no test connection

src/lib/services/textGenerationClient.ts

@@ -29,84 +29,20 @@ class TextGenerationManager {
   }
 
   /**
-   * Test connection and determine which client to use
+   * Initialize without testing - assume Qwen3 is available and test on first real use
    */
   async initialize(): Promise<void> {
     if (this.connectionTested) return;
 
-    console.log('🔍 Testing Qwen3 connection with improved verification...');
+    console.log('🔧 Initializing text generation manager - using Qwen3 but will fallback to Zephyr-7B on failure');
     
-    try {
-      if (this.primaryClient.testConnection) {
-        // Add timeout wrapper for connection test
-        const connectionTestPromise = this.primaryClient.testConnection();
-        const timeoutPromise = new Promise<boolean>((_, reject) => {
-          setTimeout(() => reject(new Error('Connection test timeout')), 15000); // 15 second timeout
-        });
-
-        const qwen3Available = await Promise.race([connectionTestPromise, timeoutPromise]);
-        
-        if (qwen3Available) {
-          console.log('✅ Qwen3 client is available and will be used for text generation');
-          this.useQwen3 = true;
-          
-          // Perform a secondary validation test to ensure real functionality
-          await this.performSecondaryValidation();
-        } else {
-          console.log('⚠️ Qwen3 client is not available, falling back to Zephyr-7B');
-          this.useQwen3 = false;
-        }
-      }
-    } catch (error) {
-      console.error('Failed to test Qwen3 connection:', error);
-      console.log('⚠️ Falling back to Zephyr-7B due to connection error');
-      this.useQwen3 = false;
-    }
-
+    // Default to using Qwen3, test will happen on first predict() call
+    this.useQwen3 = true;
     this.connectionTested = true;
+    
+    console.log('✅ Text generation manager initialized - ready to use Qwen3 (with fallback to Zephyr-7B)');
   }
 
-  /**
-   * Perform secondary validation to ensure Qwen3 is actually working
-   */
-  private async performSecondaryValidation(): Promise<void> {
-    try {
-      console.log('🔧 Performing secondary validation of Qwen3 functionality...');
-      
-      // Test with a specific request that should return predictable content
-      const validationResult = await Promise.race([
-        this.primaryClient.predict('/chat', [
-          'Respond with exactly this text: "VALIDATION_SUCCESS"',
-          [],
-          'You are a helpful assistant. Follow instructions exactly as given.',
-          20, // Very small token limit
-          0.1, // Low temperature for deterministic output
-          0.9,
-          10,
-          1.0
-        ]),
-        new Promise<any>((_, reject) => {
-          setTimeout(() => reject(new Error('Validation timeout')), 10000); // 10 second timeout
-        })
-      ]);
-
-      const response = validationResult?.data?.[0] || '';
-      const isValidResponse = typeof response === 'string' && 
-                             response.length > 0 && 
-                             !response.includes('temporarily unavailable') &&
-                             !response.includes('using_fallback');
-
-      if (!isValidResponse) {
-        console.warn('⚠️ Secondary validation failed - Qwen3 responses seem invalid, switching to fallback');
-        this.useQwen3 = false;
-      } else {
-        console.log('✅ Secondary validation passed - Qwen3 is fully functional');
-      }
-    } catch (error) {
-      console.warn('⚠️ Secondary validation failed with error, switching to fallback:', error);
-      this.useQwen3 = false;
-    }
-  }
 
   /**
    * Get the active client for text generation
@@ -123,7 +59,7 @@ class TextGenerationManager {
   }
 
   /**
-   * Predict method with automatic fallback
+   * Predict method with automatic fallback - tests on first failure
    */
   async predict(endpoint: string, params: any[]): Promise<{data: any[]}> {
     // Ensure initialization has been attempted
@@ -144,11 +80,12 @@ class TextGenerationManager {
       
       // If primary client fails and we have a fallback, try it
       if (this.useQwen3 && this.fallbackClient) {
-        console.log('🔄 Qwen3 failed, trying fallback to Zephyr-7B...');
+        console.log('🔄 Qwen3 failed, switching to fallback Zephyr-7B...');
         try {
           const fallbackResult = await this.fallbackClient.predict(endpoint, params);
           // Mark for future calls to use fallback
           this.useQwen3 = false;
+          console.log('✅ Fallback to Zephyr-7B successful - will use Zephyr-7B for future requests');
           return fallbackResult;
         } catch (fallbackError) {
           console.error('Fallback client also failed:', fallbackError);