Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 11:44

    axkr on master

    LogicNG 1.6.2 (compare)

  • Jan 17 17:35

    axkr on master

    improve boolean functions Negat… Improve ComplexSym#pow() Fix bug for Sum() and Product()… and 1 more (compare)

  • Jan 17 16:19

    axkr on master

    mathcell 1.8.6 (compare)

  • Jan 16 16:50

    axkr on master

    use mathcell 1.8.5 Javadoc / markdown (compare)

  • Jan 16 16:47
    axkr closed #160
  • Jan 12 14:21

    axkr on master

    Links to implementations (compare)

  • Jan 12 03:05

    axkr on master

    Improve interval arithmetic Min… Improve Expectation (compare)

  • Jan 11 05:48

    axkr on master

    Clean up/refactor Limit() (compare)

  • Jan 11 05:35

    axkr on master

    Improve evaluation of Piecewise… (compare)

  • Jan 10 16:39

    axkr on master

    Improve apfloat handling (compare)

  • Jan 09 21:44

    axkr on master

    added AbsoluteCorrelation for v… Improve calculations for multi … Clean up hypergeometric functio… and 1 more (compare)

  • Jan 08 18:28

    axkr on master

    markdown (compare)

  • Jan 07 00:03

    axkr on master

    Eliminate unary EqQ and NeQ cal… Eliminate EqQ and NeQ calls call hypergeometricSeries() ins… (compare)

  • Jan 06 17:03

    axkr on master

    call hypergeometricSeries() whe… (compare)

  • Jan 05 17:55

    axkr on master

    Prepare for apfloat nextDown() … (compare)

  • Jan 05 15:17

    axkr on master

    clean up (compare)

  • Jan 05 15:16

    axkr on master

    Implement ND() numerical differ… (compare)

  • Jan 05 12:22

    axkr on master

    Better accuracy for hypergeomet… (compare)

  • Jan 05 11:45

    axkr on master

    small performance improvement B… (compare)

  • Jan 05 06:47
    axkr commented on 1777f9a
Antoine Tran
@Tran-Antoine
Such a huge project ! I'm in admiration, how did you manage to do this (almost) on your own ? You've been working on this for quite a while I suppose
Ax Kr
@axkr
Yes almost on my own
But if you’re interested I‘m always looking for supporters
Antoine Tran
@Tran-Antoine
haha I'm pretty busy working on my own library, though it's a small one
but why not one day
Paul Masson
@paulmasson
Axel, I'll be at my computer for the next couple hours if you have questions about MathCell
Ax Kr
@axkr
I've seen you also have a math.js library
does it make sense to avoid the step of converting to JavaScript functions in Symja
Paul Masson
@paulmasson
Yes, for complex special functions
Ax Kr
@axkr
for exampe generate directly cos, sin, cot without Math. prefix
Paul Masson
@paulmasson
If you're only doing real-valued math then the JS functions will be a bit faster. If you want the option of complex numbers then use mine.
Ax Kr
@axkr
Overwise my javascript converter has to define code for example for a cot or coth function. So it will be easier to call the math.js functions directly
for both real and complex valued functions
Paul Masson
@paulmasson
Yup! That's why I wrote that library.
Ax Kr
@axkr
For the current issue I think http://localhost:8080/media/js/prototype/prototype.js line 403 calls the eval() function
Paul Masson
@paulmasson
Just to warn you those functions are also all global. I came at the problem after contributing to SageMath where it works that way, and I wanted the same convenience. Then again, there should only be one wat to define sin or coth, so global math makes sense to me.
Ax Kr
@axkr
Could this be the problem?
You have talked about an <iframe> solution in sagemath
Can I use the same solution here?
Paul Masson
@paulmasson
You're using eval() for injecting scripts, right? It doesn't make sense that it would alter JSON behavior, but as a dedicated JS person I never use eval() because it's considered unsafe, so maybe there's something I'm missing.
And yes, an <iframe> would take care of injection issues. The <script> still needs to be in a <div> of class=mathcell, then inside a <body> tag and that inside <html>. The source for the library can be in the body or the head. Then set the srcdoc of the iframe equal to the HTML string.
Ax Kr
@axkr
I must admit that the script is from mathics and modified for symja project
Paul Masson
@paulmasson
Would you like links to how this is done in SageMath?
Ax Kr
@axkr
As I said I'm not very good at javascript programming
Yes the sagemath solutioin would be nice to hava
Ax Kr
@axkr
@paulmasson Thanks in advance for your help. I have to leave now.
Paul Masson
@paulmasson
Iframe template is here. Returned full embedded document is here.
Or even simpler (without the bother of Python), here is how I do it in MathCell itself. Cheers!
The <iframe> approach will also take care of wrapping my global functions in a local context.
Ax Kr
@axkr
@paulmasson do you have any suggestions for JavaScript libraries which visualize graph theory structures?
Ax Kr
@axkr
@paulmasson I update the website https://symjaweb.appspot.com/ How can I calculate optimized "height" parameter for example if Manipulate shows a formula like Manipulate(Factor(x^n + 1), {n, 1, 5, 1})
Paul Masson
@paulmasson
For graph theory I'd suggest D3.js, although I don't use it myself. There is an implementation in SageMath, with JSON assembled in the python and rendered in HTML
Paul Masson
@paulmasson
Don't have a really good answer for automatic height, since it depends on what's in the cell. I generally add half an inch for each input line, plus four inches for 2D or 3D plots. If you know the output is text reduce that to one or two inches.
Ax Kr
@axkr
If you would like to help testing the latest Symja engine in an Android app, please register as a Beta tester in the Google play store for this app: https://play.google.com/apps/testing/com.duy.calculator.free
Евгений Буйволов
@ereborDeveloper

Hello, I'm using your library for my rest-application for education purposes. I'm using a lot symbolic algebra (math modeling of constructions).
Please tell me:

  • Is any possibility to speed up Simplify? (About 2 millions chars, get StackOverflow error)
  • What about ExpandAll? (work slow for 20-30 variables)
  • You have bug: evaluating of (0.17583680.41125407852711480.0*9.986163076795545E-11) gives -11, that's wrong, because simplifying of expression should give 0.0. Maybe I must create issue at github or smth else?

Thank you!

Ax Kr
@axkr
Can you give some examples for Simplify or ExpandAll. If possible open an issue on github:
I can add a label "Performance" afterwards
Why are there multiple dots in the first number ?
Maybe this is a problem in number parsing?
0.17583680.41125407852711480.0
Or is this because of gitter autoformatting?
Ax Kr
@axkr
I changed the parsing for floating-point numbers similar to Mathematica:
Multiple "dots" in numbers should now create multiple floating point numbers
>> 0.17583681.41125407852.0 // HoldForm // FullForm

HoldForm(Times(Times(0.17583681, 0.41125407852), 0.0))
Note: to use scientific notation 0.1E-11 you should set Config.EXPLICIT_TIMES_OPERATOR=true otherwise Ewill be interpreted as constant:
Евгений Буйволов
@ereborDeveloper
@axkr Sorry, I forgot about markdown
Please, try this
0.1758368*0.4112540785271148*0.0*9.986163076795545E-11
Евгений Буйволов
@ereborDeveloper
@axkr thanks! I created an issue on github axkr/symja_android_library#160
Ax Kr
@axkr
@ereborDeveloper I added your example to axkr/symja_android_library#18 to show the difference if you set Config.EXPLICIT_TIMES_OPERATOR=true See commit: axkr/symja_android_library@de8fc2a