From 45e4a5e735826953be15c0a67d33f6bc8af48df0 Mon Sep 17 00:00:00 2001
From: "Rodrigo Rodriguez (Pragmatismo)" <me@rodrigorodriguez.com>
Date: Wed, 5 Nov 2025 12:46:08 -0300
Subject: [PATCH] feat: improve prompt formatting and system metrics

- Update prompt formatting in BotOrchestrator to use clearer labels (SYSTEM/CONTEXT) with emphasis markers
- Remove unused token_ratio field from SystemMetrics struct
- Increase default context size (2048->4096) and prediction length (512->1024) in config
- Clean up metrics calculation by removing redundant token ratio computation

The changes improve readability of system prompts and simplify metrics collection while increasing default model capacity.
---
 src/bot/mod.rs                                 | 4 ++--
 src/nvidia/mod.rs                              | 8 --------
 templates/default.gbai/default.gbot/config.csv | 4 ++--
 3 files changed, 4 insertions(+), 12 deletions(-)

diff --git a/src/bot/mod.rs b/src/bot/mod.rs
index 4372872c..36f48d40 100644
--- a/src/bot/mod.rs
+++ b/src/bot/mod.rs
@@ -393,10 +393,10 @@ impl BotOrchestrator {
 
         let mut prompt = String::new();
         if !system_prompt.is_empty() {
-            prompt.push_str(&format!("AI:{}\n", system_prompt));
+            prompt.push_str(&format!("SYSTEM: *** {} *** \n", system_prompt));
         }
         if !context_data.is_empty() {
-            prompt.push_str(&format!("CTX:{}\n", context_data));
+            prompt.push_str(&format!("CONTEXT: *** {} *** \n", context_data));
         }
         for (role, content) in &history {
             prompt.push_str(&format!("{}:{}\n", role, content));
diff --git a/src/nvidia/mod.rs b/src/nvidia/mod.rs
index 6525b79e..9cfc5e7a 100644
--- a/src/nvidia/mod.rs
+++ b/src/nvidia/mod.rs
@@ -7,7 +7,6 @@ use sysinfo::{System};
 pub struct SystemMetrics {
     pub gpu_usage: Option<f32>,
     pub cpu_usage: f32,
-    pub token_ratio: f32,
 }
 
 /// Gets current system metrics
@@ -25,17 +24,10 @@ pub fn get_system_metrics(current_tokens: usize, max_tokens: usize) -> Result<Sy
         None
     };
 
-    // Calculate token ratio
-    let token_ratio = if max_tokens > 0 {
-        current_tokens as f32 / max_tokens as f32 * 100.0
-    } else {
-        0.0
-    };
 
     Ok(SystemMetrics {
         gpu_usage,
         cpu_usage,
-        token_ratio,
     })
 }
 
diff --git a/templates/default.gbai/default.gbot/config.csv b/templates/default.gbai/default.gbot/config.csv
index 44a51a68..67abe9bd 100644
--- a/templates/default.gbai/default.gbot/config.csv
+++ b/templates/default.gbai/default.gbot/config.csv
@@ -19,8 +19,8 @@ llm-server-host,0.0.0.0
 llm-server-port,8081
 llm-server-gpu-layers,0
 llm-server-n-moe,0
-llm-server-ctx-size,2048
-llm-server-n-predict,512
+llm-server-ctx-size,4096
+llm-server-n-predict,1024
 llm-server-parallel,6
 llm-server-cont-batching,true
 llm-server-mlock,false