Delete qwen-embedding-service-docker.ts

qwen-embedding-service-docker.ts

@@ -1,209 +0,0 @@
-// Qwen Embedding Service using Docker-based Hugging Face Space
-// This version uses direct HTTP calls instead of Gradio client for better stability
-
-const QWEN_API_URL = process.env.QWEN_API_URL || 'https://your-username-qwen-embedding-api.hf.space';
-
-// Helper function to call Qwen Embeddings API via HTTP
-export async function generateQwenEmbeddings(texts: string[]): Promise<number[][]> {
-  try {
-    console.log(`Calling Qwen API for ${texts.length} texts...`);
-    
-    const response = await fetch(`${QWEN_API_URL}/api/predict`, {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-      },
-      body: JSON.stringify({
-        data: [texts] // Wrap in array for batch processing
-      }),
-    });
-
-    if (!response.ok) {
-      throw new Error(`HTTP error! status: ${response.status}`);
-    }
-
-    const data = await response.json();
-    
-    if (data.error) {
-      throw new Error(`API Error: ${data.error}`);
-    }
-
-    // The response should be in the format: { data: [embeddings] }
-    const embeddings = data.data[0];
-    
-    if (!Array.isArray(embeddings)) {
-      throw new Error('Invalid embeddings format received from Qwen API');
-    }
-
-    // Validate embeddings
-    for (let i = 0; i < embeddings.length; i++) {
-      if (!Array.isArray(embeddings[i])) {
-        throw new Error(`Embedding ${i} is not an array`);
-      }
-      if (embeddings[i].length === 0) {
-        throw new Error(`Embedding ${i} is empty`);
-      }
-    }
-
-    console.log(`Successfully generated ${embeddings.length} embeddings`);
-    return embeddings;
-    
-  } catch (error) {
-    console.error('Error calling Qwen embeddings API:', error);
-    throw error;
-  }
-}
-
-// Helper function to generate single embedding
-export async function generateSingleQwenEmbedding(text: string): Promise<number[]> {
-  try {
-    console.log('Calling Qwen API for single text...');
-    
-    const response = await fetch(`${QWEN_API_URL}/api/predict`, {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-      },
-      body: JSON.stringify({
-        data: [text] // Single text
-      }),
-    });
-
-    if (!response.ok) {
-      throw new Error(`HTTP error! status: ${response.status}`);
-    }
-
-    const data = await response.json();
-    
-    if (data.error) {
-      throw new Error(`API Error: ${data.error}`);
-    }
-
-    // The response should be in the format: { data: [embedding] }
-    const embedding = data.data[0];
-    
-    if (!Array.isArray(embedding)) {
-      throw new Error('Invalid embedding format received from Qwen API');
-    }
-
-    if (embedding.length === 0) {
-      throw new Error('Empty embedding received from Qwen API');
-    }
-
-    console.log('Successfully generated single embedding');
-    return embedding;
-    
-  } catch (error) {
-    console.error('Error calling Qwen single embedding API:', error);
-    // Fallback to batch processing
-    const embeddings = await generateQwenEmbeddings([text]);
-    return embeddings[0];
-  }
-}
-
-// Health check function
-export async function checkQwenAPIHealth(): Promise<boolean> {
-  try {
-    const response = await fetch(`${QWEN_API_URL}/health`, {
-      method: 'GET',
-    });
-
-    if (!response.ok) {
-      return false;
-    }
-
-    const data = await response.json();
-    return data.status === 'healthy' && data.model_loaded === true;
-    
-  } catch (error) {
-    console.error('Health check failed:', error);
-    return false;
-  }
-}
-
-// Retry mechanism for Qwen API
-async function generateQwenEmbeddingsWithRetry(texts: string[], maxRetries: number = 3): Promise<number[][]> {
-  let lastError: Error | null = null;
-  
-  for (let attempt = 1; attempt <= maxRetries; attempt++) {
-    try {
-      console.log(`Attempt ${attempt}/${maxRetries} to generate embeddings...`);
-      return await generateQwenEmbeddings(texts);
-    } catch (error) {
-      lastError = error as Error;
-      console.warn(`Attempt ${attempt} failed:`, error);
-      
-      if (attempt < maxRetries) {
-        const delay = Math.pow(2, attempt) * 1000; // Exponential backoff
-        console.log(`Waiting ${delay}ms before retry...`);
-        await new Promise(resolve => setTimeout(resolve, delay));
-      }
-    }
-  }
-  
-  throw lastError || new Error('Qwen API failed after all retries');
-}
-
-// Fallback to Jina if Qwen fails
-export async function generateEmbeddingsWithFallback(texts: string[]): Promise<number[][]> {
-  try {
-    // Check API health first
-    const isHealthy = await checkQwenAPIHealth();
-    if (!isHealthy) {
-      throw new Error('Qwen API is not healthy');
-    }
-
-    // Try Qwen first with retry
-    return await generateQwenEmbeddingsWithRetry(texts);
-  } catch (qwenError) {
-    console.warn('Qwen API failed after retries, falling back to Jina:', qwenError);
-    
-    // Fallback to Jina
-    const JINA_API_KEY = process.env.JINA_API_KEY;
-    const JINA_EMBEDDINGS_MODEL = process.env.JINA_EMBEDDINGS_MODEL || 'jina-embeddings-v3';
-
-    if (!JINA_API_KEY) {
-      throw new Error('Both Qwen and Jina APIs failed. JINA_API_KEY not available for fallback.');
-    }
-
-    const response = await fetch('https://api.jina.ai/v1/embeddings', {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-        'Authorization': `Bearer ${JINA_API_KEY}`,
-      },
-      body: JSON.stringify({
-        model: JINA_EMBEDDINGS_MODEL,
-        input: texts,
-      }),
-    });
-
-    if (!response.ok) {
-      const errorText = await response.text();
-      throw new Error(`Jina API error: ${response.status} ${response.statusText} - ${errorText}`);
-    }
-
-    const data = await response.json();
-    return data.data.map((item: any) => item.embedding);
-  }
-}
-
-// Main function that uses Qwen with Jina fallback
-export async function generateEmbeddings(texts: string[]): Promise<number[][]> {
-  // For single text, use the optimized single embedding endpoint
-  if (texts.length === 1) {
-    try {
-      const embedding = await generateSingleQwenEmbedding(texts[0]);
-      return [embedding];
-    } catch (error) {
-      console.warn('Single embedding failed, falling back to batch processing:', error);
-      // Fall through to batch processing
-    }
-  }
-
-  // Use batch processing with fallback
-  return await generateEmbeddingsWithFallback(texts);
-}
-
-// Export the single embedding function for compatibility
-export const generateSingleEmbedding = generateSingleQwenEmbedding;