const { createClient } = require('@supabase/supabase-js');

// Load environment variables
require('dotenv').config();

const supabaseUrl = process.env.SUPABASE_URL;
const supabaseServiceKey = process.env.SUPABASE_SERVICE_KEY;

const supabase = createClient(supabaseUrl, supabaseServiceKey);

async function testVectorSearch() {
    console.log('🔍 Testing vector search function...');
    
    // Create a test embedding (1536 dimensions with small random values)
    const testEmbedding = new Array(1536).fill(0).map(() => Math.random() * 0.1);
    
    console.log('📊 Test embedding created with', testEmbedding.length, 'dimensions');
    
    // Test the vector search function
    const { data, error } = await supabase.rpc('match_document_chunks', {
        query_embedding: testEmbedding,
        match_threshold: 0.1,
        match_count: 5
    });
    
    if (error) {
        console.log('❌ Vector search function error:', error);
        
        if (error.code === '42883') {
            console.log('📝 match_document_chunks function does not exist');
            console.log('');
            console.log('🛠️  Please create the function in Supabase SQL Editor:');
            console.log('');
            console.log(`-- First enable pgvector extension
CREATE EXTENSION IF NOT EXISTS vector;

-- Create vector similarity search function
CREATE OR REPLACE FUNCTION match_document_chunks(
    query_embedding VECTOR(1536),
    match_threshold FLOAT DEFAULT 0.7,
    match_count INTEGER DEFAULT 10
)
RETURNS TABLE (
    id UUID,
    document_id TEXT,
    content TEXT,
    metadata JSONB,
    chunk_index INTEGER,
    similarity FLOAT
)
LANGUAGE SQL STABLE
AS $$
    SELECT
        document_chunks.id,
        document_chunks.document_id,
        document_chunks.content,
        document_chunks.metadata,
        document_chunks.chunk_index,
        1 - (document_chunks.embedding <=> query_embedding) AS similarity
    FROM document_chunks
    WHERE document_chunks.embedding IS NOT NULL
    AND 1 - (document_chunks.embedding <=> query_embedding) > match_threshold
    ORDER BY document_chunks.embedding <=> query_embedding
    LIMIT match_count;
$$;`);
        }
    } else {
        console.log('✅ Vector search function works!');
        console.log('📊 Search results:', data ? data.length : 0, 'matches found');
        if (data && data.length > 0) {
            console.log('First result:', data[0]);
        }
    }
    
    // Also test basic insert with embedding
    console.log('🧪 Testing insert with embedding...');
    
    const testChunk = {
        document_id: 'test-doc-with-embedding',
        content: 'This is a test chunk with an embedding vector',
        chunk_index: 1,
        embedding: testEmbedding,
        metadata: { test: true, hasEmbedding: true }
    };
    
    const { data: insertData, error: insertError } = await supabase
        .from('document_chunks')
        .insert(testChunk)
        .select();
    
    if (insertError) {
        console.log('❌ Insert with embedding failed:', insertError);
    } else {
        console.log('✅ Insert with embedding successful!');
        console.log('Inserted chunk ID:', insertData[0].id);
        
        // Test search again with data
        console.log('🔍 Testing search with actual data...');
        const { data: searchData, error: searchError } = await supabase.rpc('match_document_chunks', {
            query_embedding: testEmbedding,
            match_threshold: 0.5,
            match_count: 5
        });
        
        if (searchError) {
            console.log('❌ Search with data failed:', searchError);
        } else {
            console.log('✅ Search with data successful!');
            console.log('Found', searchData ? searchData.length : 0, 'results');
            if (searchData && searchData.length > 0) {
                console.log('Best match similarity:', searchData[0].similarity);
            }
        }
        
        // Clean up test data
        const { error: deleteError } = await supabase
            .from('document_chunks')
            .delete()
            .eq('document_id', 'test-doc-with-embedding');
            
        if (deleteError) {
            console.log('⚠️  Could not clean up test data:', deleteError.message);
        } else {
            console.log('🧹 Test data cleaned up');
        }
    }
}

testVectorSearch();