fix: merge logUsage + saveUsageStats to prevent duplicate usage stats for streaming requests

- Enhanced saveUsageStats to include cache tokens (cache_read, cache_creation), reasoning tokens, and estimated flag
- Replaced logUsage calls in stream.js with saveUsageStats (2 locations)
- Removed duplicate saveUsageStats call from streamingHandler.js onStreamComplete
- Removed logUsage function and unused imports from usageTracking.js
- Each streaming request now writes exactly 1 usage record instead of 2

open-sse/handlers/chatCore/requestDetail.js

@@ -80,14 +80,29 @@ export function saveUsageStats({ provider, model, tokens, connectionId, apiKey,
 
   if (inTokens === 0 && outTokens === 0) return;
 
+  // Extract cache/reasoning tokens (unified from different formats)
+  const cacheRead = tokens.cache_read_input_tokens || tokens.cached_tokens || tokens.prompt_tokens_details?.cached_tokens || 0;
+  const cacheCreation = tokens.cache_creation_input_tokens || 0;
+  const reasoning = tokens.reasoning_tokens || 0;
+
   const time = new Date().toLocaleTimeString("en-US", { hour12: false, hour: "2-digit", minute: "2-digit", second: "2-digit" });
   const accountSuffix = connectionId ? ` | account=${connectionId.slice(0, 8)}...` : "";
-  console.log(`${COLORS.green}[${time}] 📊 [${label}] ${provider.toUpperCase()} | in=${inTokens} | out=${outTokens}${accountSuffix}${COLORS.reset}`);
+  
+  let msg = `${COLORS.green}[${time}] 📊 [${label}] ${provider?.toUpperCase() || "UNKNOWN"} | in=${inTokens} | out=${outTokens}${accountSuffix}`;
+  if (tokens.estimated) msg += ` ${COLORS.yellow}(estimated)${COLORS.reset}`;
+  if (cacheRead) msg += ` | cache_read=${cacheRead}`;
+  if (cacheCreation) msg += ` | cache_create=${cacheCreation}`;
+  if (reasoning) msg += ` | reasoning=${reasoning}`;
+  msg += `${COLORS.reset}`;
+  console.log(msg);
 
-  // Normalize to OpenAI token shape for storage
+  // Normalize to OpenAI token shape for storage (include all token types)
   const normalized = {
     prompt_tokens: tokens.prompt_tokens ?? tokens.input_tokens ?? 0,
-    completion_tokens: tokens.completion_tokens ?? tokens.output_tokens ?? 0
+    completion_tokens: tokens.completion_tokens ?? tokens.output_tokens ?? 0,
+    cache_read_input_tokens: cacheRead,
+    cache_creation_input_tokens: cacheCreation,
+    reasoning_tokens: reasoning,
   };
 
   saveRequestUsage({