mjhjmtu

I'm on macOS Mojave.

I hate having to name images when adding them to my website (Jekyll, for anyone who cares). Instead, I think a reasonable strategy is to generate a hash based on the contents of the file and use the first N characters of the hash as the file name. This strategy also helps with versioning my assets when serving the website out of Amazon S3 and CloudFront.

Example

Given a file somescreenshot.png, I would run this TBD command and the result is the renamed file 8dbm2hz1.png.

Progress

I've made progress on this front but currently it's a multi-step approach.

1. Generate the hash, take the first eight characters and send to the clipboard.
   shasum -a 256 -b inputfile.png | head -c8 | pbcopy
   


2. Rename the file.
   mv inputfile.png [paste from #1]
   

I need to make this command an efficient piped, single step process.

Why not use:

mv -- inputfile.png "$(shasum -a 256 -b inputfile.png | head -c8)"

It will move your file to the output of $(shasum -a 256 -b inputfile.png | head -c8) i.e. first 8 character of the hash.

Use command substitution to save the hash in a shell variable:

file=inputfile.png
h=$(shasum -a 256 -b "$file" | head -c8)

Then you can use it for renaming, printing to the terminal, sending to pbcopy, or whatever.

mv -- "$file" "$h"
echo "$h" | pbcopy
echo "'$file' renamed to '$h'"

If you want to keep the file extension, use something like this:

newname=$h.$file##*.
mv -- "$file" "$newname"

($var##*. removes everything up to the last dot.)

If you want to rename all files, then use xargs. This is an incantation of GNU tools, check your man pages to see if the corresponding options exist.

printf "%s" * |
 xargs -0 -I sh -c 'echo mv "$1" "$(shasum -a 256 -b "$1"| head -c8)"' sh 

This uses the null byte to separate filenames, which makes working with filenames that contain whitespace a safe operation.

The echo command is to be removed to actually move the files.