file-converter

name	file-converter
description	Convert & transform files - images (resize, format, HEIC), markdown (PDF/HTML), data (CSV/JSON/YAML/TOML/XML), SVG, base64, text encoding. Cross-platform, single & batch mode. This skill should be used when converting file formats, resizing images, generating PDFs from markdown, or transforming data between formats.
author	George Khananaev

File Converter

Convert files between formats w/ single & batch support. All scripts use consistent CLI patterns.

When to Use

• Convert images between PNG, JPG, WEBP, BMP, TIFF, GIF, ICO, AVIF, HEIC/HEIF
• Resize/crop images w/ fit modes (contain, cover, fill, inside, outside)
• Convert markdown -> PDF or HTML w/ themes
• Convert HTML -> markdown (w/ tag stripping control)
• Transform CSV <-> JSON <-> YAML <-> TOML <-> XML
• SVG <-> raster conversion (PNG, JPG, WEBP, BMP, TIFF)
• Base64 encode/decode files (w/ data URI support, stdin)
• Fix text encoding issues (detect, convert w/ error handling strategies)

Quick Routing

Task	Script	Deps
Image convert/resize	convert_image.py	Pillow, pillow-heif (opt)
Markdown -> HTML	md_to_html.py	markdown, pygments
Markdown -> PDF	md_to_pdf.py	markdown + weasyprint|pdfkit
HTML -> Markdown	html_to_md.py	markdownify, bs4
CSV/JSON/YAML/TOML/XML	csv_json_yaml.py	pyyaml, tomli-w, xmltodict, dicttoxml (per format)
SVG convert	svg_convert.py	cairosvg, Pillow
Base64 encode/decode	base64_codec.py	(none)
Text encoding	text_encoding.py	chardet (opt)
Cross-platform utils	platform_utils.py	(none) - shared by pdf/svg scripts

Install Deps

# All deps (recommended)
pip install Pillow markdown pygments weasyprint markdownify beautifulsoup4 cairosvg pyyaml chardet tomli-w xmltodict dicttoxml

# Optional
pip install pillow-heif                    # HEIC/HEIF support

# Minimal (per task)
pip install Pillow                          # Images only
pip install markdown pygments               # MD -> HTML only
pip install markdown weasyprint             # MD -> PDF only (macOS: brew install pango)
pip install markdownify beautifulsoup4      # HTML -> MD only
pip install pyyaml                          # YAML support
pip install tomli-w                         # TOML write (read: Python 3.11+ built-in)
pip install xmltodict dicttoxml             # XML support
pip install cairosvg                        # SVG -> raster (macOS: brew install cairo)
pip install chardet                         # Encoding detection

CLI Patterns

All scripts share consistent arg patterns:

# Single file
python3 scripts/<script>.py input.ext output.ext

# Batch (directory or glob)
python3 scripts/<script>.py *.ext --output-dir ./out
python3 scripts/<script>.py ./dir/ --output-dir ./out --format ext

1. Image Convert & Resize

# Format conversion
python3 scripts/convert_image.py photo.png photo.webp
python3 scripts/convert_image.py photo.jpg photo.avif --quality 80

# Resize
python3 scripts/convert_image.py photo.jpg thumb.jpg --width 300
python3 scripts/convert_image.py photo.png banner.png --width 1200 --height 400 --fit cover

# Batch convert directory
python3 scripts/convert_image.py ./photos/ --output-dir ./webp --format webp --width 1200 --quality 85
python3 scripts/convert_image.py *.png --output-dir ./thumbs --format jpg --width 300 --height 300 --fit cover

Fit modes:

Mode	Behavior
contain	Fit inside bounds, preserve ratio (def)
cover	Fill bounds, crop overflow
fill	Stretch to exact dimensions
inside	Like contain, but only shrink (never enlarge)
outside	Like cover, but never crop

Supported: PNG, JPG, WEBP, BMP, TIFF, GIF, ICO, AVIF, HEIC/HEIF (w/ pillow-heif)

Auto-fixes EXIF orientation. Guards against decompression bombs (300M pixel limit).

2. Markdown -> HTML

# Single file
python3 scripts/md_to_html.py README.md readme.html
python3 scripts/md_to_html.py doc.md doc.html --theme dark

# Batch
python3 scripts/md_to_html.py ./docs/ --output-dir ./site --theme github

Themes: github (def), dark, minimal, print

Features: fenced code blocks, syntax highlighting, tables, TOC w/ permalinks, responsive CSS.

3. Markdown -> PDF

# Single file
python3 scripts/md_to_pdf.py report.md report.pdf
python3 scripts/md_to_pdf.py spec.md spec.pdf --theme report

# Batch
python3 scripts/md_to_pdf.py ./docs/ --output-dir ./pdfs --theme report

Themes: default, report (formal w/ serif), minimal

PDF engines: weasyprint (preferred, no external deps on macOS) or pdfkit (requires wkhtmltopdf). Script auto-detects available engine.

4. HTML -> Markdown

# Single file
python3 scripts/html_to_md.py page.html page.md

# Strip unwanted tags (default: script, style, noscript)
python3 scripts/html_to_md.py page.html page.md --strip script style nav footer

# Keep all HTML tags (no stripping)
python3 scripts/html_to_md.py page.html page.md --keep-all

# Batch
python3 scripts/html_to_md.py ./site/ --output-dir ./docs

5. Data Formats (CSV/JSON/YAML/TOML/XML)

# Any direction
python3 scripts/csv_json_yaml.py data.csv data.json
python3 scripts/csv_json_yaml.py data.json data.yaml
python3 scripts/csv_json_yaml.py config.yaml config.json
python3 scripts/csv_json_yaml.py config.toml config.json
python3 scripts/csv_json_yaml.py data.json data.xml

# Batch
python3 scripts/csv_json_yaml.py *.csv --output-dir ./json --format json

Supported: CSV, JSON, YAML (.yaml/.yml), TOML (.toml), XML (.xml). All directions supported where deps are installed.

6. SVG Conversion

# SVG -> raster
python3 scripts/svg_convert.py icon.svg icon.png --width 512
python3 scripts/svg_convert.py logo.svg logo.jpg --width 1024 --quality 90

# Raster -> SVG (embedded image wrapper)
python3 scripts/svg_convert.py photo.png photo.svg

# Batch
python3 scripts/svg_convert.py *.svg --output-dir ./png --format png --width 256

7. Base64 Encode/Decode

# Encode to stdout
python3 scripts/base64_codec.py encode image.png

# Encode to data URI (for HTML/CSS embedding)
python3 scripts/base64_codec.py encode image.png --data-uri

# Encode to file
python3 scripts/base64_codec.py encode image.png -o image.b64

# Decode
python3 scripts/base64_codec.py decode image.b64 -o image.png

# Batch
python3 scripts/base64_codec.py encode *.png --output-dir ./b64

8. Text Encoding

# Detect encoding
python3 scripts/text_encoding.py detect file.txt
python3 scripts/text_encoding.py detect *.txt

# Convert encoding (single mode requires -o to prevent accidental overwrite)
python3 scripts/text_encoding.py convert file.txt --to utf-8 -o output.txt
python3 scripts/text_encoding.py convert file.txt --from latin-1 --to utf-8 -o output.txt

# Handle unmappable characters
python3 scripts/text_encoding.py convert file.txt --to ascii --errors replace -o clean.txt
python3 scripts/text_encoding.py convert file.txt --to ascii --errors ignore -o clean.txt

# Batch convert to UTF-8
python3 scripts/text_encoding.py convert *.txt --to utf-8 --output-dir ./utf8

Error modes: strict (default, fail on unmappable), replace (use ? placeholder), ignore (skip unmappable chars).

Common Workflows

Web optimization pipeline

# Convert photos to WEBP, resize for web, generate thumbnails
python3 scripts/convert_image.py ./photos/ --output-dir ./web --format webp --width 1200 --quality 80
python3 scripts/convert_image.py ./photos/ --output-dir ./thumbs --format webp --width 300 --height 300 --fit cover --quality 75

Documentation pipeline

# Generate HTML site from markdown docs
python3 scripts/md_to_html.py ./docs/ --output-dir ./site --theme github

# Generate PDF reports
python3 scripts/md_to_pdf.py ./docs/ --output-dir ./pdfs --theme report

Data migration pipeline

# CSV -> JSON for API import
python3 scripts/csv_json_yaml.py ./exports/ --output-dir ./json --format json

# JSON config -> YAML
python3 scripts/csv_json_yaml.py config.json config.yaml

Icon generation pipeline

# SVG -> multiple PNG sizes for app icons
for size in 16 32 64 128 256 512; do
  python3 scripts/svg_convert.py icon.svg "icon-${size}.png" --width $size
done

Error Handling

All scripts:

• Print errors per-file in batch mode, continue w/ remaining files
• Exit 1 on fatal errors (missing deps, no input)
• Print size before/after for each conversion
• Create output directories automatically
• Handle KeyboardInterrupt gracefully (exit 130)

Cross-Platform Support

Scripts work on macOS, Linux, and Windows. Native library paths (cairo, pango, gobject) are auto-configured via platform_utils.py:

• macOS: /opt/homebrew/lib, /usr/local/lib
• Linux: /usr/local/lib, /usr/lib/x86_64-linux-gnu
• Windows: GTK runtime, MSYS2, Conda paths + os.add_dll_directory()

Stdin Support

md_to_html.py and base64_codec.py accept - for stdin input:

cat README.md | python3 scripts/md_to_html.py - output.html
cat file.bin | python3 scripts/base64_codec.py encode - -o file.b64

Integration

Pairs with: token-optimizer (compress markdown before PDF), code-quality (validate scripts)