Batch Research

POST /research/batch

{
  "research_id": "a7f3c8e9-4b2d-4a1e-8c5f-9d7e6f8a3b2c",
  "total_companies": 2,
  "search_strategies_generated": 12,
  "total_searches_executed": 24,
  "processing_time_ms": 34200,
  "results": [
    {
      "domain": "stripe.com",
      "confidence_score": 0.92,
      "evidence_sources": 3,
      "findings": {
        "technologies": [
          "tensorflow",
          "python",
          "kubernetes",
          "radar",
          "machine-learning"
        ],
        "evidence": [
          {
            "url": "https://stripe.com/blog/radar-2.0",
            "title": "Introducing Radar 2.0: Advanced fraud detection with machine learning",
            "snippet": "Stripe Radar uses adaptive machine learning algorithms to detect and prevent fraud in real-time across millions of transactions...",
            "source_name": "google_search"
          },
          {
            "url": "https://stripe.com/jobs/listing/machine-learning-engineer-fraud/5678",
            "title": "Machine Learning Engineer - Fraud Detection",
            "snippet": "Build and deploy ML models for real-time fraud detection using TensorFlow and Python. Work on Stripe Radar...",
            "source_name": "jobs_search"
          },
          {
            "url": "https://newsapi.org/stripe-announces-100m-investment",
            "title": "Stripe Announces $100M Investment in AI Fraud Prevention",
            "snippet": "Payment processor Stripe today announced a major investment in artificial intelligence capabilities for fraud detection...",
            "source_name": "news_search"
          }
        ],
        "signals_found": 8
      }
    },
    {
      "domain": "paypal.com",
      "confidence_score": 0.85,
      "evidence_sources": 2,
      "findings": {
        "technologies": [
          "deep-learning",
          "java",
          "scala",
          "risk-management"
        ],
        "evidence": [
          {
            "url": "https://www.paypal.com/us/webapps/mpp/security/fraud-protection",
            "title": "PayPal Fraud Protection - Advanced Security",
            "snippet": "Our advanced AI and machine learning systems monitor transactions 24/7 to detect and prevent fraudulent activity...",
            "source_name": "google_search"
          },
          {
            "url": "https://newsapi.org/paypal-fraud-detection-ai",
            "title": "PayPal Enhances Fraud Detection with Deep Learning",
            "snippet": "PayPal has deployed new deep learning models that reduce false positives by 30% while catching more fraud...",
            "source_name": "news_search"
          }
        ],
        "signals_found": 6
      }
    }
  ],
  "search_performance": {
    "queries_per_second": 18.5,
    "failed_requests": 2
  }
}

// High precision - only strong matches
const strictSearch = {
  confidence_threshold: 0.85,
  // ...
};

// High recall - more potential matches
const broadSearch = {
  confidence_threshold: 0.6,
  // ...
};

import asyncio
import aiohttp

BATCH_SIZE = 50

async def research_batch(domains, research_goal):
    async with aiohttp.ClientSession() as session:
        for i in range(0, len(domains), BATCH_SIZE):
            batch = domains[i:i + BATCH_SIZE]
            payload = {
                'research_goal': research_goal,
                'company_domains': batch,
                'search_depth': 'standard',
                'max_parallel_searches': 20,
                'confidence_threshold': 0.7,
            }
            async with session.post(
                'http://localhost:8000/research/batch',
                json=payload
            ) as resp:
                data = await resp.json()
                yield data

# Process 500 companies in batches of 50
domains = [f"company{i}.com" for i in range(500)]
async for result in research_batch(domains, "Find AI startups"):
    print(f"Batch complete: {len(result['results'])} companies")

def filter_high_confidence(response, min_confidence=0.8):
    """Filter results by confidence score."""
    return [
        result for result in response['results']
        if result['confidence_score'] >= min_confidence
    ]

data = response.json()
high_confidence = filter_high_confidence(data, min_confidence=0.85)
print(f"Found {len(high_confidence)} high-confidence matches")

function extractAllTechnologies(response) {
  const allTechs = new Set();
  
  response.results.forEach(result => {
    result.findings.technologies.forEach(tech => {
      allTechs.add(tech);
    });
  });
  
  return Array.from(allTechs);
}

const technologies = extractAllTechnologies(data);
console.log('Technologies found:', technologies);
// Output: ['tensorflow', 'python', 'kubernetes', 'java', ...]

import csv

def export_to_csv(response, filename='research_results.csv'):
    """Export research results to CSV."""
    with open(filename, 'w', newline='') as f:
        writer = csv.writer(f)
        writer.writerow([
            'Domain',
            'Confidence',
            'Evidence Sources',
            'Technologies',
            'Signals Found'
        ])
        
        for result in response['results']:
            writer.writerow([
                result['domain'],
                result['confidence_score'],
                result['evidence_sources'],
                ', '.join(result['findings']['technologies']),
                result['findings']['signals_found']
            ])

export_to_csv(data)