codenuk_backend_mine/services/ai-analysis-service/test_data_storage.py

#!/usr/bin/env python3
"""
Test data storage in all databases for AI Analysis Service
"""

import os
import psycopg2
import redis
import pymongo
import json
from datetime import datetime
from dotenv import load_dotenv

# Load environment variables
load_dotenv()

def test_postgres_data_storage():
    """Test PostgreSQL data storage"""
    try:
        conn = psycopg2.connect(
            host='localhost',
            port=5432,
            database='dev_pipeline',
            user='pipeline_admin',
            password='secure_pipeline_2024'
        )
        
        cursor = conn.cursor()
        
        # Check repositories
        cursor.execute("SELECT COUNT(*) FROM all_repositories;")
        repo_count = cursor.fetchone()[0]
        
        # Check analysis sessions
        cursor.execute("SELECT COUNT(*) FROM analysis_sessions;")
        session_count = cursor.fetchone()[0]
        
        # Check file analysis history
        cursor.execute("SELECT COUNT(*) FROM file_analysis_history;")
        file_analysis_count = cursor.fetchone()[0]
        
        # Check code embeddings
        cursor.execute("SELECT COUNT(*) FROM code_embeddings;")
        embedding_count = cursor.fetchone()[0]
        
        cursor.close()
        conn.close()
        
        print(f"📊 PostgreSQL Data Storage:")
        print(f"   📁 Repositories: {repo_count}")
        print(f"   🔍 Analysis Sessions: {session_count}")
        print(f"   📄 File Analyses: {file_analysis_count}")
        print(f"   🧠 Code Embeddings: {embedding_count}")
        
        return True
        
    except Exception as e:
        print(f"❌ PostgreSQL data check failed: {e}")
        return False

def test_redis_data_storage():
    """Test Redis data storage"""
    try:
        r = redis.Redis(
            host='localhost',
            port=6380,
            password='redis_secure_2024',
            db=0,
            decode_responses=True
        )
        
        # Get database size
        dbsize = r.dbsize()
        
        # Get all keys
        keys = r.keys('*')
        
        print(f"📊 Redis Data Storage:")
        print(f"   🔑 Total Keys: {dbsize}")
        if keys:
            print(f"   📋 Sample Keys: {keys[:5]}")
        else:
            print(f"   📋 No keys found")
        
        return True
        
    except Exception as e:
        print(f"❌ Redis data check failed: {e}")
        return False

def test_mongodb_data_storage():
    """Test MongoDB data storage"""
    try:
        client = pymongo.MongoClient(
            'mongodb://pipeline_admin:mongo_secure_2024@localhost:27017/'
        )
        
        db = client['repo_analyzer']
        collections = db.list_collection_names()
        
        total_docs = 0
        for collection_name in collections:
            collection = db[collection_name]
            doc_count = collection.count_documents({})
            total_docs += doc_count
            print(f"   📄 {collection_name}: {doc_count} documents")
        
        print(f"📊 MongoDB Data Storage:")
        print(f"   📁 Collections: {len(collections)}")
        print(f"   📄 Total Documents: {total_docs}")
        
        return True
        
    except Exception as e:
        print(f"❌ MongoDB data check failed: {e}")
        return False

def test_analysis_reports():
    """Test analysis reports storage"""
    try:
        reports_dir = "/home/tech4biz/Desktop/prakash/codenuk/backend_new/codenuk_backend_mine/services/ai-analysis-service/reports"
        
        if not os.path.exists(reports_dir):
            print(f"❌ Reports directory not found: {reports_dir}")
            return False
        
        report_files = [f for f in os.listdir(reports_dir) if f.endswith('.json')]
        
        print(f"📊 Analysis Reports:")
        print(f"   📁 Reports Directory: {reports_dir}")
        print(f"   📄 Report Files: {len(report_files)}")
        
        if report_files:
            # Check the latest report
            latest_report = max(report_files, key=lambda x: os.path.getctime(os.path.join(reports_dir, x)))
            report_path = os.path.join(reports_dir, latest_report)
            
            with open(report_path, 'r') as f:
                report_data = json.load(f)
            
            print(f"   📋 Latest Report: {latest_report}")
            print(f"   📊 Repository ID: {report_data.get('repository_id', 'N/A')}")
            print(f"   📁 Total Files: {report_data.get('total_files', 'N/A')}")
            print(f"   📄 Total Lines: {report_data.get('total_lines', 'N/A')}")
            print(f"   🎯 Quality Score: {report_data.get('code_quality_score', 'N/A')}")
        
        return True
        
    except Exception as e:
        print(f"❌ Analysis reports check failed: {e}")
        return False

def main():
    """Test all data storage systems"""
    print("🔍 Testing Data Storage Systems...")
    print("=" * 60)
    
    postgres_ok = test_postgres_data_storage()
    print()
    
    redis_ok = test_redis_data_storage()
    print()
    
    mongodb_ok = test_mongodb_data_storage()
    print()
    
    reports_ok = test_analysis_reports()
    print()
    
    print("=" * 60)
    print(f"📊 Storage Summary:")
    print(f"   PostgreSQL: {'✅' if postgres_ok else '❌'}")
    print(f"   Redis: {'✅' if redis_ok else '❌'}")
    print(f"   MongoDB: {'✅' if mongodb_ok else '❌'}")
    print(f"   Reports: {'✅' if reports_ok else '❌'}")
    
    if all([postgres_ok, redis_ok, mongodb_ok, reports_ok]):
        print("🎉 All data storage systems working!")
    else:
        print("⚠️  Some data storage systems have issues")

if __name__ == "__main__":
    main()