Ricardo Decal's Tools

Simple, self-contained, and immediately runnable

A collection of simple, self-contained tools in two flavors:

Scripts — Python CLI tools runnable with uv run https://tools.ricardodecal.com/python/foo.py, which spawns a self-contained and ephemeral uv environment.
Pages — Single-file HTML tools that run entirely in your browser at https://tools.ricardodecal.com/html/foo.html.

This is an experiment in low-stakes vibe coding. The code lives in crypdick/tools.

Inspired by Simon Willison's tools collection.

Available Tools

📊 Data Processing (7 scripts, 2 pages)

convert_arrow_to_parquet_streaming.py python

Output of uv run https://tools.ricardodecal.com/python/convert_arrow_to_parquet_streaming.py --help:


 Usage: convert_arrow_to_parquet_streaming.py [OPTIONS]

 Convert Arrow shards to Parquet.

 - Discovers all .arrow files under a given source directory
 - Converts each file to Parquet
 - Uses streaming in order to keep memory bounded and convert files larger than
 available RAM
 - Handles both Arrow IPC File and Stream formats (tries file, falls back to
 stream)

 Notes:
 - Use --preserve-subdirs to mirror the input directory tree under the output
 dir.
 - Use --overwrite to re-create files; otherwise existing outputs are skipped.

 Arguments:
     SOURCE_DIR: Directory containing .arrow files.
     OUTPUT_DIR: Directory to write .parquet files.

 Examples:

     uv run https://tools.ricardodecal.com/python/convert_arrow_to_parquet_streaming.py --source-dir
 ./arrow_data --output-dir ./parquet_data

     uv run https://tools.ricardodecal.com/python/convert_arrow_to_parquet_streaming.py --source-dir
 ./arrow_data --output-dir ./parquet_data --preserve-subdirs

╭─ Options ────────────────────────────────────────────────────────────────────╮
│ *  --source-dir              DIRECTORY  Directory containing .arrow files    │
│                                         [required]                           │
│    --output-dir              DIRECTORY  Directory to write .parquet files    │
│                                         [default: parq_convert]              │
│    --overwrite                          Overwrite existing parquet files     │
│    --preserve-subdirs                   Preserve input subdirectory          │
│                                         structure inside output dir          │
│    --help                               Show this message and exit.          │
╰──────────────────────────────────────────────────────────────────────────────╯

count_parquet_rows.py python

Output of uv run https://tools.ricardodecal.com/python/count_parquet_rows.py --help:


 Usage: count_parquet_rows.py [OPTIONS] DATASET_PATH

 Count the number of rows in a parquet file/dataset without reading data into
 memory.

 Works by reading just the metadata headers. Supports:
 - Single parquet files
 - Directories of parquet shards
 - Hive-style partitioned datasets
 - Local paths and S3 URIs

 Arguments:
     DATASET_PATH: Local file path or S3 URI to the parquet dataset.

 Examples:

     # Local file
     uv run https://tools.ricardodecal.com/python/count_parquet_rows.py
 ./data.parquet

     # Directory of shards
     uv run https://tools.ricardodecal.com/python/count_parquet_rows.py
 ./data_dir/

     # S3 URI
     uv run https://tools.ricardodecal.com/python/count_parquet_rows.py
 s3://my-bucket/data.parquet

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    dataset_path      TEXT  Local file path or S3 URI. [required]           │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --help          Show this message and exit.                                  │
╰──────────────────────────────────────────────────────────────────────────────╯

dedup_dirs.py python

Output of uv run https://tools.ricardodecal.com/python/dedup_dirs.py --help:


 Usage: dedup_dirs.py [OPTIONS] OLD_DIR NEW_DIR

 Find duplicate files between OLD_DIR and NEW_DIR, optionally deleting from
 OLD_DIR.

 Compares files by path and content. For large files (>10MB), uses sampling
 for speed. For smaller files, compares MD5 hashes. Runs in parallel for
 performance on large directory trees.

 Without --delete, runs in dry-run mode showing what would be deleted.
 With --delete, removes identical files from OLD_DIR and cleans up empty
 directories.

 Arguments:
     OLD_DIR: Source directory to deduplicate (files deleted from here).
     NEW_DIR: Reference directory to compare against (untouched).

 Examples:

     # Dry run - see what would be deleted
     uv run https://tools.ricardodecal.com/python/dedup_dirs.py ~/old-backup
 ~/new-backup

     # Actually delete duplicates
     uv run https://tools.ricardodecal.com/python/dedup_dirs.py ~/old-backup
 ~/new-backup --delete

     # Use more workers for faster processing
     uv run https://tools.ricardodecal.com/python/dedup_dirs.py ~/old ~/new
 --delete -w 16

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    old_dir      PATH  Source directory to deduplicate (files deleted from  │
│                         here).                                               │
│                         [required]                                           │
│ *    new_dir      PATH  Reference directory to compare against (untouched).  │
│                         [required]                                           │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --delete                    Actually delete identical files (default is dry  │
│                             run).                                            │
│ --workers  -w      INTEGER  Number of parallel workers. [default: 8]         │
│ --help                      Show this message and exit.                      │
╰──────────────────────────────────────────────────────────────────────────────╯

download_video.py python

Output of uv run https://tools.ricardodecal.com/python/download_video.py --help:


 Usage: download_video.py [OPTIONS] URL

 Download a video from a supported platform (Twitter/X, YouTube, etc.).

 Uses yt-dlp to download videos from a wide variety of websites.
 Twitter "GIFs" are actually MP4 videos, which this tool can also download.

 Arguments:

     URL: The URL of the video page (e.g., Twitter post, YouTube video).

 Examples:

     uv run https://tools.ricardodecal.com/python/download_video.py
 https://x.com/SemiAnalysis_/status/1990449859321888935

     uv run https://tools.ricardodecal.com/python/download_video.py
 https://www.youtube.com/watch?v=dQw4w9WgXcQ --output my_video.mp4

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    url      TEXT  The URL of the video page. [required]                    │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --output  -o      TEXT  Output filepath (file or directory). Defaults to     │
│                         'Title [ID].mp4' in current directory.               │
│ --help                  Show this message and exit.                          │
╰──────────────────────────────────────────────────────────────────────────────╯

html_to_text.html page

https://tools.ricardodecal.com/html/html_to_text.html
Paste HTML and extract readable plain text. Runs entirely in your browser.

html_to_text.py python

Output of uv run https://tools.ricardodecal.com/python/html_to_text.py --help:


 Usage: html_to_text.py [OPTIONS] URL

 Fetch a webpage and convert its readable content to plain text.

 Uses Readability-style boilerplate removal by default, with conservative
 fallbacks.
 Then uses inscriptis to render HTML to text while preserving basic structure.
 Automatically adds https:// if no scheme is provided.

 Arguments:

     URL: The webpage URL to fetch (e.g., example.com or https://example.com).

 Examples:

     uv run https://tools.ricardodecal.com/python/html_to_text.py example.com

     uv run https://tools.ricardodecal.com/python/html_to_text.py
 https://news.ycombinator.com --timeout 30

     uv run https://tools.ricardodecal.com/python/html_to_text.py
 wikipedia.org/wiki/Python --raw

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    url      TEXT  The webpage URL to fetch. [required]                     │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --timeout    -t      INTEGER                     Request timeout in seconds. │
│                                                  [default: 15]               │
│ --mode       -m      [auto|readability|heuristi  Extraction mode: auto       │
│                      c|full]                     (default), readability      │
│                                                  (boilerplate removal),      │
│                                                  heuristic (semantic tags),  │
│                                                  or full (whole page).       │
│                                                  [default: auto]             │
│ --min-chars          INTEGER RANGE [x>=0]        Minimum extracted text      │
│                                                  length to accept before     │
│                                                  falling back (auto mode).   │
│                                                  [default: 200]              │
│ --min-ratio          FLOAT RANGE [0.0<=x<=1.0]   Minimum extracted/full text │
│                                                  length ratio to accept      │
│                                                  before falling back (auto   │
│                                                  mode).                      │
│                                                  [default: 0.2]              │
│ --raw                                            Skip whitespace cleanup     │
│                                                  (preserve original          │
│                                                  formatting).                │
│ --help                                           Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯

strip_pdf_metadata.html page

https://tools.ricardodecal.com/html/strip_pdf_metadata.html
Remove author, title, timestamps, and other metadata from PDF files. Runs entirely in your browser — files never leave your device.

strip_pdf_metadata.py python

Output of uv run https://tools.ricardodecal.com/python/strip_pdf_metadata.py --help:


 Usage: strip_pdf_metadata.py [OPTIONS] INPUT_FILE [OUTPUT_FILE]

 Strip metadata from a PDF file.

 If OUTPUT_FILE is not provided, writes to 'stripped_<INPUT_FILE>'.

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    input_file       FILE           Input PDF file. [required]              │
│      output_file      [OUTPUT_FILE]  Output PDF file. Defaults to            │
│                                      'stripped_<INPUT_FILE>'.                │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --help          Show this message and exit.                                  │
╰──────────────────────────────────────────────────────────────────────────────╯

yt_transcript.py python

Output of uv run https://tools.ricardodecal.com/python/yt_transcript.py --help:


 Usage: yt_transcript.py [OPTIONS] URL [OUTPUT_FILE]

 Download transcripts from a YouTube URL (video or playlist) to a single file.

 Arguments:
     URL: YouTube video or playlist URL.
     OUTPUT_FILE: Path to save the transcript text. Defaults to transcript.txt.

 Examples:

     uv run https://tools.ricardodecal.com/python/yt_transcript.py "https://youtu.be/..."

     uv run https://tools.ricardodecal.com/python/yt_transcript.py "https://youtube.com/playlist?list=..."
 out.txt

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    url              TEXT           YouTube video or playlist URL.          │
│                                      [required]                              │
│      output_file      [OUTPUT_FILE]  Path to save the transcript text.       │
│                                      Defaults to transcript.txt.             │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --lang  -l      TEXT  Language codes to prefer (e.g. -l en -l fr)            │
│ --help                Show this message and exit.                            │
╰──────────────────────────────────────────────────────────────────────────────╯

🛠️ Development (2 scripts)

burn_iso.py python

Output of uv run https://tools.ricardodecal.com/python/burn_iso.py --help:


 Usage: burn_iso.py [OPTIONS] ISO_FILE

 Burn an ISO file to a USB drive with safety checks and progress monitoring.

 This tool will verify the ISO file, check the target device, unmount any
 mounted partitions, and burn the ISO using dd with progress reporting.

 DANGER: This will completely erase all data on the target USB device!

 Arguments:

     ISO_FILE: Path to the ISO file to burn.

 Examples:

     uv run https://tools.ricardodecal.com/python/burn_iso.py --list

     uv run https://tools.ricardodecal.com/python/burn_iso.py ubuntu.iso
 --device /dev/sdb --dry-run

     uv run https://tools.ricardodecal.com/python/burn_iso.py
 ~/Downloads/debian.iso -d /dev/sdc

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    iso_file      FILE  Path to the ISO file to burn. [required]            │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --device   -d      TEXT  USB device path (e.g., /dev/sdb). Use --list to see │
│                          available devices.                                  │
│ --list                   List available block devices and exit.              │
│ --dry-run  -n            Show what would be done without actually doing it.  │
│ --force    -f            Skip confirmation prompts.                          │
│ --help                   Show this message and exit.                         │
╰──────────────────────────────────────────────────────────────────────────────╯

ipynb_to_py_sphinx.py python

Output of uv run https://tools.ricardodecal.com/python/ipynb_to_py_sphinx.py --help:


 Usage: ipynb_to_py_sphinx.py [OPTIONS] NOTEBOOK

 Convert a Jupyter notebook to a Sphinx Gallery Python script.

 This tool converts a .ipynb file to a .py file formatted for Sphinx Gallery.
 It converts Markdown cells to RST (using pypandoc) and comments them out,
 while preserving code cells. It also handles magic commands by commenting them
 out.

 Based on: https://gist.github.com/chsasank/7218ca16f8d022e02a9c0deb94a310fe

 Arguments:
     NOTEBOOK: The path to the input Jupyter notebook (.ipynb).

 Examples:

     uv run https://tools.ricardodecal.com/python/ipynb_to_py_sphinx.py
 notebook.ipynb

     uv run https://tools.ricardodecal.com/python/ipynb_to_py_sphinx.py
 notebook.ipynb --output my_gallery_script.py

╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ *    notebook      PATH  The path to the input Jupyter notebook (.ipynb).    │
│                          [required]                                          │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --output  -o      PATH  Output Python file path. Defaults to notebook name   │
│                         with .py extension.                                  │
│ --help                  Show this message and exit.                          │
╰──────────────────────────────────────────────────────────────────────────────╯