examples.py

"""
Example usage of tablediff-arrow library.

Install from PyPI: pip install tablediff-arrow

This script demonstrates various ways to use the tablediff-arrow package
for comparing tables with different formats and options.
"""

import tempfile
from pathlib import Path

import pyarrow as pa
import pyarrow.parquet as pq

from tablediff_arrow import TableDiff
from tablediff_arrow.reports import generate_csv_report, generate_html_report


def example_basic_comparison():
    """Basic table comparison example."""
    print("=" * 60)
    print("Example 1: Basic Table Comparison")
    print("=" * 60)

    # Create sample tables
    left = pa.table(
        {
            "id": [1, 2, 3, 4],
            "name": ["Alice", "Bob", "Charlie", "David"],
            "amount": [100.0, 200.0, 300.0, 400.0],
        }
    )

    right = pa.table(
        {
            "id": [1, 2, 3, 5],
            "name": ["Alice", "Bob", "Charlie", "Eve"],
            "amount": [100.0, 205.0, 300.0, 500.0],
        }
    )

    # Compare tables
    differ = TableDiff(key_columns=["id"])
    result = differ.compare_tables(left, right)

    # Print summary
    print("\n" + result.summary())
    print()


def example_with_tolerance():
    """Table comparison with numeric tolerance."""
    print("=" * 60)
    print("Example 2: Comparison with Numeric Tolerance")
    print("=" * 60)

    left = pa.table({"id": [1, 2, 3], "value": [100.0, 200.0, 300.0]})

    right = pa.table({"id": [1, 2, 3], "value": [100.01, 200.02, 300.03]})

    # Without tolerance
    print("\nWithout tolerance:")
    differ = TableDiff(key_columns=["id"])
    result = differ.compare_tables(left, right)
    print(f"Changed rows: {result.changed_rows}")

    # With absolute tolerance
    print("\nWith absolute tolerance of 0.05:")
    differ = TableDiff(key_columns=["id"], tolerance={"value": 0.05})
    result = differ.compare_tables(left, right)
    print(f"Changed rows: {result.changed_rows}")
    print()


def example_file_comparison():
    """Compare files from disk."""
    print("=" * 60)
    print("Example 3: File Comparison")
    print("=" * 60)

    with tempfile.TemporaryDirectory() as tmpdir:
        tmppath = Path(tmpdir)

        # Create test files
        left = pa.table({"id": [1, 2, 3], "value": [10, 20, 30]})
        right = pa.table({"id": [1, 2, 3], "value": [10, 21, 30]})

        left_path = tmppath / "left.parquet"
        right_path = tmppath / "right.parquet"

        pq.write_table(left, left_path)
        pq.write_table(right, right_path)

        # Compare files
        differ = TableDiff(key_columns=["id"])
        result = differ.compare_files(left_path, right_path)

        print("\n" + result.summary())
        print()


def example_html_report():
    """Generate an HTML report."""
    print("=" * 60)
    print("Example 4: HTML Report Generation")
    print("=" * 60)

    left = pa.table(
        {
            "id": [1, 2, 3, 4],
            "category": ["A", "B", "C", "D"],
            "value": [100, 200, 300, 400],
        }
    )

    right = pa.table(
        {
            "id": [1, 2, 3, 5],
            "category": ["A", "B", "C", "E"],
            "value": [100, 210, 300, 500],
        }
    )

    differ = TableDiff(key_columns=["id"])
    result = differ.compare_tables(left, right)

    with tempfile.TemporaryDirectory() as tmpdir:
        report_path = Path(tmpdir) / "report.html"
        generate_html_report(result, report_path)
        print(f"\nHTML report generated at: {report_path}")
        print(f"Report size: {report_path.stat().st_size} bytes")
        print()


def example_csv_reports():
    """Generate CSV reports."""
    print("=" * 60)
    print("Example 5: CSV Reports Generation")
    print("=" * 60)

    left = pa.table(
        {
            "year": [2020, 2020, 2021],
            "month": [1, 2, 1],
            "sales": [1000, 2000, 3000],
        }
    )

    right = pa.table(
        {
            "year": [2020, 2020, 2021],
            "month": [1, 2, 1],
            "sales": [1000, 2100, 3000],
        }
    )

    differ = TableDiff(key_columns=["year", "month"])
    result = differ.compare_tables(left, right)

    with tempfile.TemporaryDirectory() as tmpdir:
        generate_csv_report(result, tmpdir, prefix="sales_diff")
        print("\nCSV reports generated:")
        for csv_file in Path(tmpdir).glob("*.csv"):
            print(f"  - {csv_file.name} ({csv_file.stat().st_size} bytes)")
        print()


def example_relative_tolerance():
    """Compare with relative tolerance."""
    print("=" * 60)
    print("Example 6: Relative Tolerance")
    print("=" * 60)

    left = pa.table(
        {
            "id": [1, 2],
            "small_value": [10.0, 20.0],
            "large_value": [1000.0, 2000.0],
        }
    )

    right = pa.table(
        {
            "id": [1, 2],
            "small_value": [10.1, 20.1],  # 1% difference
            "large_value": [1010.0, 2020.0],  # 1% difference
        }
    )

    # With absolute tolerance
    print("\nWith absolute tolerance of 0.5:")
    differ = TableDiff(key_columns=["id"], tolerance={"small_value": 0.5, "large_value": 0.5})
    result = differ.compare_tables(left, right)
    print(f"Changed rows: {result.changed_rows}")
    print(f"Changes: {result.column_changes}")

    # With relative tolerance
    print("\nWith relative tolerance of 2% (0.02):")
    differ = TableDiff(
        key_columns=["id"],
        relative_tolerance={"small_value": 0.02, "large_value": 0.02},
    )
    result = differ.compare_tables(left, right)
    print(f"Changed rows: {result.changed_rows}")
    print(f"Changes: {result.column_changes}")
    print()


if __name__ == "__main__":
    print("\n" + "=" * 60)
    print("tablediff-arrow Examples")
    print("=" * 60 + "\n")

    example_basic_comparison()
    example_with_tolerance()
    example_file_comparison()
    example_html_report()
    example_csv_reports()
    example_relative_tolerance()

    print("=" * 60)
    print("All examples completed successfully!")
    print("=" * 60)