#clojure – 2016-05-17

2016-05-17

01:30	ilevd	How I can setup timbre to print exceptions?
01:31	dysfun	it doesn't "print exceptions", but you can log an exception object you have caught
01:31	ilevd	And uncaught exceptions?
01:34	ilevd	Hm, it seems I need to set stdout, stderr to file to do this
01:35	dysfun	you can register a global uncaught exception handler
01:35	dysfun	and log it
01:38	ilevd	Thanks, I think it's what I need
01:39	dysfun	but you really should try and catch exceptions in your code
01:43	ilevd	I know, now I need quickly find an existing error
01:43	dysfun	ah :)
03:11	ilevd	Errors in pmap aren't caught by setDefaultUncaughtExceptionHandler -_-
03:12	dysfun	another reason not to use pmap
03:13	ilevd	Or i forgot to use doall..
04:00	ilevd	I do (future (try ... (catch Exception e (log e))), but message isn't infromative: java.util.concurrent.ExecutionException: java.lang.NullPointerException
04:00	ilevd	Any ideas?
04:01	dysfun	nullpointerexception means you've tried to call something as a function and it was null, or a function argument to a java function was null
04:02	ilevd	I expected some stacktrace, maybe line number
04:04	amalloy	blame whatever that log function is
04:12	ilevd	I can blame only myself
04:19	ilevd	.getStackTrace, .getMessage, .getCause return nil or java.lang.NullPointerException
04:53	TMA	ilevd: the java.util.concurrent.ExecutionException wraps java.lang.NullPointerException
04:53	TMA	ilevd: therefore .getCause returns the wrapped java.lang.NullPointerException
04:53	TMA	ilevd: try .getStackTrace and .getMessage on the java.lang.NullPointerException from .getCause
05:40	ilevd	@(future 10)
05:40	ilevd	,@(future 10)
05:40	clojurebot	#error {\n :cause "no threads please"\n :via\n [{:type java.lang.SecurityException\n :message "no threads please"\n :at [clojurebot.sandbox$enable_security_manager$fn__857 invoke "sandbox.clj" 94]}]\n :trace\n [[clojurebot.sandbox$enable_security_manager$fn__857 invoke "sandbox.clj" 94]\n [clojurebot.sandbox.proxy$java.lang.SecurityManager$Door$f500ea40 checkAccess nil -1]\n [java.lang.Threa...
05:50	paolo_	hey guys, i'm having problems using eval: i can't use clojure.math functions inside the expression to be eval-ed, can anyone help me?
05:57	TEttinger	clojure.math may need to be required again in the evaled code, but I'm half asleep what am I doing
06:09	Guest48	if I'm using map to execute a function a bunch of times on a sequence, and my call to map immediately finishes and the function I'm executing does not executed (as the log statement never runs) - what fundamental point is it likely I'm missing? I can log the sequence just before I run the call to map, and it has some elements in it
06:10	TMA	Guest48: map returns a lazy seq; if you do not force the realization, it will never be realized
06:10	Guest48	ah so I should (doall (map ... ?
06:11	Guest48	the behaviour seems different in the repl
06:11	TMA	Guest48: if you map for the side effects, then yes.
06:11	TMA	Guest48: because the result is realized for printing out
06:12	TMA	,(= nil (map println [1 2 3]))
06:12	clojurebot	false
06:12	Guest48	TMA: thanks loads. I had a feeling there was some fundamental thing I was not grasping
06:12	TMA	Guest48: see -- if the result is not printed, then it is not realized even in the REPL
06:13	TMA	,(= nil (doall (map println [1 2 3])))
06:13	clojurebot	1\n2\n3\nfalse
06:13	Guest48	oooh
06:13	Guest48	It is taking me a while to grasp these fundamentals despite an embarrassing amount of time tinkering with clojure
06:14	Guest48	thanks for your very useful explanation and example
06:14	TMA	Guest48: it'll just click together sooner or later, keep trying :)
06:14	Guest48	:) yup will get there in the end!
06:14	Guest48	yes everything is now exactly working as expected
06:14	Guest48	thanks!
06:15	TMA	you are welcome
10:35	sdegutis	How are you?
10:43	sdegutis	Okay.
10:45	fijro	When evaling a var, you don't have to type out the whole namespace. I type `+` instead of `clojure.core/+`. In my program I need to simulate this behaviour - and would like to know where in the Clojure source code I can study how that works? In what function is the 'full' name of a var completed?
10:46	justin_smith	fijro: eval uses the bindings of the current namespace to figure out how to resolve a symbol
10:46	justin_smith	fijro: you don't need to simulate the behaviour, you can call resolve
10:46	justin_smith	,(resolve '+)
10:46	clojurebot	#'clojure.core/+
10:47	hiredman	fijro: in the compiler if I recall there is a method named resolveSymbol
10:48	fijro	justin_smith: thank you, that's great to know. I need to reimplement that function, because ns in my app is a list that begins with a list that begins with a symbol 'ns. Don't know if I have misunderstood clojure, but feels like the ns part of it was not programmed in the clojure way of 'everything is data'...
10:48	hiredman	fijro: but you are likely to be more interested in resolve and ns-resolve
10:48	justin_smith	fijro: ns is a macro, it creates a namespace object
10:48	justin_smith	fijro: a namespace object owns some hash-maps from symbols to vars
10:49	fijro	hiredman: justin_smith: great to know. Thank you. The coffeeshop I'm coding in is closing and I've got to get going, but thanks!
11:53	sdegutis	Found a nice way of saying "(:quantity x) > 1"
11:53	sdegutis	,(map (comp pos? dec) (range 3))
11:53	sdegutis	Hahahhahaha
11:53	clojurebot	(false false true)
11:53	sdegutis	I know it's cheating but totally worth it
11:54	dysfun	not (comp pos? dec :quantity) ?
11:54	sdegutis	Yeah that.
11:54	sdegutis	But I got lazy and didn't want to make maps.
11:55	sdegutis	I wonder if there's a shortcut for (comp dec dec)
11:55	sdegutis	Man, I should make one. Fricken, what is it again? The thing that adds them together.
11:55	dysfun	yes, but it's longer
11:55	sdegutis	Repeatedly?
11:55	sdegutis	iterate?
11:55	sdegutis	Yeah apply iterate repeat dec n
11:55	sdegutis	or something
11:57	sdegutis	,((->> dec (repeat 2) (apply comp)) 3)
11:57	clojurebot	1
11:58	sdegutis	,(def dec-by #(->> dec (repeat %) (apply comp)))
11:58	clojurebot	#'sandbox/dec-by
11:58	sdegutis	,((dec-by 2) 7)
11:58	clojurebot	5
11:58	sdegutis	Surely it can be prettier.
12:01	sdegutis	Granted this is only a stepping stone for pos?
12:01	sdegutis	Something like greater-by?
12:03	sdegutis	Oh wait that reads unlogical.
12:03	sdegutis	And that ladies and gentlemen is why I do not write helper functions.
12:40	momerath	contains? is broken for transient sets; can anyone recommend a working equivalent? Allocation profiling definitely justifies trying to make this set transient.
12:42	hiredman	https://docs.oracle.com/javase/7/docs/api/java/util/concurrent/ConcurrentSkipListSet.html
12:43	hiredman	why people use transients for anything other that fast creation of a collection, I have no idea
12:43	momerath	ah- yeah, I guess I could use that; thanks hiredman!
13:14	sdegutis	momerath: transients should only be used briefly and turned into non-transients as soon as possible
13:15	sdegutis	momerath: the purpose of transients is to increase performance in creation of data sets in spots where performance matters
13:15	momerath	sure- that's true of the set in question, though the function is complex; it's very much an inner-loop
13:17	momerath	I used a plain java.utils.HashSet and I'm now allocating about 15% less in what might be called a cycle
13:18	dysfun	if you're in a tight loop, it's what mutables were made for
13:32	justin_smith	dysfun: I'm gonna get a decorative noose to hang on the wall and have a plaque by it that says "The Tight Loop".
13:32	justin_smith	momento mori kind of thing
13:36	dysfun	justin_smith: that's just asking to be involved in some sort of event where a coworker goes crazy and hangs your manager with your tight loop
13:36	dysfun	and you're gonna be prime suspect
13:36	justin_smith	that's why I wouldn't have a petard (which are notorious for hoisting their owners)
13:37	dysfun	in six months time, you'd better hope we start the "free justin_smith" campaign
13:37	dysfun	"free justin_smith, the rest of us have had to answer clojure questions"
13:54	sdegutis	hello
13:55	will_sm	hello
14:01	sdegutis	Fk freeing justin_smith. I refuse to answer Clojure questions.
14:01	sdegutis	Wait, how's that go?
14:21	sdegutis	What's a thing for like has-fraction?
14:21	sdegutis	Like, if you were to divine things, then it's intergral.
14:21	sdegutis	Like, (f 1/3) is false, but (f 2/4) is true
14:21	dysfun	wut?
14:22	justin_smith	dysfun: he's asking for a test for rationals vs. irrational numbers
14:22	dysfun	oh
14:22	dysfun	then he should have said so!
14:22	sdegutis	I just mean 'does this thing have anythig beside szeroes after the decimal?"
14:22	MJB47	i though ti was a test if a fraction could be simplified or not?
14:22	MJB47	ah ok
14:22	sdegutis	How do you say this mplore simply?
14:23	justin_smith	sdegutis: how does 2/4 satisfy but not 1/3 ?
14:23	MJB47	justin_smith said it well i think
14:23	sdegutis	"Is this thing really an integer?"
14:23	sdegutis	justin_smith: 2/4 is 2.0 1/3 is 0.33
14:23	MJB47	2/4 is not an integer
14:23	luma	2/4 is 0.5
14:23	justin_smith	sdegutis: wat
14:23	idev	where can I read an analysis of datomic vs relational model ?
14:23	sdegutis	Maybe (float x) = (int x) is what I want?
14:23	justin_smith	oh, he meant 4/2
14:23	sdegutis	justin_smith: yes, 4/2
14:23	sdegutis	Thanks
14:23	idev	I'm trying to figure out if EAV or Relational Model is the right appraoch for my needs
14:23	MJB47	oh
14:23	sdegutis	I do not understand math very well.
14:24	justin_smith	,(defn integral [n] (== (int n) n))
14:24	clojurebot	#'sandbox/integral
14:24	justin_smith	,(defn integral? [n] (== (int n) n))
14:24	clojurebot	#'sandbox/integral?
14:24	justin_smith	,(integral? (/ 4 2))
14:24	clojurebot	true
14:24	justin_smith	,(integral? (/ 2 4))
14:24	clojurebot	false
14:24	sdegutis	Yeah that's what I was thinking of a spossible solution
14:25	sdegutis	But is it the RIGHR one?
14:25	will_sm	you're testing a string?
14:25	justin_smith	sdegutis: definitely use == instead of =
14:25	sdegutis	I'm testing a denomenator and a numerator
14:25	sdegutis	I have both individually
14:25	sdegutis	They may be combined into a ratio if need be.
14:25	luma	there's integer?, which returns true for int, long, bigint and biginteger
14:26	sdegutis	luma: but I have two ints
14:26	sdegutis	Actually two longs.
14:27	sdegutis	e.g. 11500/9
14:27	luma	,(integer? 4/2)
14:27	clojurebot	true
14:27	luma	,(integer? (/ 4 2))
14:27	clojurebot	true
14:27	sdegutis	Why the crap does this warn on reflection (= (double 11500/9) (long 11500/9))
14:27	sdegutis	,(= (double 11500/9) (long 11500/9))
14:27	clojurebot	false
14:28	sdegutis	Ah, == not =
14:28	will_sm	is this the same as checking if (= 1 (gcd n d))
14:28	sdegutis	will_sm: hmm interesting idea
14:28	sdegutis	will_sm: what does gcd do? it's not in clojure.core
14:28	will_sm	greatest common divisor/factor, not in clojure
14:29	luma	,(integer (/ 11500 9))
14:29	clojurebot	#error {\n :cause "Unable to resolve symbol: integer in this context"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.RuntimeException: Unable to resolve symbol: integer in this context, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6688]}\n {:type java.lang.RuntimeException\n :message "Unable to resolve symbol: integer i...
14:29	luma	,(integer? (/ 11500 9))
14:29	clojurebot	false
14:29	sdegutis	luma: int
14:29	luma	sdegutis, why isn't integer? good enough?
14:29	sdegutis	Oh heck yeah, interger? works!
14:29	justin_smith	,(integer? (/ 4 2))
14:29	clojurebot	true
14:29	will_sm	, (defn gcd [a b] (if (zero? b) a (recur b (mod a b))))
14:29	sdegutis	,(map integer? [4/1 4/2 4/3 4/4])
14:29	clojurebot	#'sandbox/gcd
14:29	clojurebot	(true true false true)
14:30	sdegutis	Heck yeah!!!
14:30	justin_smith	,(integer? (/ 4 2.0))
14:30	clojurebot	false
14:30	sdegutis	Great job everyone.
14:30	justin_smith	sdegutis: as long as you never see floats, that should work
14:30	sdegutis	Never ever.
14:30	sdegutis	I have two longs.
14:31	sdegutis	What the crap?
14:31	sdegutis	(defn integer? [n] (or (instance? Integer n) (instance? Long n) (instance? clojure.lang.BigInt n) (instance? BigInteger n) (instance? Short n) (instance? Byte n)))
14:31	sdegutis	That's it's real definition, no foolin
14:32	sdegutis	Oh, it's because of /
14:32	sdegutis	,(type (/ 4 2))
14:32	clojurebot	java.lang.Long
14:32	sdegutis	fricken,
14:32	sdegutis	whatthecrap
14:32	justin_smith	,(type (+ (/ 1 2) (/ 1 2)))
14:32	clojurebot	clojure.lang.BigInt
14:32	sdegutis	,(for [i (range 5)] (class (/ 4 (inc i))))
14:32	clojurebot	(java.lang.Long java.lang.Long clojure.lang.Ratio java.lang.Long clojure.lang.Ratio)
14:33	engblom	,(type (/ 1 2))
14:33	clojurebot	clojure.lang.Ratio
14:34	sdegutis	,(for [i (range 5)] (class (/ (inc i) 4)))
14:34	clojurebot	(clojure.lang.Ratio clojure.lang.Ratio clojure.lang.Ratio java.lang.Long clojure.lang.Ratio)
14:35	sdegutis	,(clojure.lang.Numbers/divide 4 2)
14:35	clojurebot	2
14:35	sdegutis	That's why.
14:35	will_sm	sdegutis, so have you solved the problem?
14:35	sdegutis	https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Numbers.java#L155-L160
14:36	sdegutis	will_sm: technically yes, but I won't be satisfied until I understand why '/ gives me the answer I want so easily.
14:38	sdegutis	which uses https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Numbers.java#L1012-L1033
14:39	sdegutis	which in my case of two longs always uses https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Numbers.java#L417 for ops() ctor
14:43	sdegutis	so (/) with two Longs uses https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/Numbers.java#L489-L506
14:44	sdegutis	And oh look, it has a variable named "gcd" just lik will_sm was talkin bout (maybe-ish)
14:44	luma	yes, it's the same function
14:44	luma	greatest common divisor
14:44	will_sm	(gcd 2 4)
14:45	will_sm	,(defn gcd [a b] (if (zero? b) a (recur b (mod a b))))
14:45	will_sm	(gcd 2 4)
14:45	clojurebot	#'sandbox/gcd
14:45	will_sm	,(gcd 2 4)
14:45	clojurebot	2
14:45	will_sm	,(gcd 3 4)
14:45	clojurebot	1
14:46	will_sm	,(gcd 6 8)
14:46	clojurebot	2
14:46	sdegutis	Yall are good at math
14:47	will_sm	I almost became a math major, glad I chose to earn a good living instead
14:47	sdegutis	i have no fricken clue what this code does in my last link
14:48	sdegutis	but whatever, i trust that its doing the right thing based on what will_sm said plus his anecdotal qualifications
14:50	sdegutis	WHY WILL NOT THIS WORK
14:50	sdegutis	,(clojure.lang.Numbers$LongOps.)
14:50	clojurebot	#error {\n :cause "No matching ctor found for class clojure.lang.Numbers$LongOps"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.IllegalArgumentException: No matching ctor found for class clojure.lang.Numbers$LongOps, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyzeSeq "Compiler.java" 6875]}\n {:type java.lang.IllegalArgumentException\n :m...
14:50	bikeshedr	will_sm, hey, don't discourage me :p
14:52	sdegutis	I almost worked at factory third shift for the rest of my life, until my hobby combined with a series of strange events led to me becoming a competent software engineer.
14:52	kwladyka	wow.... i was rejected by Cognitect because i didn't know emac shortcuts.... recruitment process is totally random monster
14:52	sdegutis	Sh!t, 90 are integers but 29 are not.
14:53	hfaafb	how many does Cognitect employ
14:53	sdegutis	kwladyka: I was rejected by them because I am awkward and lack proper social skills.
14:53	sdegutis	kwladyka: but my employer is pretty happy with me so I guess cognitect missed out :P
14:53	will_sm	bikeshedr, if it makes you feel better, I will be working in hell/imperative programming for at least the next year
14:54	sdegutis	kwladyka: meaning, don't worry if they rejected you, they probably missed out on a great employee by not hiring you :P
14:54	kwladyka	sdegutis yeah... and i get in the same time feedback i have very strong skills etc..... but emacs...
14:54	sdegutis	will_sm: what lang?
14:54	dysfun	imperative programming isn't necessarily hell
14:54	will_sm	Haven't started yet but they use C++, Python, Java, & Go
14:54	bikeshedr	will_sm, a little :)
14:54	sdegutis	kwladyka: meh it's just one editor out of many, granted it's pretty darn great
14:54	sdegutis	will_sm: oh wow, you'll have fun with that one
14:54	dysfun	it just makes it a lot easier to make mess
14:55	sdegutis	will_sm: life is a waterslide where you die at the end
14:55	sdegutis	Hmm. How to figure out how many decimal places a thing has
14:55	kwladyka	i really cared about that job and put a lot of energy in that, but during one pair session i was talking with true emacs user
14:56	sdegutis	(= (map f [1/4 2/4 3/4 4/4]) [2 1 2 0])
14:56	sdegutis	What's f?
14:56	kwladyka	damn... do you know good Companies coding in Clojure which looking remote contractor? :)
14:56	dysfun	what is a "true" emacs user anyway?
14:56	kwladyka	dysfun only emacs matter, your editor sucks
14:57	kwladyka	dysfun but i think you have distance to it :) ( i hope ;) )
14:57	dysfun	no, it's just that i deal with so much shit software on a daily basis that emacs is comparatively unannoying except when it randomly breaks during upgrade
14:58	dysfun	all software sucks, all hardware sucks
14:58	dysfun	anyone telling you otherwise is a liar
14:58	kwladyka	i was trying emacs last 7 days very intensively. In my opinion emacs had great time, but time has changed :)
14:58	kwladyka	dysfun true
14:59	MJB47	my biggest issue with emacs was i spent too much time configuring emacs
14:59	MJB47	instead of actually working on whatever i needed to work on
14:59	MJB47	:(
14:59	dysfun	you do eventually get it into a steady state where it works 'enough'
15:00	will_sm	I limited myself to tweaking emacs for 10 min, every other week
15:01	ben_vulpes	> emacs
15:02	ben_vulpes	> works enough
15:02	ben_vulpes	how true
15:02	ben_vulpes	it helps to only scratch one itch at a time
15:02	kwladyka	dysfun what you will stack with things like colour syntax for .less files :)
15:02	kwladyka	*but
15:02	ben_vulpes	"dang it would be neat if focus jumped to the compile buffer on compilation failures"
15:02	kwladyka	i am trying atom editor, it looks very good, first impression is very strong
15:02	dysfun	you will hate it before long
15:03	ben_vulpes	that's the webkit one?
15:03	dysfun	text editors are not meant to run in the browser
15:03	dysfun	i think it's browser and/or node-webkit
15:03	MJB47	electron
15:04	MJB47	which is like node webkit
15:04	kwladyka	dysfun i don't have to use all features
15:05	dysfun	yeah, emacs people think that for the first week or so
15:05	kwladyka	but for using it as emacs it looks very good, but still i spend only few hours with it
15:05	kwladyka	*it - atom editor
15:06	dysfun	'using it as emacs'
15:06	dysfun	er, no. you have missed the point of emacs
15:06	engblom	kwladyka: Somehow I think you spend too much time thinking about the tools rather than coding :)
15:06	BamBalaam	Hi everyone! Does anyone know of a Clojure and/or Overtone that does music pitch detection?
15:07	kwladyka	engblom i was forced to it as you can see from my recrutiment process
15:07	BamBalaam	If not, any hint on any library for signal processing?
15:07	dysfun	hrm, i can't help but notice there's a big lack of signal processing libraries listed on clojure toolbox
15:08	will_sm	BamBalaam, IDK, but if straight pitch detection can't be done, you can use a Fourier transform to pick out the most dominant frequencies
15:09	dysfun	https://sourceforge.net/projects/dsplaboratory/
15:09	engblom	BamBalaam: If you get something like that done, please drop me on PM a link to the project. I know a person needing a such library
15:10	dysfun	http://www.source-code.biz/dsp/java/
15:11	BamBalaam	will_sm: it could be a good idea, for a quick and dirty first draft. I see how I could implement a FFT but not how to split open the music track in clojure
15:11	BamBalaam	dysfun: thanks for the links, will go through them
15:12	will_sm	BamBalaam, what is the source of the audio? Is it files?
15:12	BamBalaam	engblom: Whatever I achieve will never be extremely good
15:13	BamBalaam	will_sm: Yes, music files
15:14	BamBalaam	I think I'll start with really simple tracks just with basic notes/chords. If the solution works with more complex tracks, better, but not strictly necessary
15:15	will_sm	Found this lib for working with audio files, https://github.com/candera/dynne
15:17	BamBalaam	will_sm: thanks, will have a look at it ! :)
15:30	kwladyka	dysfun no, i didn't but as i said i was using atom only 2-3 hours and i don't know too much about it yet. But generally it looks like you can do with it everything like with emacs. Only one thing what you can't do is run on ssh.
15:35	sdegutis	How do you tell if a double only has 2 digits after the decimal?
15:38	justin_smith	will_sm: even fourier is tricky - the stuff actually implemented (fft) doesn't give you the actual frequency, it gives you a set of evenly spaced bins and you need to figure out the centroid from that using statistical methods
15:40	justin_smith	BamBalaam: overtone is just a DSL for the supercollider API, it can use all the UGens that supercollider defines and they are well documented, including multiple pitch detection methods
15:41	justin_smith	BamBalaam: this article compares three methods https://ccrma.stanford.edu/~kermit/website/scpitch.html
15:42	justin_smith	but frankly given examples and documented usage, it's a lot easier to just use sc directly and overtone just makes using it more complex
15:42	will_sm	justin_smith, gives a histogram of frequencies over some time interval. But yeah, given you still have lots of other problems such as start/end, chords, bin parameters and smoothing... how do ppl do it!
15:43	justin_smith	will_sm: but the frequencies have no relationship to the input
15:43	justin_smith	will_sm: they are just evenly spaced based on the algorithm parameters
15:43	justin_smith	finding the actual frequencies in the input requires statistical analysis
15:43	BamBalaam	justin_smith: I'm not extremely familiar with the supercollider API since I'm just starting with Overtone (or clojure for that matter!). It has pitch detection?
15:44	BamBalaam	justin_smith: thanks for the article, will read it, it's interesting!
15:44	justin_smith	BamBalaam: it has multiple pitch detection algorithms
15:44	BamBalaam	justin_smith: ok great, thanks for the information! :)
15:45	will_sm	justin_smith, wat. Fourier transform takes a signal and decomposes it to sine waves
15:45	justin_smith	will_sm: for example, if your bin is enough samples for 1/20th of a second, your detected frequencies will always be 0, 20, 40, 60, 80, 100, .... hz
15:46	justin_smith	to detect a 90hz input you either naively assign 80 or 100, or you use the spectrum spread behavior of the algorithm to derive the actual centroid
15:48	BamBalaam	oops, quit irssi by mistake
15:48	will_sm	justin_smith, use a higher resolution than, in practice I haven't had issues with aliasing, but then again, i don't have much practice
15:48	BamBalaam	but thanks everyone for the pointers in the good direction :)
15:49	justin_smith	will_sm: my point is that the fft only gives specific frequencies in the output, and will give you data per bin, and the frequencies of the bins are fixed by the algorithm, and have nothing to do with your input
15:50	justin_smith	will_sm: often the naiive usage is sufficient, but with a little interpolation / analysis you can do a lot better
15:50	justin_smith	but no, fft never actually gives you "the frequency of the input", the algorithm simply isn't capable of such a thing
15:53	will_sm	justin_smith, interpolation on results of fourier or resample audio? Have any source?
15:54	justin_smith	interpolations of fft to find centroids that are the actual input frequencies
15:56	justin_smith	will_sm: even the wikipedia page for fft will tell you that if your bin spacing is 20 hz, there will be no 90 hz bin - it's not even possible. The energy of a 90 hz input will be split evenly into the 80 hz bin and 100 hz bin. If you naiively scan for the bin with the most energy, that can cause a fundamental at 90 hz not to be detected
15:56	will_sm	or double the samplerate
15:57	justin_smith	will_sm: that's outside the algorithm, and you have the same problem again if you try a 95 hz input
15:57	will_sm	I mean, take your audio and change the sample rate
15:57	justin_smith	point is that fft never gives you the actual input frequency directly, but statistical methods on fft results can
15:58	justin_smith	will_sm: all that matters is the bin resolution in hz, but sure, changing input frequency can help you get more bin precision
15:59	justin_smith	but I think this is all a bit OT for this channel
16:00	will_sm	yeah, and we're getting nowhere lol. I might research it more later
16:01	BamBalaam	your discussion is super interesting though haha
16:01	BamBalaam	I'm just an extreme noob in the subject
16:02	BamBalaam	and my teacher simply totaly overevaluated the difficulty of the subject, given how we have close to no knowledge in signal processing
16:02	BamBalaam	(I know the basic math behind it, and that's...pretty much it)
16:03	BamBalaam	But I just might take the easy way out and try to use supercollider functions in Overtone (still gotta check how that's possible) and that might just be it
16:04	will_sm	when I installed supercollider, it came with the SuperCollider IDE, you can try that. I have yet to try it. Think I'll stick to standalone music hardware
16:04	justin_smith	BamBalaam: overtone definstrument is a macro that generates supercollider instrument definitions that are loadable by the supercollider synth server, that runs separately (it's a c++ program)
16:04	justin_smith	and then overtone talks to the scsynth server via UDP
16:06	justin_smith	BamBalaam: also, there are very friendly interactive tools for this stuff in matlab too
16:06	BamBalaam	justin_smith: I see. So in theory I could read the music file into a buffer, and use definstrument to interface with supercollider? I'll look into it
16:07	BamBalaam	justin_smith: I know. Unfortunately I really HAVE to use Overtone. It's... Silly.
16:07	BamBalaam	shrugs University
16:07	justin_smith	yeah, sounds about right - though supercollider is much more oriented to real-time, while mathematica makes it easier to do offline analysis
16:07	justin_smith	ahh, got it
16:07	dysfun	julia is gonna shit all over mathematica one day
16:08	will_sm	justin_smith, you know if there is any significant latency from overtone->supercollider or supercollider->jack
16:09	will_sm	I was really dissapointed with julia. Well, it was mostly the arrays. Weren't as good as numpy
16:09	sobel	i want to like Overtone more but i own a copy of Ableton Live
16:09	justin_smith	well, overtone doesn't even touch the signal data, and jack is going to get the best latency performance available usually
16:10	justin_smith	sobel: that's like saying "I want to like clojure but I already have excel" - these are tools of different levels of power designed for different kinds of tinkering
16:11	dysfun	justin_smith: i want to like clojure but i already have a banana
16:11	sobel	it's like saying "i want to like clojure but i already have excel AND know all its macros, have friends who regularly send me excel sheets, know how to extend it in C++ when needed, know how to interface it to all the services i need to integrate with my organization, ...
16:12	sobel	i want to like clojure but i have a hammock
16:12	dysfun	then go lie in your hammock and stop trying to write code
16:12	dysfun	it won't make you happy
16:12	sobel	. o O (one day)
16:12	justin_smith	sobel: if you needed to get an array of detected pitches in hash-maps as edn from ableton live, how would you do that?
16:13	will_sm	export to midi and convert to edn
16:13	sobel	justin_smith: probably export detected pitches through some interface in Max
16:13	dysfun	yeah, and how's the emacs support for ableton?
16:14	sobel	it's pointless
16:14	sobel	that's how it is ;)
16:14	justin_smith	I'm just saying, these are different kinds of tools - you can get a csv out of excel too of course, but they are really oriented for different approaches
16:14	justin_smith	max is just supercollider with point and click instead of text
16:14	dysfun	it's 2030 and the libreoffice import csv function still doesn't work by defualt
16:15	sobel	i used to do tons of CLI music/synthesis but i could not even come close to a good creative workflow that way
16:15	justin_smith	and ableton began life as a crazy max patch
16:15	Gh0stInTheShell	"Now it's the year three thousand and thirty, everybody wants to be an MC."
16:18	will_sm	Don't have windows anymore, I miss my DAWs
16:18	dysfun	can we invent our own expansions for acronyms we don't recognise?
16:22	sdegutis	Why is it that (/ 23000 18) must exist
16:22	sdegutis	,(/ 23000 18)
16:22	clojurebot	11500/9
16:22	sdegutis	,(double (/ 23000 18))
16:22	clojurebot	1277.777777777778
16:22	sdegutis	fricken..
16:22	dysfun	you'd love number theory sdegutis
16:23	sdegutis	dysfun: I'm guessing I'd actually find it incredibly inconvenient
16:24	will_sm	your hatred for these numbers is irrational
16:25	dysfun	inconvenient?
16:25	dysfun	i dunno, a lot of number theory is pretty important
16:26	dysfun	e.g. RSA derives its hardness from number theory
16:26	will_sm	Are you trying to count infinite decimal places now?
16:31	will_sm	sdegutis, so what project are you working on?
16:32	sdegutis	will_sm: a commercial website
17:08	sdegutis	,(double 95/15)
17:08	clojurebot	6.333333333333333
17:08	sdegutis	thanks OBAMA
17:25	kenrestivo	that's interesting. the compiler has no problem with [:timestamp] but pukes on [:@timestamp]
17:26	kenrestivo	tells me that there's an unmatched delimiter [
17:26	kenrestivo	not being able to handle :@timestamp puts a cramp in using elasticsearch from clojure
17:26	kenrestivo	,(select-keys {} [:timestamp])
17:26	clojurebot	{}
17:26	kenrestivo	,(select-keys {} [:@timestamp])
17:26	clojurebot	#<RuntimeException java.lang.RuntimeException: Invalid token: :>
17:27	kenrestivo	:(
17:27	justin_smith	kenrestivo: that's terrible, but you can use [(keyword "@timestamp")] but the better option is not keywordizing
17:27	kenrestivo	,(select-keys {} [(keyword "@timestamp")])
17:27	clojurebot	{}
17:27	kenrestivo	i like keywords
17:27	justin_smith	but you want to use "keywords" with invalid names
17:27	kenrestivo	also clj-http does that when you {:as :json} but i can manually work around that
17:28	kenrestivo	is :@ invalid?
17:28	justin_smith	,(select-keys {(keyword "@timestamp") ::yes} [(keyword "@timestamp")])
17:28	clojurebot	{:@timestamp :sandbox/yes}
17:28	TEttinger	I'd assume so, since @ is the deref macro
17:28	justin_smith	yeah, you are messing with the reader there
17:28	kenrestivo	oh, reader macro. grrr..
17:28	TEttinger	at-timestamp
17:28	kenrestivo	i can mangle the names on ingress to remove @'s
17:29	TEttinger	can you use strings?
17:29	justin_smith	so you can use the keyword function, or accept that you want to look things up that are not valid keywords and switch to string lookup
17:29	justin_smith	I guess there's mangling too, but that's the least simple option
17:29	kenrestivo	i hate strings so i'd rather sanitize the keys.
17:30	justin_smith	invalid keywords are worse than strings
17:30	justin_smith	,(keyword "\n\n\n")
17:30	clojurebot	:\n\n\n
17:30	kenrestivo	hahaha
17:30	kwladyka	what is the difference between paredit and parinfer?
17:30	justin_smith	paredit has commands for moving parens, parinfer moves parens when you indent
17:30	TEttinger	,(keyword "@$&^\ufeff()")
17:30	clojurebot	:@$&^()
17:31	kwladyka	justin_smith thx
18:08	sdegutis	How do you say "this double has no more than 2 decimal places, I'm sure of that"?
18:08	sdegutis	Or better yet, "this double ends with a repeating decimal place"?
18:14	sdegutis	Ah
18:14	sdegutis	(integer? (* x 100))
18:14	sdegutis	Got it.
18:15	amalloy	.(integer? (* 1.05 100))
18:16	amalloy	,(integer? (* 1.05 100))
18:16	clojurebot	false
18:17	sdegutis	,(integer? (* 100 (/ 52373 10)))
18:17	clojurebot	true
18:17	sdegutis	,(integer? (* 100 (/ 23000 18)))
18:17	clojurebot	false
18:26	amalloy	those aren't doubles
19:18	toastycj	I read or heard somewhere that clojure syntax is a derivative of another general language ? I couldn't find reference to it anymore after googling a bit.
19:18	toastycj	If somebody can let me know, it will be great.
19:18	toastycj	It's not lisp, from what I heard.
19:19	justin_smith	the closest relative is scheme
19:24	ben_vulpes	is let a recur target? i'm losing my mind over here
19:24	ben_vulpes	it shouldn't be right RIGHT?
19:24	justin_smith	no
19:25	amalloy	that question is curiously urgent when lacking punctuation
19:26	amalloy	like you're not just emphasizing the question, you're grabbing me by my collar and shaking me. "right RIGHT?"
19:27	ben_vulpes	yeah well when you feel the world moving under your feet and the only sane answer is "no you're wrong and stupid for making the mistake in the first place"...
19:31	TEttinger	toastycj, clojure can be described as a dialect of lisp, in the same way scheme is a dialect of lisp, and common lisp is a dialect of lisp (usually when people say Lisp without qualification they mean common lisp or the early lisps that predate common lisp)
19:32	TEttinger	clojure makes more departures from scheme and lisps in general than many other implementations, syntax-wise
22:53	sdegutis	amalloy: sometimes they are