fr3

Performs freq-analysis of "words", on directories and files.

Most of the code has been copy-pasted from myself, lol

Usage

Install

This needs a Rust toolchain. Recommended command:

cargo install --path . --config 'build.rustflags="-C target-cpu=native"'

Assuming you've downloaded and cded into the repo

Run

Invoke the program by passing the paths whose stats you want to get:

# example
fr3 file.txt directory/
file.txt
	test 3
	bruh 1
directory/
	the 64
	wiki 12
	yeet_3 6
# non-UTF8 paths are supported

Or simply pass nothing, if you want stats about WD, identical to fr3 ..

Important

The default regex is for human use only. It'll always be API-unstable, as it can change across patch versions and maybe even across runs of the same version!

Note

Tie-breaking is unspecified. That is, if 2 or more words have the same count, they will be printed in arbitrary order within their partition.

I'm considering to lexicographically sort ties

You can define what a "word" is by passing a regex:

fr3 --re '[^,\n\r]+' table.csv
table.csv
	my record 3
	joe 2
	zayda 1
	1 1
	0 1
	2 1
	id 1
	name 1
	column 1

Prepend (?i) for case-insensitive regex. Note that counting is always case-sensitive:

fr3 --re '(?i)\bthe\b' prose.md
prose.md
	the 6
	The 3
	THE 1

The output format, while rudimentary, is mostly unambiguous. The only way (that I know) it can be ambiguous, is if your words contain \n, but that's just asking for trouble, so please let me pretend \n doesn't exist :3

Here's one last example usage; list your top 8 most used cmds:

fr3 -r.+ ~/.bash_history -strue | head -8
# equivalent to
sort ~/.bash_history | uniq -c | sort -nr | head -8

Note the -strue (--sort=true), this is needed because pipes aren't TTYs

etc

This program is single-threaded, as it's IO-bound.

I'm considering to support matching delimiters by regex (split)

See also

freq