test/test_file_upload_handler.py

#!/usr/bin/env python3
"""
Test the FileUploadHandler functionality
"""

import os
import sys
from io import BytesIO

# Add parent directory to path
sys.path.append(os.path.dirname(__file__))

from web_app.utils import FileUploadHandler

def test_tmp_directory():
    """Test /tmp directory accessibility"""
    print("Testing /tmp directory...")
    
    if os.path.exists("/tmp"):
        print("✅ /tmp exists")
        print(f"  - Writable: {os.access('/tmp', os.W_OK)}")
        print(f"  - Readable: {os.access('/tmp', os.R_OK)}")
        
        # Test write
        try:
            test_file = "/tmp/test_write.txt"
            with open(test_file, 'w') as f:
                f.write("test")
            os.remove(test_file)
            print("  - Write test: ✅ Success")
        except Exception as e:
            print(f"  - Write test: ❌ Failed - {e}")
    else:
        print("❌ /tmp directory does not exist!")

def test_file_upload_handler():
    """Test FileUploadHandler with mock uploaded file"""
    print("\nTesting FileUploadHandler...")
    
    # Create mock uploaded file
    content = b"word\tfreq\trank\nthe\t69868\t1\nof\t36426\t2\nand\t28891\t3"
    mock_file = BytesIO(content)
    mock_file.name = "test_data.tsv"
    mock_file.type = "text/tab-separated-values"
    mock_file.size = len(content)
    
    # Add getbuffer method
    mock_file.getbuffer = lambda: content
    
    print(f"Mock file created: {mock_file.name} ({len(content)} bytes)")
    
    # Test save to temp
    print("\n1. Testing save_to_temp...")
    temp_path = FileUploadHandler.save_to_temp(mock_file, prefix="test")
    if temp_path:
        print(f"✅ Saved to: {temp_path}")
        print(f"  - File exists: {os.path.exists(temp_path)}")
        print(f"  - File size: {os.path.getsize(temp_path)} bytes")
    else:
        print("❌ Failed to save to temp")
        return
    
    # Test read from temp
    print("\n2. Testing read_from_temp...")
    read_content = FileUploadHandler.read_from_temp(temp_path)
    if read_content:
        print(f"✅ Read content: {len(read_content)} bytes")
        print(f"  - Content matches: {read_content == content}")
        print(f"  - First line: {read_content.decode('utf-8').split('\\n')[0]}")
    else:
        print("❌ Failed to read from temp")
    
    # Test get_file_content
    print("\n3. Testing get_file_content...")
    mock_file.seek(0)  # Reset position
    text_content = FileUploadHandler.get_file_content(mock_file, as_text=True)
    if text_content:
        print(f"✅ Got text content: {len(text_content)} chars")
        print(f"  - First line: {text_content.split('\\n')[0]}")
    else:
        print("❌ Failed to get file content")
    
    # Test cleanup
    print("\n4. Testing cleanup_temp_file...")
    FileUploadHandler.cleanup_temp_file(temp_path)
    exists_after = os.path.exists(temp_path)
    print(f"✅ Cleanup: File removed = {not exists_after}")
    
    # Test file size validation
    print("\n5. Testing validate_file_size...")
    mock_file.size = 100 * 1024 * 1024  # 100 MB
    valid = FileUploadHandler.validate_file_size(mock_file, max_size_mb=300)
    print(f"✅ 100MB file validation: {valid}")
    
    mock_file.size = 400 * 1024 * 1024  # 400 MB
    valid = FileUploadHandler.validate_file_size(mock_file, max_size_mb=300)
    print(f"✅ 400MB file validation: {valid} (should be False)")

def test_with_actual_file():
    """Test with an actual file from the data directory"""
    print("\n\nTesting with actual file...")
    
    test_files = [
        "data/word_freq.txt",
        "data/COCA_5000.txt",
        "data/jpn_word_freq.txt"
    ]
    
    for file_path in test_files:
        if os.path.exists(file_path):
            print(f"\nTesting with {file_path}...")
            
            # Read file content
            with open(file_path, 'rb') as f:
                content = f.read()
            
            # Create mock uploaded file
            mock_file = BytesIO(content)
            mock_file.name = os.path.basename(file_path)
            mock_file.type = "text/tab-separated-values"
            mock_file.size = len(content)
            mock_file.getbuffer = lambda: content
            
            print(f"  - File size: {len(content):,} bytes")
            
            # Test the upload handler
            temp_path = FileUploadHandler.save_to_temp(mock_file, prefix="actual")
            if temp_path:
                print(f"  - ✅ Saved to temp: {temp_path}")
                
                # Read back
                read_content = FileUploadHandler.read_from_temp(temp_path)
                if read_content:
                    print(f"  - ✅ Read back: {len(read_content):,} bytes")
                    print(f"  - ✅ Content matches: {read_content == content}")
                
                # Cleanup
                FileUploadHandler.cleanup_temp_file(temp_path)
                print(f"  - ✅ Cleaned up")
            
            break  # Test just one file
        else:
            print(f"  - File not found: {file_path}")

if __name__ == "__main__":
    print("=" * 60)
    print("FileUploadHandler Test Suite")
    print("=" * 60)
    
    test_tmp_directory()
    test_file_upload_handler()
    test_with_actual_file()
    
    print("\n" + "=" * 60)
    print("Test completed!")