has anybody tried to use a textcell as input for stringtemplate? My backticks get removed :(

scratch that, I just made a typo like an idiot

reversecolor is so crap

@CarlLange It's really hard to get this right in an existing application that always used positive coloring. I worked hard on the UI design of JabRef and although JavaFX uses CSS and many colors are "derived" from some basic colors, it takes a lot of testing and tweaking.

@halirutan I totally agree. But if SW used dark mode once in a while, it would be perfect.

And now think about Mathematica and how many hidden explicit RGBColor values are very likely hidden in the code.

Who knows what SW does but in essence you need to re-create every single icon that is used. Painful.

Definitely, I appreciate that it would be frustrating and likely near-impossible to make it work well. Still sucks though.

@CarlLange The solution is to just not provide a dark scheme if it isn't perfect. Think about how long MacOS needed to provide one.

@halirutan Yes, I suppose I agree!

Hey, so I did this thing where you can back an entitystore with wikidata

It's not very smart or fancy right now but I think it's pretty cool.

notebook in case anyone's interested wolframcloud.com/objects/e4782d00-46c6-499c-ab23-eccbd48f92f8

so you can create an entitystore for basically anything that wikidata knows about

Many cool low-hanging fruit to eat there as well - converting quantities, dates, countries and so on into regular entities for example...

@CarlLange I'll have to google those things -- I must officially be a dinosaur

@channel in more paclet server news you can now submit a paclet directly from a simple webform (unfortunately I can't implement OAuth so we're using Basic Auth): paclets.github.io/PacletServer/pages/submit.html

If anyone knows how to get OAuth on the client side I will switch over in a heartbeat

@C.E. how does this look: raw.githubusercontent.com/paclets/PacletServer/master/docs/…

For some reason GitHub doesn't want to include that JSON file on the actual website so it might need to serve from there.

@Nasser I'm pretty sure the problem is well-posed. It's the initial condition that makes the solution unique. Actually this is a problem mentioned in many PDE text books, for example, 《Introduction to Partial Differential Equations》 by Peter J. Olver:

Another example:

@Nasser Also, notice the solution given by your Maple is different from Mariusz's. The coefficient of t is different. Yours is correct.

@b3m2a1 That is good, I will put up a cron job that reads that URL periodically for the information.

@CarlLange You should write a Q/A on that. It's a great idea and undoubtedly very useful.

@C.E. Sounds good. If you want it in a different form I can do that too. I just decided to dump literally all of the information per paclet that I had on the server.

I figured better too much than too little.

@CarlLange I'm writing a wrapper layer on top of your stuff (I also have some experience working with EntityStore in 11.1 and I'll try to add convenience layers there if possible).

Is the limit argument to all the functions necessary?

I understand it cleans stuff up but I'm imagining we could generally do without it and if people want to get a humongous dataset well...they'll just abort the calculation.

One other question is whether it's possible to cut down on the number of queries that are made. It looks like there will be a bunch per dataset and I'd expect the API provides some way to vectorize these calls.

@xzczd thanks for the links, they are useful. Good point about the initial conditions. Hopefully Mathematica 12 will be able to solve this one also like Maple 2018.

@b3m2a1 The limit is there only to be nice to Wikidata's servers. mediawiki.org/wiki/Wikidata_Query_Service/…

@b3m2a1 There is definitely a way to do this - the propertyQuery should be a subquery of the entityQuery, I just didn't get to it

I kind of wanted to see if it was possible to do on-demand queries rather than one big query ahead of time, sort of like what they're doing with the entity stuff in 12

If I was going to do that I'd want to wait for 12 anyway though, since it'll have nice native sparql stuff instead of me writing the queries with StringTemplates

Oh, huh, I didn't realise this exists wikidata.org/wiki/Property:P4839

Thus, we could back all already existing entities in WL with wikidata also to enrich the data that exists

A little bit of awkward WKT parsing and you can get crazy things like the locations of every deck arch bridge...

Or, I dunno, computer company headquarters

I'm excited to write the community post for this!

@CarlLange Looking forward to reading it

Hi everyone, I want to draw a well, which actually the function pot[x_]:=Which[x<0,Infinity,0<=x<=a,-v,x>a,0] .. using Plot I am getting a picture which doesn't have the end points connected, I think using `Plot' I can't get the picture I want.

I want it like a well.. i.e; an extra vertical line connecting {a,-v} and {a,0}

In short how can I plot multiple lines in a single figure with Graphics[Line{...}]

Or Plot[pot[x], {x, -1, 3}, Exclusions -> False, PlotRange -> {-2, 10}]

@CarlLange I've been trying to get the WikiData API to work with POST, but can't wrangle it. Do you know if it does in fact work?

@b3m2a1 I do not know if it does in fact work, give me a few minutes

I once upon a time used the same underlying triple store, let me check

@b3m2a1 This works:

wikidataQuery[query_] :=
 ImportString[
  URLRead[
   HTTPRequest[<|
     Method -> "POST",
     "Scheme" -> "https",
     "Domain" -> "query.wikidata.org",
     "Path" -> "bigdata/namespace/wdq/sparql",
     "Body" -> URLQueryEncode[<|"query" -> StringTrim@query|>]
     |>, <|
     "Headers" -> <|"Accept" -> "text/csv"|>
     |>],
   "Body"],
  "CSV", "SkipLines" -> 1]

They want to body to have the full querystring in it

oh, it doesn't work work though

the response isn't CSV

@CarlLange I think you need the method and body in the second argument

That spits out like 80k things that I don't know how to parse

yeah, it spits out RDF instead of CSV

Or well I know how to parse them as JSON (I had it return me JSON)

But it wasn't clear if it was what I wanted or not

Yeah, hang on

hmm

ah

URLRead[
 HTTPRequest[<|
   "Scheme" -> "https",
   "Domain" -> "query.wikidata.org",
   "Path" -> "bigdata/namespace/wdq/sparql"
   |>, <|
   Method -> "POST",
   "Body" ->
    URLQueryEncode[<|"query" -> StringTrim@entityQuery["Q146"]|>],
   "Headers" -> <|"Accept" -> "text/csv"|>,
   "ContentType" -> "application/x-www-form-urlencoded"
   |>
  ],
 "Body"]

that's the one

ContentType

wikidataQuery[query_] :=
 ImportString[
  URLRead[
   HTTPRequest[<|
     "Scheme" -> "https",
     "Domain" -> "query.wikidata.org",
     "Path" -> "bigdata/namespace/wdq/sparql"
     |>, <|
     Method -> "POST",
     "Body" -> URLQueryEncode[<|"query" -> StringTrim@query|>],
     "Headers" -> <|"Accept" -> "text/csv"|>,
     "ContentType" -> "application/x-www-form-urlencoded"
     |>
    ],
   "Body"],
  "CSV", "SkipLines" -> 1]

Ah I see. Didn't realize we needed to supply the "ContentType". Presumably the post will be cleaner.

The upside is that larger queries will work, which is a good thing I suppose.

One other note, I took some of your stuff to make multiple property queries in one call, but I don't know how to aggregate the properties to a specific entity :|

Do you know how to do this?

Yeah, aggregating is quite awkward in SPARQL

Maybe, I'm happy to give it a try. Do you want to take this to a separate room?

Sure