rename-simple

A small Rust CLI tool that renames files and directories to clean, ASCII-safe slugs.

Warning

Breaking change since 0.4.0. The directory-scan mode and the -a/--all and -r/--recursive options have been removed. rename-simple now operates only on the paths you give it, like rename(1). Use your shell's globbing to select entries: rename-simple * instead of rename-simple -a, and rename-simple **/*.pdf instead of -r.

What it does

• Transliterates accented and extended Latin characters to ASCII (é → e, ç → c, œ → oe, ß → ss…)
• Expands typographic ligatures and compatibility forms (fi → fi, fullwidth Ｆｉｌｅ → file, superscripts ² → 2, ™ → tm, № → no)
• Romanises Greek and Cyrillic letters (Ελληνικά → ellinika, Москва → moskva)
• Lowercases everything
• Replaces spaces and special characters with -; collapses consecutive separators
• Preserves _; cleans up _- and -_ sequences to _
• Strips leading and trailing - / _ before the extension
• Preserves known compound extensions (.tar.gz, .tar.bz2, .tar.xz, .tar.zst)
• Keeps extensions separate only when they are ASCII alphanumeric and ≤10 characters (e.g. .tét → absorbed as -tet; .abcdefghijkl (12 chars) → absorbed as -abcdefghijkl)
• Skips hidden files (.gitignore, .DS_Store…) and flags naming conflicts
• Optionally (-D) deletes a source file that is a byte-for-byte duplicate of an already existing target
• Optional cleanup fixes: repairs mojibake (-U), strips HTML tags and entities (-H), or both (-A)

Installation

Requires Rust 1.85+.

# From crates.io
cargo install rename-simple

# Or from source
cargo install --path .

With Nix

The repository is a flake, so no Rust toolchain is needed:

# Run without installing
nix run github:darkone-linux/rename-simple

# Or install into your profile
nix profile install github:darkone-linux/rename-simple

Usage

rename-simple [OPTIONS] <files>...

Each argument is an entry renamed itself, like the traditional rename command — globbing is left to the shell (rename-simple *.jpg, or rename-simple dir/**/*.pdf with zsh / bash globstar). Files and directories are both renamed by default; -f / -d act as a type filter. With no arguments, rename-simple prints this help.

Option	Description
<files>...	Entries to rename (files and/or directories)
-f, --files-only	Rename files only
-d, --dirs-only	Rename directories only
-U, --fix-unicode	Repair mojibake (UTF-8 misread as Latin-1/CP1252) before renaming
-H, --fix-html	Strip HTML tags and decode HTML entities before renaming
-A, --fix-all	Apply every cleanup fix (currently -U + -H)
-D, --delete-duplicates	Delete a source that is a byte-for-byte duplicate of an existing target
-n, --dry-run	Preview renames without touching any file
-q, --quiet	Print nothing at all
-v, --verbose	Show details of each rename
-h, --help	Print help
-V, --version	Print version

Examples

Preview all renames (dry-run)

$ rename-simple --dry-run ~/Downloads/*

[R] .: 01_ Introduction au Projet.PDF -> 01_introduction-au-projet.pdf
[R] .: Réunion d'équipe (2024).docx -> reunion-d-equipe-2024.docx
[R] .: backup.TAR.GZ -> backup.tar.gz
[R] .: Café Montréal.jpg -> cafe-montreal.jpg
[R] .: à faire .tét -> a-faire-tet
[R] .: notes.cuicuicuicui -> notes-cuicuicuicui
6 entries matched, 6 entries renamed, 0 error.

Rename files only

$ rename-simple -f ~/Downloads/*

Directories among the arguments are left untouched; only files are renamed.

Rename directories only

$ rename-simple -d ~/Projects/*

Files among the arguments are left untouched; only directories are renamed.

Rename everything

$ rename-simple ~/Downloads/*

Process a whole tree

With a shell that supports recursive globs you can select entries at any depth:

$ rename-simple ~/Documents/**/*.pdf

Repair damaged names (mojibake, HTML)

Names coming from broken downloads or web scrapers often carry mojibake (UTF-8 read as Latin-1/CP1252) or HTML markup. The cleanup fixes repair the name before the slug pipeline runs:

$ rename-simple -A ~/Downloads/*

[R] .: Café Montréal.jpg -> cafe-montreal.jpg
[R] .: Tom & Jerry.mp4 -> tom-jerry.mp4
[R] .: <b>Été 2024.pdf -> ete-2024.pdf
3 entries matched, 3 entries renamed, 0 error.

-U/--fix-unicode is all-or-nothing per name: a file that is already correctly named (café.jpg) is never re-decoded, so the fix is safe to apply everywhere. Double mojibake (Café) is repaired too. -H/--fix-html strips tags and decodes named (é), decimal (é) and hex (é) entities. Block-level tags (<br>, <p>, <div>, <li>, <h1>–<h6>, …) are replaced by a space so the surrounding words stay separated (data<br>client → data-client), while inline tags (<b>, <i>, <span>, …) vanish with no gap (client<b>s → clients). -A/--fix-all applies every cleanup fix.

Conflicts and duplicates

When the cleaned name is already taken, the source is never overwritten and the clash is reported. The message tells you whether the two entries hold the same bytes:

$ rename-simple *

[E] .: IMG 9572.mov -> File name already exists (identical duplicate, use -D to delete the source)
[E] .: VID 20260819.mp4 -> File name already exists (2 different files)

-D/--delete-duplicates acts on the first case: when source and target are two regular files with strictly identical content (compared byte for byte, not by name or size), the rename would only produce a copy of what is already there, so the redundant source is deleted:

$ rename-simple -D *

[W] .: IMG 9572.mov -> Identical duplicate, source deleted
[E] .: VID 20260819.mp4 -> File name already exists (2 different files)
4 entries matched, 0 entry renamed, 1 duplicate removed, 1 error.

Nothing is ever deleted without -D, and -A/--fix-all does not imply it: -A only repairs names. -D combined with --dry-run announces the deletion without performing it. Everything else stays an error and is left untouched: differing content, a directory, or two names for the same entry (hard link, case-insensitive filesystem).

Symlinks are left out of the comparison entirely — neither side may be one. A link is not a redundant copy of the bytes it points at, and content reached through a symlinked target lives outside the directory being cleaned up: deleting the source there would drop the last real copy.

Verbose output

Verbose mode prints each rename and a summary:

$ rename-simple -v ~/Downloads/*

[R] .: 01_ Introduction au Projet.PDF -> 01_introduction-au-projet.pdf
[R] .: Réunion d'équipe (2024).docx -> reunion-d-equipe-2024.docx
[R] .: backup.TAR.GZ -> backup.tar.gz
[R] .: Café Montréal.jpg -> cafe-montreal.jpg
[R] .: à faire .tét -> a-faire-tet
[R] .: notes.cuicuicuicui -> notes-cuicuicuicui
6 entries matched, 6 entries renamed, 0 error.

Tips

Since the directory-scan mode is gone, "rename everything in the current directory" is now just rename-simple *. A shell alias makes it a habit:

alias rsa='rename-simple *'

Add this line to your ~/.bashrc or ~/.zshrc, then:

rsa                      # rename everything in the current directory
rename-simple -n *       # dry-run preview
rename-simple **/*.pdf   # rename every PDF in the tree (zsh / bash globstar)

Running the tests

cargo test

License

MIT