Is there any way to wrap a function a promise quickly? So if I have a fn called
setIsOpen = (open) => this.isOpen = true is there a way to wrap it so it ends up being setIsOpen = (open) => () => this.isOpen = true?
I gotta say, @Bradcomp.... R.sequence and data.task brought up a very clean workflow over here. Thanks!
Is there a function that will remap keys of an object to new keys with the same values?
@tjpeden like this? https://github.com/ramda/ramda/wiki/Cookbook#rename-keys-of-an-object
@lucasschejtman exactly, thanks!
@JasonShin
guys, in Immutablejs, how do you compose its functions?
I want to use Map.map, Map.toArray, etc using lodash/fp flowRight / ramda.compose
so it would look like
guys, in Immutablejs, how do you compose its functions?
I want to use Map.map, Map.toArray, etc using lodash/fp flowRight / ramda.compose
so it would look like
flowRight(
toArray
map(myMapFunc)
)(myMap)
anyone?
=)..
Ramda’s
pipe function
@i-am-tom Why
pipe over compose?
OH, I misread - thought the question was specifically left-to-right composition
Why not? pipe reads left-> right, top->bottom
@JasonShin
flowRight(
x => x.toArray(),
x => x.map(myMapFunc)
)(myMap)
nice
thanks
Hi. Are there some fun exercises to play around with Ramda?
I like to play “refactor legacy codebase” :sweat_smile:
lol
I've asked the same question in Bacon.js, but there is almost no activity there, so I'll ask here too, please excuse me those that are in both channels
if I have Stream Stream Array , how do I convert it to Stream Array?
there is no
I'd like to flatten the contents of the current stream
the thing is I'm using
there is no
flatten method, and flatMap, flatMapConcat, and others need a second streamI'd like to flatten the contents of the current stream
the thing is I'm using
Bacon.fromPoll(x, () => Bacon.fromPromise(returnPromise()))
flatMap identity perhaps?
R.unnest should do the trick (if Bacon is compatible with old versions of Fantasy Land).
@galileopy
flatMap(x => x) try it
Is
R.unnest the highest level of R.unny and R.unner? ;)
@xgrommx thanks it worked
flowRight(
invoker(0, 'toArray'),
invoker(1, 'map')(myMapFunc)
)(myMap)
@justrag :joy:
how complex could become working with canvas in a functional way? I have a polling function that will update the list of locations that need to be shown periodically
I see 2 approaches, one is to create a marker for every location every time polling has finished
the other is to keep a list of references to existing markers, adding new marker only as needed, and modify their state whenever a new falue is available
a third option would be to create a stream of streams, each updating itself as needed, and the stream of streams will create a new stream every time there is a new marker to keep track of
that third option involves the creation of a new api endpoint that allows me to query each marker individually, which would result in a lot more api calls
it might not involve that, I could simply split the result and do the above
oh R.memoize allows me to just create one
that one is nice
TLDR: What is unit testing like with Futures and chaining? Do you have any examples of unit tests covering things like that? Do you even bother with it?
I’m not sure if this is an appropriate place to ask this question, so forgive me if that’s the case. Right now my team is experimenting with utilizing a more functional approach to software development. We all come from a TDD background with all kinds of mock and stubbing experience. We also don’t like the mocks so much :)
Our general architecture is moving toward composition at the outer edges of the system and then pushing branching logic into pure functions (except IO currently). We’re mostly on board with not testing those compose functions at the unit level, but rather just using integration or acceptance tests. But the thing that is causing some worry is general omitted code where we can’t determine correctness through outside behavior alone. Here’s an example:
getData(dataIds) {
return getDataFromCache(dataIds)
//.then(filterExpiredData) //Whoops, we should be calling this.
.then(getMissingDataFromApi)
}My question is, will taking the next step into things like futures and chaining them together help our testing situation? Will we be able to test our compositions without mocks? Or would we still not test compositions at the unit level?
Any insight would be greatly appreciated.
@bpetersen testing compositions is fairly straightforward if you think of the entire composition as a single function
with Futures, (I use mocha), running async tests using the
done argument is like testing a promise
I use a utility to put in the rejected callback on the future that will fail the test
import identity from 'ramda/src/identity'
import partial from 'ramda/src/partial'
import compose from 'ramda/src/compose'
import tap from 'ramda/src/tap'
const FAIL_MSG = chalk.bold.red('Test Executed Unreachable Code')
/**
* Create a function to use as contingency for unreachable code
* in an asynchronous test.
*
* @param doneCallback
* @return {*}
*/
export default function unreachableTestCode(doneCallback, verbose = 0) {
return compose(
partial(
doneCallback,
[new Error(FAIL_MSG)]
),
!verbose ? identity : tap(
console.log.bind(console,
`\n${ chalk.bold.underline('TEST STDOUT') + chalk.bold(' >> ') }\n`)
)
)
}
That could be overkill, a test would look like this:
import unreachable from '../helpers/unreachableTestCode'
it('testing of a future', function (done) {
someFutureReturingFn.fork(
unreachable(done/* optional boolean argument to logout error */),
(resolvedVal) => {
/* some tests go here */
done()
}
)
})
@bpetersen as for the mocking, you would only need to test the input/out of a composition. So you shouldn't have to mock anything
Thanks for the reply @mrosata. I really like the idea of treating the composition as a single function with inputs and outputs like any other function. Also, the Future test example was very helpful. Question on the point of mocking/stubbing. Does that mean that if one of our functions in the composition did some IO, like fetch from an api for example, we would just go ahead and let it do it? Or is there another way we can verify the correct behavior without going over the wire?
@bpetersen If your doing I/O in a function, I think you should probably test that function in isolation. I use
chai-fetch with API calls, so then I use a composition that contains an API call in a test, but I'm not really testing the composition as much as testing the side-effect, does it call the correct url, does it call it with the correct properties. I'm not a testing guru or anything, but it seems to work and with chai fetch there's no bloat in the code. I'll send you a snippet in a PM
Are there any plans to support valueOf like R.toString? For example, R.maxBy(valueOf) .. Or is there already a way to do this?
Anyone know if there’s a REPL for sublime text 3?
@joeyfigaro Not sure about a REPL within sublime, but you can always setup a Build system to run your code
Using babel and the build process isn’t quick enought to make that super attractive :/
@joeyfigaro use jsbin :smile: or ramda repl
@joeyfigaro was talking about the Sublime build system .. you can set it up to run your code in NodeJS for example
use vscode
@xgrommx vscode has a repl? :0
yes
vscode has terminal
built-in
@xgrommx I'm a big fan of the online REPL too .. ramda and ramda-fantasy are available .. its awesome
@bijoythomas probably u will be fan of https://github.com/princejwesley/Mancy
Mancy is super cool.
yes
I need support for elm,purescript or haskell :smile:
but I have HFM
@xgrommx thanks! It does look pretty kool
@bijoythomas yeah) nice tool
My earlier post might have gotten lost so asking again
Are there any plans to support valueOf like R.toString? For example, R.maxBy(valueOf) .. Or is there already a way to do this?
Are there any plans to support valueOf like R.toString? For example, R.maxBy(valueOf) .. Or is there already a way to do this?
You could use
R.invoker if you wanted to. maxBy(invoker(0, 'valueOf'))
@Bradcomp thanks. Never used it before but now see how it can be handy
Yeah. Another option is a simple lambda:
maxBy(x => x.valueOf())
:thumbsup:
@xgrommx Anything you’ve run across for using es6/7 in vscode’s integrated terminal?
@joeyfigaro if I need dev mode, I use babel-node (you can install it via
npm i -g babel-cli)
Oh, you’re running one module/file at a time through babel-node to evaluate?
for dev mode - yes
Ahhh, word. I’ve got my terminal already - looking for breakpoints or inline evaluation (or as close to as possible)
I like Mancy but for vscode users I've found quokkajs really nice as well
Thanks @lucasschejtman - checking that out now. :)
@lucasschejtman Quokka is fantastic. Thanks!