test_scraper.py

"""Test script to validate scraper with 50 reviews from Chase Mobile"""

import time
from pathlib import Path

from data_collection.scraper import PlayStoreScraper
from data_collection.config import BANKING_APPS, TEST_REVIEW_COUNT


def main():
    """Run a quick test scrape of Chase Mobile"""

    print("=" * 60)
    print("GOOGLE PLAY STORE SCRAPER - TEST RUN")
    print("=" * 60)
    print(f"\nTarget: {TEST_REVIEW_COUNT} reviews from Chase Mobile")
    print("This will help validate:")
    print("  1. Review quality (truncation rate)")
    print("  2. Scrape speed (estimate time for full dataset)")
    print("  3. Schema validation")
    print("  4. Data distribution\n")

    input("Press Enter to start scraping...")

    # Initialize scraper
    scraper = PlayStoreScraper()

    # Test with just Chase Mobile
    test_app_name = "Chase Mobile"
    test_app_id = BANKING_APPS[test_app_name]

    # Start timing
    start_time = time.time()

    # Scrape
    reviews = scraper.scrape_app_reviews(
        app_name=test_app_name,
        app_id=test_app_id,
        target_count=TEST_REVIEW_COUNT,
        checkpoint_path=Path("data/raw/test_checkpoint.csv")
    )

    # End timing
    duration = time.time() - start_time

    # Save test results
    output_path = Path("data/raw/test_reviews_chase.csv")
    scraper.save_reviews(reviews, output_path)

    # Analysis
    print("\n" + "=" * 60)
    print("TEST RESULTS")
    print("=" * 60)

    if len(reviews) > 0:
        truncated_count = sum(1 for r in reviews if r.is_truncated)
        truncation_rate = truncated_count / len(reviews)

        avg_length = sum(len(r.review_text) for r in reviews) / len(reviews)

        # Rating distribution
        rating_counts = {}
        for r in reviews:
            rating_counts[r.star_rating] = rating_counts.get(r.star_rating, 0) + 1

        print(f"\n✓ Reviews collected: {len(reviews)}/{TEST_REVIEW_COUNT}")
        print(f"✓ Time taken: {duration:.1f} seconds ({duration/len(reviews):.2f}s per review)")
        print(f"\nData Quality:")
        print(f"  - Truncation rate: {truncation_rate:.1%} ({truncated_count}/{len(reviews)})")
        print(f"  - Avg review length: {avg_length:.0f} characters")
        print(f"\nRating Distribution:")
        for rating in sorted(rating_counts.keys()):
            count = rating_counts[rating]
            pct = count / len(reviews) * 100
            bar = "█" * int(pct / 2)
            print(f"  {rating} stars: {count:3d} ({pct:5.1f}%) {bar}")

        # Time estimate for full dataset
        reviews_needed = 5000 * 5  # 5 apps, 5K each
        estimated_time = (duration / len(reviews)) * reviews_needed
        print(f"\nEstimated time for full dataset ({reviews_needed:,} reviews):")
        print(f"  {estimated_time / 3600:.1f} hours")

        # Quality warnings
        print("\n" + "=" * 60)
        if truncation_rate > 0.3:
            print("⚠️  WARNING: High truncation rate (>30%)")
            print("   Consider: This may limit feature extraction quality")
        if avg_length < 50:
            print("⚠️  WARNING: Short average review length (<50 chars)")
            print("   Consider: Filter or oversample longer reviews")

        # Check for class imbalance
        one_star = rating_counts.get(1, 0)
        five_star = rating_counts.get(5, 0)
        total = len(reviews)
        if (one_star + five_star) / total > 0.7:
            print("⚠️  WARNING: Rating distribution skewed toward 1-star and 5-star")
            print("   Consider: Oversample 2-4 star reviews for balance")

        print("=" * 60)
        print(f"\n✓ Test data saved to: {output_path}")
        print("\nNext steps:")
        print("  1. Review the CSV file to inspect review quality")
        print("  2. If quality looks good, update config.py:")
        print("     - Change TEST_REVIEW_COUNT to REVIEWS_PER_APP (5000)")
        print("  3. Run full scrape with all 5 banking apps")
        print("  4. Proceed to feature extraction with Phi-3/BERT")

    else:
        print("❌ No reviews collected. Check logs/scraper.log for errors")

    print("=" * 60)


if __name__ == "__main__":
    main()