a couple thoughts

[kedashoe]

I was thinking of a similar project but you beat me to it (by a mile :smile: ) Looks really great! A couple thoughts I had:

• support different input types. I would say maybe change -i, --inspect to -p, --pretty-print and use -i, --input-type and support at least csv/tsv out of the box. Also would be cool to support "white-space columns" input type (not sure what the accepted name is for that), eg the output of ls -l, top -n 1, ps aux etc. This may be a little more work than csv as I couldn't find any packages that already do this but I also am not sure exactly what terms to search for. Others?
• support different output types. I see you've added -r, --raw-output, maybe get rid of that and add -o, --output-type to be a companion to -i (and handle -o raw) and support all the same input types
• add an option to "save" your snippet. This is of course somewhat unnecessary (bash history, good old copy paste), but it hints to users/potential users how awesome ramda-cli can be as they can easily reuse the awesome cli stuff they've built. So I could do something like cat friends.json |\ ramda --save-as ~/bin/first-names.ramda 'first-word=(head . split " "); prop(\friends) >> map(first-word . prop(\fullName)) >> sortBy length' which would write the associated snippet to the file, and then in the future I can just run cat friends.json | first-names.ramda and I'm good to go :smile:

I was going to submit a PR but I've never used livescript and while it looks neat I don't have much free time at the moment.

[raine]

I like the idea of different input and output formats. Here's an example how formatting as CSV could work using a function and raw output together.

csv is to-csv. It takes a list of objects and the keys of the first object become the headers.

# outputs a string formatted as JSON
$ cat latest.json | R 'take 5' | R 'map pick <[ name category ]>' | R csv                                                                            
"name,category\n__,Function\r\nadd,Math\r\nadjust,List\r\nalways,Function\r\naperture,List\r\n"

# add -r for raw output. boom
$ cat latest.json | R 'take 5' | R 'map pick <[ name category ]>' | R csv -r                                                                           
name,category
__,Function
add,Math
adjust,List
always,Function
aperture,List

I like the idea that ramda-cli provides tools to use whatever output formatting function you want to use by allowing raw output, but on the other hand -o csv is an easier interface to approach so it's worth a try.

Pertaining to your third point, I was playing around with this idea:

~/.config/ramda-cli.{ls,js}:

var R = require('ramda')
exports.shout = R.compose(R.toUpper, R.add(R.__, '!'))

Functions exported in the configuration file would be available as such:

$ echo '"foo"' | R shout
"FOO!"

Also, something like this could be useful.

get-names.js:

var R = require('ramda');

module.exports = R.pluck('name');

cat latest.json | ramda -f get-names.js
[
  "__",
  "add",
  "adjust",
...

[raine]

--output-type is now implemented with possible values pretty, csv, tsv, raw

[kedashoe]

Very cool :) And feel free to close whenever you like, just thoughts rather than "issues"

[raine]

Sure. Thanks for the feedback, appreciate it a lot.