Pipeline d'évaluation des fournisseurs d'API de recherche

import requests, os, json, time

SCAVIO_KEY = os.environ["SCAVIO_API_KEY"]
SH = {"x-api-key": SCAVIO_KEY, "Content-Type": "application/json"}

TEST_QUERIES = [
    {"query": "best crm for startups 2026", "category": "informational"},
    {"query": "sony wh-1000xm6 price", "category": "product"},
    {"query": "python asyncio tutorial", "category": "technical"},
    {"query": "restaurants near times square", "category": "local"},
    {"query": "latest ai agent frameworks", "category": "trending"},
]

def test_scavio(query: str) -> dict:
    start = time.time()
    resp = requests.post(
        "https://api.scavio.dev/api/v1/search",
        headers=SH,
        json={"query": query, "country_code": "us"},
        timeout=15,
    )
    latency = time.time() - start
    data = resp.json()
    return {
        "provider": "scavio",
        "latency_ms": round(latency * 1000),
        "result_count": len(data.get("organic_results", [])),
        "has_knowledge_graph": "knowledge_graph" in data,
        "has_ai_overview": "ai_overview" in data,
        "has_related_questions": "related_questions" in data,
        "cost_per_query": 0.005,
    }

def run_evaluation():
    results = []
    for tq in TEST_QUERIES:
        scavio_result = test_scavio(tq["query"])
        scavio_result["category"] = tq["category"]
        scavio_result["query"] = tq["query"]
        results.append(scavio_result)

    avg_latency = sum(r["latency_ms"] for r in results) / len(results)
    avg_results = sum(r["result_count"] for r in results) / len(results)
    kg_rate = sum(1 for r in results if r["has_knowledge_graph"]) / len(results)

    print(f"Scavio Evaluation ({len(results)} queries):")
    print(f"  Avg latency: {avg_latency:.0f}ms")
    print(f"  Avg results: {avg_results:.1f}")
    print(f"  Knowledge Graph rate: {kg_rate:.0%}")
    print(f"  Cost: {len(results) * 0.005:.3f} USD")
    return results

evaluation = run_evaluation()

const SH = {'x-api-key': process.env.SCAVIO_API_KEY, 'Content-Type': 'application/json'};

const TEST_QUERIES = [
  {query:'best crm for startups 2026', category:'informational'},
  {query:'sony wh-1000xm6 price', category:'product'},
  {query:'python asyncio tutorial', category:'technical'},
  {query:'restaurants near times square', category:'local'},
  {query:'latest ai agent frameworks', category:'trending'},
];

async function testScavio(query) {
  const start = Date.now();
  const r = await fetch('https://api.scavio.dev/api/v1/search', {method:'POST', headers:SH, body:JSON.stringify({query, country_code:'us'})});
  const latency = Date.now() - start;
  const data = await r.json();
  return {provider:'scavio', latencyMs:latency, resultCount:(data.organic_results||[]).length, hasKg:'knowledge_graph' in data, hasAio:'ai_overview' in data, hasRelated:'related_questions' in data, costPerQuery:0.005};
}

async function runEvaluation() {
  const results = [];
  for (const tq of TEST_QUERIES) {
    const result = await testScavio(tq.query);
    result.category = tq.category;
    result.query = tq.query;
    results.push(result);
  }
  const avgLatency = results.reduce((s,r)=>s+r.latencyMs,0)/results.length;
  const avgResults = results.reduce((s,r)=>s+r.resultCount,0)/results.length;
  const kgRate = results.filter(r=>r.hasKg).length/results.length;
  console.log('Scavio Evaluation ('+results.length+' queries):');
  console.log('  Avg latency: '+avgLatency.toFixed(0)+'ms');
  console.log('  Avg results: '+avgResults.toFixed(1));
  console.log('  KG rate: '+(kgRate*100).toFixed(0)+'%');
  console.log('  Cost: $'+(results.length*0.005).toFixed(3));
  return results;
}

const evaluation = await runEvaluation();