Any2MD-1StepMore 0.3.0

Universal document to Markdown converter with quality/fast dual-mode pipeline

Any2MD - 格式工厂分发引擎

Universal document to Markdown converter with quality/fast dual-mode pipeline

Any2MD converts documents to Markdown with two processing modes:

• Quality Mode: Uses specialized converters (pypandoc, mammoth, markdownify) for superior output
• Fast Mode: Uses markitdown for quick "good enough" results

Features

• Dual-Mode Pipeline: Quality vs Fast mode for optimal results
• PDF Heavy Channel: MarkItDown (primary) + pypdfium2 (fallback)
• Async Batch Processing: Concurrent file conversion with Semaphore control
• Smart Retry: Tenacity-based retry for transient failures
• Non-Destructive Output: Auto-incremented filenames, never overwrites
• Organized Output: Folder batch outputs to {folder}_converted/
• BAT Drag-and-Drop: Zero-config usage on Windows

Quick Start

Installation

# From source (development)
git clone https://github.com/1StepMore/Any2MD.git
cd Any2MD
pip install -e .

# Or install from PyPI (when available)
pip install any2md

Usage

Drag-and-Drop (Windows)

1. Drag a file onto bat\run.bat
2. Done! Markdown appears next to original

Drag-and-Drop Folder (Windows)

1. Drag a folder onto bat\run.bat
2. Done! Markdown files appear in {folder}_converted/

Command Line

# Single file conversion
python cli.py -i document.pdf

# With quality mode (better output)
python cli.py -i document.docx --mode quality

# Batch folder processing
python cli.py -i ./documents --concurrency 4

# Verbose logging
python cli.py -i document.pdf --verbose

# Full options
python cli.py --help

Python API

from any2md import convert_to_markdown

# Basic usage
md = convert_to_markdown("document.pdf")
print(md)

# With options
md = convert_to_markdown("document.docx", mode="quality")
md = convert_to_markdown("document.pdf", mode="fast", pdf_engine="heavy")

# Error handling
from any2md import (
    convert_to_markdown,
    FileTooLargeError,
    UnsupportedFormatError,
    ConversionError,
)

try:
    md = convert_to_markdown("document.pdf")
except FileTooLargeError:
    print("File too large (max 50MB)")
except UnsupportedFormatError:
    print("Format not supported")
except ConversionError as e:
    print(f"Conversion failed: {e}")

CLI Options

Option	Short	Description	Default
--input	-i	Input file or folder	Required
--output	-o	Output directory	Same as input
--mode	-	quality or fast	From config
--pdf-engine	-	light (markitdown) or heavy (MarkItDown)	From config
--concurrency	-	Max parallel conversions	From config
--config	-	Config file path	config.yaml
--verbose	-v	Enable debug logging	false

Configuration

Edit config.yaml:

# Output mode: quality (best) or fast (quick)
output_mode: fast

# PDF engine: light (markitdown) or heavy (MarkItDown -> pypdfium2)
pdf_engine: light

# Max file size (MB) - files larger are skipped
max_file_size: 50

# Async concurrency level
concurrency: 4

# Retry attempts for transient failures
retry_count: 3

Tip: You can also access config programmatically via from wheels.config import get_config

Architecture

Any2MD/
├── bat/
│   └── run.bat              # Windows drag-and-drop entry
├── wheels/
│   ├── converters/         # Format-specific converters
│   │   ├── converter_pandoc.py   # Quality: docx/pptx/xlsx
│   │   ├── converter_mammoth.py   # Quality: complex DOCX
│   │   ├── converter_html.py      # Quality: HTML
│   │   ├── converter_pdf.py       # Dynamic: PDF (light/heavy)
│   │   └── converter_passthrough.py # .md/.txt passthrough
│   │   └── converter_markitdown.py # Fast: csv/json/xml/yaml/epub/zip
│   ├── dispatcher.py        # Format routing
│   ├── fast_lane.py        # markitdown wrapper
│   ├── cleaner.py           # Text post-processing
│   └── logger.py           # Logging configuration
├── cli.py                  # Typer CLI entry
├── pipeline.py              # Async batch orchestration
├── config.yaml             # Configuration
└── requirements.txt        # Dependencies

Supported Formats

Format	Quality Mode	Fast Mode
.docx	pypandoc → mammoth fallback	markitdown
.pptx / .xlsx	pypandoc	markitdown
.html / .htm	markdownify	markitdown
.pdf	MarkItDown (heavy) or markitdown (light)	markitdown
.md / .txt	passthrough	passthrough
.csv / .json / .xml	-	markitdown
.yaml / .yml	-	markitdown
.epub	-	markitdown
.zip	-	markitdown

PDF Engine

Light Mode: Uses markitdown with zero external dependencies.

Heavy Mode: Uses Microsoft MarkItDown for table/layout detection, falls back to pypdfium2 for raw text extraction.

Install heavy dependencies:

# Ubuntu/Debian
sudo apt install tesseract-ocr poppler-utils

# Windows (requires admin)
winget install --id UB-Mannheim.TesseractOCR -e
winget install --id oschwartz10612.Poppler -e

Error Handling

• File too large: Files >50MB are skipped with error log
• Unsupported format: Clear error message with supported formats
• Missing dependencies: Install instructions in error message
• Transient failures: Automatic retry (up to 3 attempts)
• Batch continuation: Single file failure doesn't stop batch processing

Development

# Run CLI help
python cli.py --help

# Test with sample files
python cli.py -i sample.docx --mode quality

# Check environment
python check_env.py

License

MIT License - see LICENSE file for details

Credits

Built with these excellent open-source libraries:

• pypandoc - Pandoc wrapper
• mammoth - DOCX to Markdown
• markdownify - HTML to Markdown
• markitdown - Microsoft format converter
• pypdfium2 - PDF rendering
• tenacity - Retry logic
• Typer - CLI framework