dgx-spark-playbooks/nvidia/txt2kg/assets/frontend/app/rag/page.tsx
2025-12-02 19:43:52 +00:00

591 lines
26 KiB
TypeScript

//
// SPDX-FileCopyrightText: Copyright (c) 1993-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
// SPDX-License-Identifier: Apache-2.0
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
"use client";
import { useState, useEffect } from "react";
import { RagQuery, RagParams } from "@/components/rag-query";
import type { Triple } from "@/types/graph";
import Link from "next/link";
import { useRouter } from "next/navigation";
import { DatabaseConnection } from "@/components/database-connection";
import { NvidiaIcon } from "@/components/nvidia-icon";
import { ArrowLeft, BarChart2, Search as SearchIcon } from "lucide-react";
export default function RagPage() {
const router = useRouter();
const [results, setResults] = useState<Triple[] | null>(null);
const [llmAnswer, setLlmAnswer] = useState<string | null>(null);
const [isLoading, setIsLoading] = useState(false);
const [errorMessage, setErrorMessage] = useState<string | null>(null);
const [vectorEnabled, setVectorEnabled] = useState(false);
const [metrics, setMetrics] = useState<{
avgQueryTime: number;
avgRelevance: number;
precision: number;
recall: number;
queryTimesByMode?: Record<string, number>;
} | null>(null);
const [currentParams, setCurrentParams] = useState<RagParams>({
kNeighbors: 4096,
fanout: 400,
numHops: 2,
topK: 5,
useVectorSearch: false,
usePureRag: false,
queryMode: 'traditional'
});
// Initialize backend when the page loads
useEffect(() => {
// Initialize the backend services
const initializeBackend = async () => {
try {
// Check graph database connection (ArangoDB by default)
const graphResponse = await fetch('/api/graph-db', {
method: 'GET',
headers: {
'Content-Type': 'application/json',
},
});
if (!graphResponse.ok) {
const errorData = await graphResponse.json();
console.warn('Graph database connection warning:', errorData.error);
}
// Check if vector search is available
const vectorResponse = await fetch('/api/pinecone-diag/stats');
if (vectorResponse.ok) {
const data = await vectorResponse.json();
setVectorEnabled(data.totalVectorCount > 0);
}
// Fetch basic metrics
const metricsResponse = await fetch('/api/metrics');
if (metricsResponse.ok) {
const data = await metricsResponse.json();
setMetrics({
avgQueryTime: data.avgQueryTime,
avgRelevance: data.avgRelevance,
precision: data.precision,
recall: data.recall,
queryTimesByMode: data.queryTimesByMode
});
}
} catch (error) {
console.warn('Error initializing backends:', error);
}
};
initializeBackend();
}, []);
const handleQuerySubmit = async (query: string, params: RagParams) => {
setIsLoading(true);
setErrorMessage(null);
setCurrentParams(params); // Store current params for UI rendering
const startTime = Date.now();
let queryMode: 'pure-rag' | 'vector-search' | 'traditional' = 'traditional';
let resultCount = 0;
let relevanceScore = 0;
// Debug logging
console.log('🔍 Query params:', {
usePureRag: params.usePureRag,
useVectorSearch: params.useVectorSearch,
vectorEnabled,
queryMode: params.queryMode
});
try {
// If using pure RAG (Pinecone + LangChain) without graph search
if (params.usePureRag) {
queryMode = 'pure-rag';
try {
console.log('Using pure RAG with Qdrant and NVIDIA LLM for query:', query);
const ragResponse = await fetch('/api/rag-query', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
query,
topK: params.topK
})
});
if (ragResponse.ok) {
const data = await ragResponse.json();
console.log('📥 RAG Response data:', {
hasAnswer: !!data.answer,
answerLength: data.answer?.length,
documentCount: data.documentCount
});
// Handle the answer - we might need to display differently than triples
if (data.answer) {
console.log('✅ Setting answer in results:', data.answer.substring(0, 100) + '...');
// Set the LLM answer for display (same as traditional mode)
setLlmAnswer(data.answer);
// Set empty results array since Pure RAG doesn't return triples
setResults([]);
resultCount = data.documentCount || 0;
relevanceScore = data.relevanceScore || 0;
// Log the query with performance metrics
logQuery(query, queryMode, {
executionTimeMs: Date.now() - startTime,
relevanceScore,
resultCount
});
console.log(`✅ Pure RAG query completed. Retrieved ${resultCount} document chunks`);
setIsLoading(false);
return;
}
} else {
// If the RAG query fails, log but continue to try other methods
const errorData = await ragResponse.json();
throw new Error(errorData.error || 'Failed to execute pure RAG query');
}
} catch (ragError) {
console.warn('Pure RAG query error (falling back to other methods):', ragError);
// Continue to other query methods as fallback
}
}
// If we have vector embeddings AND explicitly selected vector search, use enhanced query with metadata
if (vectorEnabled && params.useVectorSearch && !params.usePureRag) {
queryMode = 'vector-search';
try {
console.log('Using enhanced RAG with LangChain for query:', query);
const enhancedResponse = await fetch('/api/enhanced-query', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
query,
kNeighbors: params.kNeighbors,
fanout: params.fanout,
numHops: params.numHops,
topK: params.topK
})
});
if (enhancedResponse.ok) {
const data = await enhancedResponse.json();
// Update the results
setResults(data.relevantTriples || []);
resultCount = data.count || 0;
relevanceScore = data.relevanceScore || 0;
// Log the query with performance metrics
logQuery(query, queryMode, {
executionTimeMs: Date.now() - startTime,
relevanceScore,
resultCount,
precision: data.precision || 0,
recall: data.recall || 0,
});
// Log to console instead of showing alert
let message = `Enhanced query completed. Found ${resultCount} relevant triples`;
if (data.metadata?.entityMatches) {
message += ` from ${data.metadata.entityMatches} matched entities`;
}
console.log(message);
setIsLoading(false);
return;
}
} catch (enhancedError) {
console.warn('Enhanced query error (falling back to traditional query):', enhancedError);
// Continue to traditional query as fallback
}
}
// Call the LLM-enhanced graph query API
console.log('✅ Using Graph Search + LLM approach');
queryMode = 'traditional';
// Get selected LLM model from localStorage
let llmModel = undefined;
let llmProvider = undefined;
try {
const savedModel = localStorage.getItem("selectedModelForRAG");
if (savedModel) {
const modelData = JSON.parse(savedModel);
llmModel = modelData.model;
llmProvider = modelData.provider;
console.log(`Using LLM: ${llmModel} (${llmProvider})`);
}
} catch (e) {
console.warn("Could not load selected LLM model, using default");
}
const response = await fetch(`/api/graph-query-llm`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
},
body: JSON.stringify({
query,
topK: params.topK || 5,
useTraditional: true,
llmModel,
llmProvider
}),
});
if (!response.ok) {
const errorData = await response.json();
throw new Error(errorData.error || 'Failed to query with LLM');
}
const data = await response.json();
// Log sample of retrieved triples for debugging
console.log('📊 Retrieved Triples (sample):', data.triples.slice(0, 3));
console.log('🤖 LLM-Generated Answer (preview):', data.answer?.substring(0, 200) + '...');
console.log('📈 Triple Count:', data.count);
// DEBUG: Check if depth/pathLength are present in received data
if (data.triples && data.triples.length > 0) {
console.log('🔍 First triple structure:', JSON.stringify(data.triples[0], null, 2));
console.log('🔍 Has depth?', 'depth' in data.triples[0]);
console.log('🔍 Has pathLength?', 'pathLength' in data.triples[0]);
}
// Update the results with the triples (for display)
setResults(data.triples || []);
resultCount = data.count || 0;
relevanceScore = 0; // No relevance score for traditional search
// Store the LLM answer for display
if (data.answer) {
console.log('✅ Setting llmAnswer state (length:', data.answer.length, 'chars)');
setLlmAnswer(data.answer);
} else {
console.log('⚠️ No answer in response');
setLlmAnswer(null);
}
// Log the query with performance metrics
logQuery(query, queryMode, {
executionTimeMs: Date.now() - startTime,
relevanceScore,
resultCount,
precision: data.precision || 0,
recall: data.recall || 0,
});
// Log to console instead of showing alert
let message = `Query completed. Found ${resultCount} relevant triples`;
if (vectorEnabled && params.useVectorSearch) {
message += ` (using standard vector search)`;
}
console.log(message);
} catch (error) {
console.error("RAG query error:", error);
setErrorMessage(error instanceof Error ? error.message : "An unknown error occurred");
setResults([]);
// Log failed query
logQuery(query, queryMode, {
executionTimeMs: Date.now() - startTime,
resultCount: 0,
error: error instanceof Error ? error.message : "Unknown error"
});
} finally {
setIsLoading(false);
}
};
// Helper function to log queries
const logQuery = async (
query: string,
queryMode: 'pure-rag' | 'vector-search' | 'traditional',
metrics: {
executionTimeMs: number;
relevanceScore?: number;
precision?: number;
recall?: number;
resultCount: number;
error?: string;
}
) => {
try {
await fetch('/api/query-log', {
method: 'POST',
headers: {
'Content-Type': 'application/json'
},
body: JSON.stringify({
query,
queryMode,
metrics
})
});
console.log('Query logged successfully');
} catch (error) {
// Non-blocking error, just log it
console.warn('Failed to log query:', error);
}
};
const clearResults = () => {
setResults(null);
setLlmAnswer(null);
setErrorMessage(null);
};
return (
<div className="min-h-screen bg-background text-foreground">
{/* Main Content */}
<main className="container mx-auto px-6 py-12">
{/* Header Section */}
<div className="flex items-center justify-between mb-8">
<Link href="/" className="inline-flex items-center gap-3 px-4 py-2 text-sm font-medium border border-border/40 hover:border-border/60 bg-background hover:bg-muted/30 rounded-lg transition-colors">
<ArrowLeft className="h-4 w-4" />
Back to Documents
</Link>
</div>
{/* Two Column Layout */}
<div className="grid grid-cols-1 lg:grid-cols-4 gap-8">
{/* Left Column - Database Connections */}
<div className="lg:col-span-1 space-y-6">
<div className="nvidia-build-card">
<DatabaseConnection />
</div>
{/* Performance Metrics Card */}
{metrics && (
<div className="nvidia-build-card">
<div className="flex items-center justify-between mb-4">
<div className="flex items-center gap-3">
<div className="w-6 h-6 rounded-md bg-nvidia-green/15 flex items-center justify-center">
<BarChart2 className="h-3 w-3 text-nvidia-green" />
</div>
<h3 className="text-base font-semibold text-foreground">Performance Metrics</h3>
</div>
<Link href="/rag/metrics" className="text-xs text-nvidia-green hover:text-nvidia-green/80 font-medium underline underline-offset-2">
View All
</Link>
</div>
<div className="space-y-3 text-sm">
{/* Query times by mode */}
{metrics.queryTimesByMode && Object.keys(metrics.queryTimesByMode).length > 0 ? (
<>
{metrics.queryTimesByMode['pure-rag'] !== undefined && (
<div className="flex justify-between">
<span className="text-muted-foreground">Pure RAG:</span>
<span className="font-medium">{(metrics.queryTimesByMode['pure-rag'] / 1000).toFixed(2)}s</span>
</div>
)}
{metrics.queryTimesByMode['traditional'] !== undefined && (
<div className="flex justify-between">
<span className="text-muted-foreground">Graph Search:</span>
<span className="font-medium">{(metrics.queryTimesByMode['traditional'] / 1000).toFixed(2)}s</span>
</div>
)}
{metrics.queryTimesByMode['vector-search'] !== undefined && (
<div className="flex justify-between">
<span className="text-muted-foreground">GraphRAG:</span>
<span className="font-medium">{(metrics.queryTimesByMode['vector-search'] / 1000).toFixed(2)}s</span>
</div>
)}
</>
) : (
<div className="flex justify-between">
<span className="text-muted-foreground">Avg. Query Time:</span>
<span className="font-medium">{metrics.avgQueryTime > 0 ? `${metrics.avgQueryTime.toFixed(2)}ms` : "No data"}</span>
</div>
)}
</div>
</div>
)}
</div>
{/* Right Column - RAG Query Interface */}
<div className="lg:col-span-3">
<RagQuery
onQuerySubmit={handleQuerySubmit}
clearResults={clearResults}
isLoading={isLoading}
error={errorMessage}
vectorEnabled={vectorEnabled}
/>
{/* LLM Answer Section */}
{llmAnswer && (
<div className="mt-8 nvidia-build-card">
<div className="flex items-center gap-3 mb-4">
<div className="w-6 h-6 rounded-md bg-nvidia-green/15 flex items-center justify-center">
<SearchIcon className="h-3 w-3 text-nvidia-green" />
</div>
<h3 className="text-lg font-semibold text-foreground">Answer</h3>
{currentParams.queryMode && (
<span className="text-xs px-2.5 py-1 rounded-full font-medium bg-nvidia-green/10 text-nvidia-green border border-nvidia-green/20">
{currentParams.queryMode === 'pure-rag' ? 'Pure RAG' :
currentParams.queryMode === 'vector-search' ? 'GraphRAG' :
'Graph Search'}
</span>
)}
</div>
<div className="prose prose-sm dark:prose-invert max-w-none">
{(() => {
// Parse <think> tags
const thinkMatch = llmAnswer.match(/<think>([\s\S]*?)<\/think>/);
const thinkContent = thinkMatch ? thinkMatch[1].trim() : null;
const mainAnswer = thinkContent
? llmAnswer.replace(/<think>[\s\S]*?<\/think>/, '').trim()
: llmAnswer;
return (
<>
{thinkContent && (
<details className="mb-4 bg-muted/10 border border-border/20 rounded-xl overflow-hidden group">
<summary className="cursor-pointer p-4 hover:bg-muted/20 transition-colors flex items-center gap-2">
<svg className="w-4 h-4 transform transition-transform group-open:rotate-90" xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path strokeLinecap="round" strokeLinejoin="round" strokeWidth={2} d="M9 5l7 7-7 7" />
</svg>
<span className="text-sm font-medium text-muted-foreground">Reasoning Process</span>
</summary>
<div className="p-4 pt-0 text-sm text-muted-foreground leading-relaxed whitespace-pre-wrap border-t border-border/10">
{thinkContent}
</div>
</details>
)}
<div className="bg-muted/20 border border-border/20 p-6 rounded-xl">
<div
className="text-foreground leading-relaxed whitespace-pre-wrap"
dangerouslySetInnerHTML={{
__html: mainAnswer
.replace(/\*\*(.*?)\*\*/g, '<strong>$1</strong>')
.replace(/\*(.*?)\*/g, '<em>$1</em>')
}}
/>
</div>
</>
);
})()}
</div>
</div>
)}
{/* Results Section */}
{results && results.length > 0 && !currentParams.usePureRag && (
<div className="mt-8 nvidia-build-card">
<div className="flex items-center gap-3 mb-6">
<div className="w-6 h-6 rounded-md bg-nvidia-green/15 flex items-center justify-center">
<SearchIcon className="h-3 w-3 text-nvidia-green" />
</div>
<h3 className="text-lg font-semibold text-foreground">
{llmAnswer ? `Retrieved Knowledge (${results.length})` : `Results (${results.length})`}
</h3>
{results.some((r: any) => r.pathLength && r.pathLength > 1) && (
<span className="text-xs px-2.5 py-1 rounded-full font-medium bg-amber-500/10 text-amber-600 dark:text-amber-400 border border-amber-500/20 flex items-center gap-1.5">
<svg className="w-3 h-3" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path strokeLinecap="round" strokeLinejoin="round" strokeWidth={2} d="M13 10V3L4 14h7v7l9-11h-7z" />
</svg>
Multi-hop enabled
</span>
)}
</div>
<div className="space-y-4">
{results.map((triple, index) => (
<div key={index} className="bg-muted/20 border border-border/20 p-4 rounded-xl">
{currentParams.usePureRag ? (
// Pure RAG display format (no subject/predicate/object columns)
<div className="p-2 rounded">
{triple.usedFallback && (
<div className="mb-2 text-sm px-3 py-1 bg-amber-500/20 text-amber-700 dark:text-amber-400 rounded-md inline-block">
Using general knowledge (no documents found)
</div>
)}
<p className="font-medium break-words">{triple.object}</p>
</div>
) : (
// Standard triple display for other modes
<div className="grid grid-cols-1 md:grid-cols-3 gap-3">
<div className="bg-background/60 border border-border/30 p-3 rounded-lg">
<p className="text-xs font-medium text-nvidia-green uppercase tracking-wider mb-1">Subject</p>
<p className="font-medium break-words text-foreground">{triple.subject}</p>
</div>
<div className="bg-background/60 border border-border/30 p-3 rounded-lg">
<p className="text-xs font-medium text-nvidia-green uppercase tracking-wider mb-1">Predicate</p>
<p className="font-medium break-words text-foreground">{triple.predicate}</p>
</div>
<div className="bg-background/60 border border-border/30 p-3 rounded-lg">
<p className="text-xs font-medium text-nvidia-green uppercase tracking-wider mb-1">Object</p>
<p className="font-medium break-words text-foreground">{triple.object}</p>
</div>
</div>
)}
{triple.confidence && !currentParams.usePureRag && (
<div className="mt-3 flex items-center gap-4 text-xs">
<div className="flex items-center gap-1.5">
<div className="w-2 h-2 rounded-full bg-nvidia-green/60"></div>
<span className="text-muted-foreground">
Confidence: <span className="font-medium text-foreground">{(triple.confidence * 100).toFixed(1)}%</span>
</span>
</div>
{triple.depth !== undefined && (
<div className="flex items-center gap-1.5">
<svg className="w-3 h-3 text-nvidia-green/60" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path strokeLinecap="round" strokeLinejoin="round" strokeWidth={2} d="M13 7l5 5m0 0l-5 5m5-5H6" />
</svg>
<span className="text-muted-foreground">
Hop: <span className="font-medium text-foreground">{triple.depth + 1}</span>
</span>
</div>
)}
{triple.pathLength !== undefined && triple.pathLength > 1 && (
<div className="flex items-center gap-1.5">
<svg className="w-3 h-3 text-amber-500/60" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path strokeLinecap="round" strokeLinejoin="round" strokeWidth={2} d="M9 20l-5.447-2.724A1 1 0 013 16.382V5.618a1 1 0 011.447-.894L9 7m0 13l6-3m-6 3V7m6 10l4.553 2.276A1 1 0 0021 18.382V7.618a1 1 0 00-.553-.894L15 4m0 13V4m0 0L9 7" />
</svg>
<span className="text-amber-600/80 dark:text-amber-400/80">
Multi-hop path (length: <span className="font-medium">{triple.pathLength}</span>)
</span>
</div>
)}
</div>
)}
</div>
))}
</div>
</div>
)}
{results && results.length === 0 && !isLoading && !currentParams.usePureRag && (
<div className="mt-8 nvidia-build-card border-dashed">
<div className="text-center py-8">
<div className="w-12 h-12 rounded-xl bg-muted/30 flex items-center justify-center mx-auto mb-4">
<SearchIcon className="h-6 w-6 text-muted-foreground" />
</div>
<p className="text-foreground font-medium mb-2">No results found for your query</p>
<p className="text-sm text-muted-foreground">Try adjusting your query or parameters</p>
</div>
</div>
)}
</div>
</div>
</main>
</div>
);
}