|e| ScrapeError::InvalidSelector(format!("'{selector_str}': {e}")))?;
+
+    let document = Html::parse_document(html_body);
+
+    for element in document.select(&selector) {
+        if element.text().any(|text| text.trim() == link_text) {
+            if let Some(link) = element.value().attr("href") {
+                return Ok(Some(link.to_string()));
+            }
+        }
+    }
+
+    Ok(None)
+}
+
+/// 解析HTML文本，查找并返回第一个匹配“编辑”按钮的链接。
+pub fn find_edit_link(html_body: &str) -> Result<Option<String>, ScrapeError> {
+    find_link_by_text(html_body, "a.btn.btn-info.btn-xs.m-bot5", "编辑")
+}
+
+/// 解析HTML文本，查找并返回“TMS配置”按钮的链接。
+pub fn find_tms_config_link(html_body: &str) -> Result<Option<String>, ScrapeError> {
+    find_link_by_text(html_body, "a.btn.btn-default.not-cinema", "TMS配置")
+}
+
+/// # 参数
+/// * `html_body`: 一个字符串切片，包含要解析的HTML内容。
+pub fn find_token(html_body: &str) -> Result<Option<String>, ScrapeError> {
+    // 使用属性选择器直接、高效地定位目标元素
+    let selector_str = r#"input[name="token"]"#;
+    let token_selector = Selector::parse(selector_str)
+        .map_err(|e| ScrapeError::InvalidSelector(format!("'{selector_str}': {e}")))?;
+
+    let document = Html::parse_document(html_body);
+
+    // 查找第一个匹配的元素并提取其 "value" 属性
+    if let Some(element) = document.select(&token_selector).next() {
+        if let Some(token_value) = element.value().attr("value") {
+            return Ok(Some(token_value.to_string()));
+        }
+    }
+
+    Ok(None)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn it_works() {
+        let result = add(2, 2);
+        assert_eq!(result, 4);
+    }
+}

src-tauri/tms_service/Cargo.toml

@@ -9,3 +9,4 @@ tokio = { version = "1", features = ["full"] }
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 openai = { path = "../openai" }
+html_scraper = { path = "../html_scraper" }

src-tauri/tms_service/src/lib.rs

@@ -40,8 +40,11 @@ pub async fn create_ticket(cookie: &str, n2p: &str, massageQ: &str, wx: Option<&
 
     println!("源码获取完成");
 
-    let system_prompt = "告诉我编辑按钮的链接，只要告诉我链接的url就可以，不要说其他的任何内容";
-    let edit_page_url = openai::ask_openai(&body, system_prompt).await?;
+    let edit_page_url = match html_scraper::find_edit_link(&body) {
+        Ok(Some(link)) => link,
+        Ok(None) => return Ok("没有找到编辑链接".to_string()),
+        Err(e) => return Ok(format!("查找编辑链接时出错: {}", e)),
+    };
     println!("平台id{}", edit_page_url);
 
     // 3. Visit the edit page to get its source
@@ -50,8 +53,11 @@ pub async fn create_ticket(cookie: &str, n2p: &str, massageQ: &str, wx: Option<&
     let body2 = String::from_utf8_lossy(&body2_bytes).to_string();
 
     // 4. Get the TMS config URL
-    let system_prompt2 = "告诉我TMS配置按钮的链接，只要告诉我链接的url就可以，不要说其他的任何内容";
-    let tms_config_url = openai::ask_openai(&body2, system_prompt2).await?;
+    let tms_config_url = match html_scraper::find_tms_config_link(&body2) {
+        Ok(Some(link)) => link,
+        Ok(None) => return Ok("没有找到TMS配置链接".to_string()),
+        Err(e) => return Ok(format!("查找TMS配置链接时出错: {}", e)),
+    };
     println!("tms配置链接{}", tms_config_url);
     
     // 5. Visit the TMS config URL to get the token
@@ -59,8 +65,11 @@ pub async fn create_ticket(cookie: &str, n2p: &str, massageQ: &str, wx: Option<&
     let body3_bytes = res3.bytes().await?;
     let body3 = String::from_utf8_lossy(&body3_bytes).to_string();
 
-    let system_prompt3 = "告诉我权限认证(token)的值，只要告诉我对应的值就行，不要说其他的任何内容";
-    let token = openai::ask_openai(&body3, system_prompt3).await?;
+    let token = match html_scraper::find_token(&body3) {
+        Ok(Some(t)) => t,
+        Ok(None) => return Ok("没有找到token".to_string()),
+        Err(e) => return Ok(format!("查找token时出错: {}", e)),
+    };
     println!("token={}", token);
 
     // 6. Visit the final URL to get the session cookies