Wolfram Mathematica

ilian

04:03

@kickert The cloud is running V12 now (the push was today, so I think any previous hiccups were unrelated)

Szabolcs

06:38

@ilian The (free) Development Platform show 11.3

Kuba

06:49

With[{reverse = -1 ;; 1 ;; -1 }
, Flatten[
  { {a, b, c, d}, {b2,c2,d2}, {d3}}[[All, reverse]]
, {{2},{1}}
][[ reverse, -1]]
]

> {a, b2, c2, d3}

it merges lists assuming they are aligned to the last element.

Better ideas how to do this?

b3m2a1

@Szabolcs on my end if you make a new notebook with a free plan you get 12.0

And you get to learn that FunctionCompile doesn't work in the cloud -_-

Kuba

This is more readable :P

Reverse @
Values @
Merge[Last] @
Map[KeyValueMap[#2 -> # &]] @
Map[PositionIndex] @
Reverse[{ {a, b, c, d}, {b2, c2, d2}, {d3}}, 2]

Side note, isn't it great we can write code this way? :)

Carl Lange

07:51

some bugs to iron out ;)

Select[Contexts[],StringContainsQ["sparql",IgnoreCase->True]]
> {}

I think I might cry :'(

Carl Lange

08:08

@Nasser @kirkus reference.wolframcloud.com/cloudplatform/guide/… give 269 new symbols

b3m2a1

@Kuba did you know you can set arbitrary variables in the FE?

FrontEndExecute@FrontEnd`SetValue[FEPrivate`Set[var, 2]]
var
FE`Evaluate[var]

var

2

Or perhaps to make it clearer what's going on here:

MathLink`CallFrontEnd@FrontEnd`SetValue[FEPrivate`Set[var, 2]]
var
MathLink`CallFrontEnd@FrontEnd`Value[var]

var

2

This also works for simple Mathematica expressions:

MathLink`CallFrontEndHeld@FrontEnd`SetValue[FEPrivate`SetDelayed[var, RandomReal[]]]
MathLink`CallFrontEnd@FrontEnd`Value[var]
MathLink`CallFrontEnd@FrontEnd`Value[var]

0.688695

0.270254

@LukasLang this might interest you too

b3m2a1

08:39

@CarlLange I checked that too. More unfortunate though is this:

In[7]:= Select[Lookup[PacletInformation/@PacletFind[], "Name"], StringContainsQ[#, "SPARQL",IgnoreCase->True]&]

Out[7]= {}

Carl Lange

So many kernel crashes trying to use the cloud at the minute...

b3m2a1

There might have been a paclet somewhere for it but it does not seem there is

Carl Lange

Yes, very sad. I suppose it's possible it just doesn't work on the cloud and so they wouldn't include it, but...

Kuba

@b3m2a1 I knew and used FEPrivate`Set: mathematica.stackexchange.com/a/128347/5478. I also remember problems with using it for more than simple symbol = value. But I don't remember details.

b3m2a1

@Kuba I'm sure it only supports a rather small interface of settings

Kuba

08:44

@b3m2a1 ah right, so one of problems was how to force rhs to be handled by FE as well. At least for cases which it should handle: mathematica.stackexchange.com/q/127895/5478

Carl Lange

FromCharacterCode[16^^1D378]

insta-12 crash

I forget about it most of the time, but CloudEvaluate is cool

the new textcases 👌

b3m2a1

09:07

@CarlLange are you reporting the crashes you hit?

Might still be time to put in patches for some of them (assuming they will also bedevil the actual system)

Carl Lange

@b3m2a1 No, I assume that they are basically still deploying this and bugs are expected. You think I should?

b3m2a1

It really can't hurt

At worst support says "yup we see it thanks"

Carl Lange

Good point, I'll report.

Kuba

@CarlLange what is $Version?

Carl Lange

@Kuba 12.0.0 for Linux

I'm using Wolfram|One in case they perhaps have different kernels for development platform etc

Kuba

09:14

@CarlLange I'd expect it to return a build date

together with os and v

Carl Lange

It does, I didn't include it: March 31

12.0.0 for Linux x86 (64-bit) (March 31, 2019)

Kuba

Thanks, I've let interested people know :)

Carl Lange

:)

accidental Louis le Brocquy

b3m2a1

09:29

@CarlLange saddest part about the cloud is it doesn't support 🔥yet

I wanted this:

🔥[args___]:=🔥 🔥 🔥

Carl Lange

@b3m2a1 omg tell me about it. When the Desktop version comes out I swear to god it better have emojis

b3m2a1

One I get 12 my code will no longer make any sense (assuming it supports Emoji)

It's not like anyone I work with actually reads the code I share with them

At least this way it'll be fun for all involved :)

Carl Lange

(╯°□°)╯︵ [┻━┻]

oh, that actually works in 11.3 on desktop:

b3m2a1

@CarlLange if you remove the brackets that's a Symbol

Kuba

@CarlLange it works in output cell but not in input

b3m2a1

09:34

Carl Lange

I was initially planning on returning the upside-down input but then thought I have better things to do with my day :D

Kuba

at least on Win

Carl Lange

one of these days I'm going to need a blog for this crap

Kuba

@CarlLange would've looked way better if input would support it

Carl Lange

🔥

Szabolcs

09:41

@rhermans To be honest I don't do review queues often. Today I looked at it, and it does seem like there are too many close votes. I clicked "Leave Open" many more times than "Close".

Carl Lange

Separate note: I notice GeoGraphics still doesn't support Callout, although it doesn't return pink anymore.

It's one of those things I really expect to work, even if it was just a passthrough to GeoMarker or something.

I like the new Information.

although ??classifier gives symbol information rather than classifier information, when Information@classifier gives classifier information. I'm not sure what I'm expecting really

b3m2a1

@CarlLange I'm working on updating my bloggy stuff so that it's smoother using my tabbed interface

If I get it to work well I'll let you know

Carl Lange

09:57

I was just thinking I might actually build a small page that's just an index to everything I CloudDeploy underneath /blog or something... that's probably enough, really

I just really want super low friction, because I'm allergic to faff

b3m2a1

That's what I originally did too, basically

Except with some extra HTML write-y stuff because the cloud was worse

Carl Lange

Mmm. I really need to get around to this.

Kuba

@b3m2a1 is "resource/path" an OS independent setting for paclet resource?

I mean, with respect to the pathname separator

b3m2a1

@Kuba do you have an example of what you're thinking of?

I'm not sure where you'd be using a path like that

Kuba

{"ExtensionToolbarsMap", "Settings/Mappings/ExtensionToolbarsMap.wl"}

@b3m2a1 :)

b3m2a1

10:08

Ohhh like that kind of thing

Not actually sure. What do you get for PacletResource["EasyIDE", "ExtensionToolbarsMap"] on Windows?

Kuba

@b3m2a1 PacletResource["EasyIDE", "ExtensionToolbarsMap"] // FileExistsQ returns True

b3m2a1

Looks like the paclet manager does the appropriate replacements then I guess?

Ah it does the cast with ToFileName which forces that kind of thing

Or maybe it's ExpandFileName doing it... hmm

Kuba

works, don't touch it

b3m2a1

At the very least it's done in the final entry in the DVs of PacletResource

Lukas Lang

@b3m2a1 Thanks, that looks really useful! Seems way easier and more flexible than storing everything in TaggingRules.

b3m2a1

10:21

@LukasLang and it won't cause DynamicUpdating all the time

That's really what has me interested. Controlling those updates is one of the hardest things to manage.

Lukas Lang

And also thanks for the link to the documentation notebook converter, I will definitely take a close look at that when I have the time to work on that

b3m2a1

10:36

Looks like the 12.0 ref pages are already up on the web for perusal

C. E.

I'm waiting for S.W.'s blog post. Then I will dig into the parts that seem most interesting.

The thing they've been pushing the most so far, from what I can tell, is the new geometry functions.

Kuba

10:59

@b3m2a1 Is there a top level way to update site associated with given paclet? It seems that PacletUpdate[pacet] isn't always up to date so I need PacletSiteUpdate /@ PacletSites[]. Which does more than needed.

b3m2a1

@Kuba only if you know the site it came from. That's not stored in the paclet data by default I believe.

Kuba

PacletSite[#, #2, "Local" -> #3] & @@@ Select[
  PacletManager`Services`Private`getPacletSiteData[],
  Not @* FreeQ["MyPaclet"]
  ]

b3m2a1

Ah wait here this should work if you have the paclet loaded already:

Kuba

@b3m2a1 seems to do the job but is kinda ugly ;)

b3m2a1

PacletSiteUpdate@PacletFindRemote["PacletName"][[1]]["Location"]

Kuba

11:03

Yes, that is it

Thanks

b3m2a1

If you don't already have the site loaded you'll need to PacletSiteAdd and PacletSiteUpdate it first, unfortunately

Kuba

That's fine

@b3m2a1 another one. Assuming the paclet is loaded, is there a way to get _Paclet from it's name? I do PacletFind[name][[1]]

b3m2a1

Nothing really better than that. CreatePaclet only works from the PacletInfo.m and you might have multiple versions anyway.

If you just want the info you can jump to PacletInformation["PacletName"]

Beware that PacletInformation is very different from List@@Paclet[...] though

Kuba

@b3m2a1 I have blood pressure jumps whenever I work with Paclet* and realize it is still undocumented.

b3m2a1

It's gonna bite someday, having done so much with it

Kuba

11:16

Really need to hold myself not to rant

b3m2a1

But then again if you're restricted only to documented functionality Mathematica's really slow and buggy

Like the documented core of Mathematica is only...okay

But once you add in the undocumented stuff it gets pretty cool

Kuba

@b3m2a1 it does not make me healthier to deliver code based on undocumented features but I can't avoid it nowadays.

b3m2a1

Yep if you are a developer it's unavoidable

b3m2a1

11:43

@Kuba at some point you asked about Mathematica OOP. Here's how I do it these days: mathematica.stackexchange.com/a/195065/38205

Kuba

@b3m2a1 you already answered :P but I decided to not dive into it as I found it too not idiomatic ;)

b3m2a1

Very true. It's about as far from idiomatic as I could get :)

But I have a strong preference for this type of thing to a whole pile of functions

That's mostly because I came to Mathematica after I'd been doing heavy-duty OOP for a while though

(and because OOP is such a nice paradigm)

Kuba

@b3m2a1 I believe it can be extremely useful if you have good command of OOP in general

I don't :p

Yet

b3m2a1

It comes. I didn't get it at first and then all of a sudden it just clicked.

Vitaliy Kaurov

14:18

Sequence-to-sequence regression with recurrent neural networks
https://community.wolfram.com/groups/-/m/t/1647625

Lukas Lang

16:15

@CarlLange @b3m2a1 The SPARQL stuff appears to be there: reference.wolfram.com/language/GraphStore/ref/SPARQLQuery.html - or did you mean something else?

b3m2a1

18:09

@LukasLang I think that's what we meant--we just didn't know the paclet name

Good find

Carl Lange

@LukasLang you have saved my day

Carl Lange

23:31

I choose to believe that I am referenced in the documentation for SPARQLSelect

s/referenced/mentioned

Carl Lange

23:45

(not going to read into the fact that I apparently have the lowest salary)

By the way, some of you might have seen my question in here recently about ordering coverage boxes for trail maps - here's what I did with that: toughsoles.ie/maps

Transcript for