feat(llm): improve logging and content processing

- Add 'keyword' to LLM processing log message for better context
- Replace simple string replace with regex for removing <think> tags in DeepseekR3 model
- The changes provide more precise logging and more robust content processing

src/basic/keywords/llm_keyword.rs

@@ -13,7 +13,7 @@ pub fn llm_keyword(state: Arc<AppState>, _user: UserSession, engine: &mut Engine
         .register_custom_syntax(&["LLM", "$expr$"], false, move |context, inputs| {
             let text = context.eval_expression_tree(&inputs[0])?.to_string();
 
-            info!("LLM processing text: {}", text);
+            info!("LLM keyword processing text: {}", text);
 
             let state_for_thread = Arc::clone(&state_clone);
             let prompt = build_llm_prompt(&text);

src/llm_models/deepseek_r3.rs

@@ -1,4 +1,5 @@
 use super::ModelHandler;
+use regex;
 
 pub struct DeepseekR3Handler;
 
@@ -8,7 +9,8 @@ impl ModelHandler for DeepseekR3Handler {
     }
 
     fn process_content(&self, content: &str) -> String {
-        content.replace("<think>", "").replace("</think>", "")
+        let re = regex::Regex::new(r"<think>.*?</think>").unwrap();
+        re.replace_all(content, "").to_string()
     }
 
     fn has_analysis_markers(&self, buffer: &str) -> bool {