OCR/backend/test_empty_layout.py

#!/usr/bin/env python
"""
測試 calculate_page_dimensions 是否正確處理 layout=[] 但 text_regions 有數據的情況
這模擬了用戶報告的 ELER-8-100HFV Data Sheet 的場景
"""

import json
from pathlib import Path
from app.services.pdf_generator_service import pdf_generator_service
import logging

# Set up logging
logging.basicConfig(level=logging.INFO, format='%(levelname)s: %(message)s')

def test_empty_layout_with_text_regions():
    """
    測試場景：
    - layout: [] (空列表)
    - text_regions: 包含高解析度 bbox 數據
    - 應該從 text_regions 推斷出正確的 OCR 尺寸
    """

    test_dir = Path("test_output_empty_layout")
    test_dir.mkdir(exist_ok=True)

    print("\n" + "="*70)
    print("測試場景：layout=[] 但 text_regions 包含數據")
    print("="*70)

    # 模擬用戶的 JSON 結構
    mock_ocr_data = {
        "status": "success",
        "file_name": "ELER-8-100HFV_Data_Sheet.pdf",
        "language": "ch",
        "layout": [],  # *** 關鍵：這是空的 ***
        "text_regions": [
            {
                "text": "義典科技",
                "bbox": [[461, 270], [819, 252], [822, 408], [464, 426]],  # 高解析度座標
                "confidence": 0.95
            },
            {
                "text": "ELER-8-100HFV",
                "bbox": [[1150, 580], [1850, 580], [1850, 680], [1150, 680]],
                "confidence": 0.93
            },
            {
                "text": "表格中的文字",
                "bbox": [[1259, 936], [1317, 936], [1317, 960], [1259, 960]],  # X=1259 超出 A4 寬度
                "confidence": 0.92
            },
            {
                "text": "底部文字",
                "bbox": [[400, 2800], [1200, 2800], [1200, 2880], [400, 2880]],  # Y=2880
                "confidence": 0.91
            }
        ],
        "total_text_regions": 4,
        "average_confidence": 0.928,
        "layout_data": None,
        "images_metadata": [],
        "markdown_content": "義典科技\nELER-8-100HFV\n表格中的文字\n底部文字",
        "processing_time": 3.2,
        "timestamp": "2025-11-17T00:00:00"
    }

    # Save mock JSON
    json_path = test_dir / "empty_layout_test.json"
    with open(json_path, "w", encoding="utf-8") as f:
        json.dump(mock_ocr_data, f, ensure_ascii=False, indent=2)

    print(f"\n✓ 創建測試 JSON: {json_path}")
    print(f"  - layout: [] (空列表)")
    print(f"  - text_regions: 4 個區域")
    print(f"  - OCR 座標範圍: X=[400..1850], Y=[252..2880]")
    print(f"  - 預期 OCR 尺寸: ~1850 x ~2880")

    # Create A4 source PDF
    from reportlab.pdfgen import canvas
    from reportlab.lib.pagesizes import A4

    source_pdf = test_dir / "source_a4.pdf"
    c = canvas.Canvas(str(source_pdf), pagesize=A4)
    c.drawString(100, 800, "Original A4 Document")
    c.save()

    print(f"✓ 創建 A4 源文件: {source_pdf}")
    print(f"  - A4 尺寸: 595 x 842 點")

    # Test PDF generation
    pdf_path = test_dir / "output.pdf"

    print(f"\n開始生成 PDF...")
    print("-" * 70)

    success = pdf_generator_service.generate_layout_pdf(
        json_path=json_path,
        output_path=pdf_path,
        source_file_path=source_pdf
    )

    print("-" * 70)

    if success:
        print(f"\n✓ PDF 生成成功: {pdf_path}")
        print(f"\n預期結果:")
        print(f"  - OCR 尺寸（從 text_regions 推斷）: ~1850 x ~2880")
        print(f"  - 目標 PDF 尺寸: 595 x 842")
        print(f"  - 預期縮放因子: X={595/1850:.3f}, Y={842/2880:.3f}")
        print(f"\n如果實際縮放因子是 1.0，說明 Bug 仍存在！")
        return True
    else:
        print(f"\n✗ PDF 生成失敗")
        return False

if __name__ == "__main__":
    import sys
    sys.path.insert(0, str(Path(__file__).parent))

    success = test_empty_layout_with_text_regions()

    print("\n" + "="*70)
    if success:
        print("✓ 測試完成")
        print("="*70)
        sys.exit(0)
    else:
        print("✗ 測試失敗")
        print("="*70)
        sys.exit(1)