markitect-main/tests/test_roundtrip_consolidated.py

"""
Consolidated Roundtrip Tests for Enhanced Explode-Implode System

This test suite consolidates and updates all roundtrip tests to work with the new
variant system, ensuring backward compatibility while testing new functionality.
"""

import pytest
import tempfile
import subprocess
from pathlib import Path
from typing import List, Dict, Any

from markitect.explode_variants import ExplodeVariant, get_variant_factory


class TestRoundtripBase:
    """Base class for roundtrip tests with common utilities."""

    def setup_method(self):
        """Set up temporary directory for each test."""
        self.temp_dir = Path(tempfile.mkdtemp())

    def teardown_method(self):
        """Clean up temporary directory after each test."""
        import shutil
        shutil.rmtree(self.temp_dir, ignore_errors=True)

    def run_markitect_command(self, args: List[str]) -> subprocess.CompletedProcess:
        """Run a markitect command and return the result."""
        cmd = ["python", "-m", "markitect.cli"] + args
        return subprocess.run(
            cmd,
            capture_output=True,
            text=True,
            cwd="/home/worsch/markitect-main"
        )

    def validate_basic_structure_preservation(self, original: str, reconstructed: str) -> Dict[str, Any]:
        """Validate that basic document structure is preserved."""
        import re

        # Extract headings from both documents
        orig_headings = re.findall(r'^(#+)\s+(.+)', original, re.MULTILINE)
        recon_headings = re.findall(r'^(#+)\s+(.+)', reconstructed, re.MULTILINE)

        return {
            'original_heading_count': len(orig_headings),
            'reconstructed_heading_count': len(recon_headings),
            'headings_preserved': len(orig_headings) == len(recon_headings),
            'original_headings': orig_headings,
            'reconstructed_headings': recon_headings
        }


class TestVariantRoundtrips(TestRoundtripBase):
    """Test roundtrips with all variants using CLI commands."""

    @pytest.fixture
    def sample_document(self):
        """Sample document for testing."""
        return """# Book Title

This is the introduction to our book.

## Chapter 1: Getting Started

Welcome to the first chapter.

### Section 1.1: Overview

Basic overview content.

### Section 1.2: Setup

Setup instructions here.

## Chapter 2: Advanced Topics

More advanced material.

### Section 2.1: Deep Dive

Detailed explanations.

# Conclusion

Final thoughts and summary.
"""

    def test_flat_variant_cli_roundtrip(self, sample_document):
        """Test flat variant roundtrip using CLI commands."""
        self._test_variant_roundtrip(sample_document, "flat")

    def test_hierarchical_variant_cli_roundtrip(self, sample_document):
        """Test hierarchical variant roundtrip using CLI commands."""
        self._test_variant_roundtrip(sample_document, "hierarchical")

    def test_semantic_variant_cli_roundtrip(self, sample_document):
        """Test semantic variant roundtrip using CLI commands."""
        self._test_variant_roundtrip(sample_document, "semantic")

    def _test_variant_roundtrip(self, content: str, variant: str):
        """Generic variant roundtrip test."""
        # Step 1: Create original file
        original_file = self.temp_dir / f"test_{variant}.md"
        original_file.write_text(content, encoding='utf-8')

        # Step 2: Explode using specific variant
        exploded_dir = self.temp_dir / f"test_{variant}.mdd"
        result = self.run_markitect_command([
            "md-explode", str(original_file),
            "--variant", variant,
            "--output-dir", str(exploded_dir)
        ])
        assert result.returncode == 0, f"Explode failed: {result.stderr}"
        assert exploded_dir.exists()

        # Verify manifest was created
        manifest_file = exploded_dir / "manifest.md"
        assert manifest_file.exists()

        # Step 3: Implode back (should auto-detect variant)
        reconstructed_file = self.temp_dir / f"reconstructed_{variant}.md"
        result = self.run_markitect_command([
            "md-implode", str(exploded_dir),
            "--output", str(reconstructed_file)
        ])
        assert result.returncode == 0, f"Implode failed: {result.stderr}"
        assert reconstructed_file.exists()

        # Step 4: Validate content preservation
        reconstructed_content = reconstructed_file.read_text(encoding='utf-8')
        validation = self.validate_basic_structure_preservation(content, reconstructed_content)

        assert validation['headings_preserved'], f"Headings not preserved in {variant} variant"

        # Verify key content is present
        assert "# Book Title" in reconstructed_content
        assert "## Chapter 1: Getting Started" in reconstructed_content
        assert "### Section 1.1: Overview" in reconstructed_content
        assert "# Conclusion" in reconstructed_content


class TestBackwardCompatibilityRoundtrips(TestRoundtripBase):
    """Test backward compatibility with legacy behavior."""

    def test_default_behavior_roundtrip(self):
        """Test that default behavior (flat variant) works like before."""
        content = """# Introduction

Basic introduction content.

## Overview

Overview section.

# Main Content

Main content here.

# Conclusion

Final thoughts.
"""

        # Create original file
        original_file = self.temp_dir / "test.md"
        original_file.write_text(content, encoding='utf-8')

        # Explode without specifying variant (should default to flat)
        result = self.run_markitect_command([
            "md-explode", str(original_file)
        ])
        assert result.returncode == 0

        # Should create .mdd directory with manifest
        exploded_dir = original_file.with_suffix('.mdd')
        assert exploded_dir.exists()
        assert (exploded_dir / "manifest.md").exists()

        # Implode back
        reconstructed_file = self.temp_dir / "reconstructed.md"
        result = self.run_markitect_command([
            "md-implode", str(exploded_dir),
            "--output", str(reconstructed_file)
        ])
        assert result.returncode == 0

        # Validate content
        reconstructed_content = reconstructed_file.read_text(encoding='utf-8')
        assert "# Introduction" in reconstructed_content
        assert "# Main Content" in reconstructed_content
        assert "# Conclusion" in reconstructed_content

    def test_legacy_exploded_directory_handling(self):
        """Test that legacy exploded directories can still be imploded."""
        # Create a structure that looks like legacy exploded content
        legacy_dir = self.temp_dir / "legacy_structure"
        legacy_dir.mkdir()

        # Create some markdown files without manifest
        (legacy_dir / "intro.md").write_text("# Introduction\n\nIntro content.")
        (legacy_dir / "chapter1.md").write_text("# Chapter 1\n\nChapter content.")
        (legacy_dir / "conclusion.md").write_text("# Conclusion\n\nFinal thoughts.")

        # Should still be able to implode
        result = self.run_markitect_command([
            "md-implode", str(legacy_dir)
        ])
        assert result.returncode == 0

        # Check that output file was created
        output_file = legacy_dir.parent / f"{legacy_dir.name}_imploded.md"
        assert output_file.exists()

        content = output_file.read_text(encoding='utf-8')
        assert "# Introduction" in content
        assert "# Chapter 1" in content
        assert "# Conclusion" in content


class TestComplexRoundtrips(TestRoundtripBase):
    """Test roundtrips with complex content and features."""

    def test_front_matter_preservation_roundtrip(self):
        """Test that front matter is preserved through roundtrips."""
        content_with_fm = """---
title: "Test Document"
author: "Test Author"
tags: ["test", "markdown"]
version: 1.0
---

# Main Content

This document has front matter.

## Section 1

Content here.

# Conclusion

End of document.
"""

        original_file = self.temp_dir / "test_fm.md"
        original_file.write_text(content_with_fm, encoding='utf-8')

        # Test with each variant
        for variant in ["flat", "hierarchical", "semantic"]:
            # Explode
            exploded_dir = self.temp_dir / f"test_fm_{variant}.mdd"
            result = self.run_markitect_command([
                "md-explode", str(original_file),
                "--variant", variant,
                "--output-dir", str(exploded_dir)
            ])
            assert result.returncode == 0

            # Implode
            reconstructed_file = self.temp_dir / f"reconstructed_fm_{variant}.md"
            result = self.run_markitect_command([
                "md-implode", str(exploded_dir),
                "--output", str(reconstructed_file)
            ])
            assert result.returncode == 0

            # Verify front matter preservation - check for semantic equivalence
            reconstructed_content = reconstructed_file.read_text(encoding='utf-8')

            # Use frontmatter parser to check semantic equivalence
            from markitect.matter_frontmatter.parser import FrontmatterParser
            parser = FrontmatterParser()
            reconstructed_fm = parser.extract_frontmatter(reconstructed_content)

            # Check that all expected values are preserved
            assert reconstructed_fm.get('title') == 'Test Document'
            assert reconstructed_fm.get('author') == 'Test Author'
            assert reconstructed_fm.get('tags') == ['test', 'markdown']
            assert reconstructed_fm.get('version') == 1.0

    def test_unicode_and_special_characters_roundtrip(self):
        """Test roundtrip with unicode and special characters."""
        unicode_content = """# Tëst Dócümënt

This document contains ünïcödë characters.

## Spëcïál Chàráctërs

- Émojis: 🚀 📝 ✅
- Symbols: © ® ™ € £ ¥
- Math: ∑ ∞ π √ ≈ ≠

### Çødë Blöck

```python
def hëllö_wörld():
    print("Hëllö, Wörld! 🌍")
```

# Cönclüsïön

End öf tëst.
"""

        original_file = self.temp_dir / "unicode_test.md"
        original_file.write_text(unicode_content, encoding='utf-8')

        # Test with flat variant
        result = self.run_markitect_command([
            "md-explode", str(original_file),
            "--variant", "flat"
        ])
        assert result.returncode == 0

        exploded_dir = original_file.with_suffix('.mdd')
        assert exploded_dir.exists()

        # Implode back
        reconstructed_file = self.temp_dir / "unicode_reconstructed.md"
        result = self.run_markitect_command([
            "md-implode", str(exploded_dir),
            "--output", str(reconstructed_file)
        ])
        assert result.returncode == 0

        # Verify unicode preservation
        reconstructed_content = reconstructed_file.read_text(encoding='utf-8')
        assert "Tëst Dócümënt" in reconstructed_content
        assert "🚀 📝 ✅" in reconstructed_content
        assert "hëllö_wörld" in reconstructed_content

    def test_large_document_roundtrip(self):
        """Test roundtrip with a large document."""
        # Generate large content
        large_content = "# Large Document Test\n\nThis tests performance with large documents.\n\n"

        for chapter in range(1, 11):  # 10 chapters
            large_content += f"# Chapter {chapter}\n\n"
            large_content += f"This is chapter {chapter} content.\n\n"

            for section in range(1, 6):  # 5 sections per chapter
                large_content += f"## Section {chapter}.{section}\n\n"
                large_content += f"Content for section {chapter}.{section}.\n\n"
                large_content += "Lorem ipsum dolor sit amet, consectetur adipiscing elit. " * 20
                large_content += "\n\n"

        large_content += "# Conclusion\n\nEnd of large document.\n"

        original_file = self.temp_dir / "large_doc.md"
        original_file.write_text(large_content, encoding='utf-8')

        # Test with hierarchical variant (most complex)
        result = self.run_markitect_command([
            "md-explode", str(original_file),
            "--variant", "hierarchical"
        ])
        assert result.returncode == 0

        exploded_dir = original_file.with_suffix('.mdd')
        assert exploded_dir.exists()

        # Verify many files were created
        md_files = list(exploded_dir.glob("**/*.md"))
        assert len(md_files) > 10  # Should have many files

        # Implode back
        reconstructed_file = self.temp_dir / "large_reconstructed.md"
        result = self.run_markitect_command([
            "md-implode", str(exploded_dir),
            "--output", str(reconstructed_file)
        ])
        assert result.returncode == 0

        # Verify structure preservation
        reconstructed_content = reconstructed_file.read_text(encoding='utf-8')
        validation = self.validate_basic_structure_preservation(large_content, reconstructed_content)
        assert validation['headings_preserved']


class TestErrorHandlingRoundtrips(TestRoundtripBase):
    """Test error handling in roundtrip scenarios."""

    def test_malformed_markdown_handling(self):
        """Test handling of malformed markdown."""
        malformed_content = """# Valid Header

Some content here.

## Another header

# Missing spacing
No space before content.

###Too many hashes without space

# Final header
"""

        original_file = self.temp_dir / "malformed.md"
        original_file.write_text(malformed_content, encoding='utf-8')

        # Should still work despite malformed content
        result = self.run_markitect_command([
            "md-explode", str(original_file)
        ])
        assert result.returncode == 0

        exploded_dir = original_file.with_suffix('.mdd')
        assert exploded_dir.exists()

        # Should be able to implode back
        result = self.run_markitect_command([
            "md-implode", str(exploded_dir)
        ])
        assert result.returncode == 0

    def test_empty_content_handling(self):
        """Test handling of empty files and sections."""
        empty_content = """# Empty Test

## Empty Section 1

## Empty Section 2

# Another Empty

"""

        original_file = self.temp_dir / "empty.md"
        original_file.write_text(empty_content, encoding='utf-8')

        # Should handle empty content gracefully
        result = self.run_markitect_command([
            "md-explode", str(original_file)
        ])
        assert result.returncode == 0

        exploded_dir = original_file.with_suffix('.mdd')
        assert exploded_dir.exists()

        result = self.run_markitect_command([
            "md-implode", str(exploded_dir)
        ])
        assert result.returncode == 0


if __name__ == '__main__':
    pytest.main([__file__, "-v"])