[sage-support] Combinatorics: Listing minimal subset covers(?)
Let L be a list of lists, like [[1,2], [2,3,4], [4,5]].
Is there a function to list minimal sets S such that for every (sub)list
of L there is at least one element in S? For example {1, 3, 5} would be
one "cover" that I am looking for. It is minimal, i.e. {1, 3}, {1, 5} and
{3, 5} would not cover all (sub)lists of L, even if it is not minimun ({2,
4} and {2, 5} are).
I suppose that this is NP-complete but can be done in small cases. Might
be that this could be formulated as a graph theory problem.
--
Jori Mäntysalo
Re: [sage-support] Parallel computation in Sage
On Mon, 16 Jul 2018, chandra chowdhury wrote: I want to calculate the determinant of a large matrix with large entries. So it is taking time. In my machine, I have 32 CPUs. Is it possible in Sage to use all CPUs parallelly to find the determinant? Are values real number or complex numbers, finite field elements...? Symbolics? -- Jori Mäntysalo
Re: [sage-support] Abstract Algebra tutorial ?
On Tue, 10 Jul 2018, Pat Browne wrote: Can anyone recommend a good tutorial for using sage Permutations and groups? I am a complete newbie and haven't been able to figure out how to do things with symmetry or permutation multiplications. I'm pretty sure that sage would be a great help in understanding these things. See http://abstract.ups.edu/sage-aata.html -- Jori Mäntysalo
Re: [sage-support] Re: gap: cannot extend the workspace any more! with tons of free memory.
On Wed, 30 May 2018, sbrandho...@web.de wrote: This is how my memory looks like before gap crashes: Every 5.0s: free -m Wed May 30 22:03:55 2018 total used free shared buff/cache available Mem: 15960 2539 11498 300 2021 12845 Swap: 8007 0 8007 I suggest that you try to increase swap to 2 times on RAM. Something like this might work (as root): head -c $((12*1024**3)) /dev/zero > /home/SWAPFILE chmod og= /home/SWAPFILE mkswap /home/SWAPFILE swapon /home/SWAPFILE -- Jori Mäntysalo
Re: [sage-support] Route inspection problem
On Sun, 1 Apr 2018, Dima Pasechnik wrote: this is also known as Chinese postman problem. There is a tutorial on doing it in networkx. http://brooksandrew.github.io/simpleblog/articles/intro-to-graph-optimization-solving-cpp/ Duh. Was more complicated than I thank. Anyways, thanks; maybe I'll implement that some day. -- Jori Mäntysalo
[sage-support] Route inspection problem
I was walking at https://kintulammi.fi/wp-content/uploads/Kintulammi_Opaskartta.jpg and hence wondered if Sagemath has a function to solve https://en.wikipedia.org/wiki/Route_inspection_problem . Didn't found any, is it hidden with some other name or module? -- Jori Mäntysalo
Re: [sage-support] Are they equivalent ?
On Sun, 1 Apr 2018, Henri Girard wrote: I am surprise because graph is different it should n't because the adjacency matrix is the same ? They are: sage: W = graphs.WheelGraph(4) sage: H = Graph(W.adjacency_matrix ()) sage: W == H True However, for almost every built-in graph or graph family Sage also knows some "nice" way to draw it. For other graphs it tries to found a "nice" drawing by partly randomized algorithm. You can also try H.show(layout='planar') -- Jori Mäntysalo
Re: [sage-support] avoiding startup time cost with multiple invocations
On Thu, 11 Jan 2018, Berkeley Churchill wrote: I'll look into using a pipe, that's a good suggestion and might be easier than the embedded interpreter. There are many ways to do this. Sage could read a pipe and write to files, another program could busy-wait to see for a file with given name to apper (of course Sage should first write to temporary file and then rename it). Or instead of busy-wait loop Sage could write to pipe when it's done. Also inotify is one possible solution. In any case, after you get rid of the startup time I suppose the next slowest part is converting data structures between Python and C. -- Jori Mäntysalo
Re: [sage-support] avoiding startup time cost with multiple invocations
On Wed, 10 Jan 2018, Berkeley Churchill wrote: (@Jori: unfortunately starting one process and doing all the computations at once won't work for us because we need to dynamically generate the n+1st computation based on the output of the nth computation. We could theoretically port all that logic to python/sage, but we don't really think that's worth it right now) Maybe you should then use pipe? First mkfifo thepipe ./sage -q < thepipe > thepipe and on the another window I tested with jm58660@j-op7010:~/sage$ echo 1+2 > thepipe jm58660@j-op7010:~/sage$ read result < thepipe jm58660@j-op7010:~/sage$ echo $result sage: 3 jm58660@j-op7010:~/sage$ tmp=$(echo $result | cut -f 2 -d ' ') jm58660@j-op7010:~/sage$ echo $tmp+3 > thepipe jm58660@j-op7010:~/sage$ read anotherresult < thepipe jm58660@j-op7010:~/sage$ echo $anotherresult sage: 6 jm58660@j-op7010:~/sage$ echo quit > thepipe Or maybe use Sage as controller part, i.e. call C++-program from Sage instead of calling Sage from C++-program? -- Jori Mäntysalo
Re: [sage-support] avoiding startup time cost with multiple invocations
On Wed, 10 Jan 2018, Berkeley Churchill wrote:
As part of a research project we're using sage as a subroutine for some matrix
computations over rings. We have hundreds or thousands of these computations,
but each computation
is fairly quick. Right now, for each computation we write python code into a .sage file,
and then start a new process with "sage somefile.sage > output". This works
fine, except
each time we incur the startup time, which is about 2.5 seconds (whereas running a
tiny python program is < 0.02s). Since we call sage hundreds of times, this
adds up pretty
quickly.
In Sage - actually in Python - you can handle exceptions:
L = [2, 0, -3]
for x in L:
try:
print(1/x)
except ZeroDivisionError:
print("Can't compute inverse of %x." % x)
so why you can't just make you C++ program to output, say, 1 of matrix
computation to a .sage file having some try-except -structure on them?
--
Jori Mäntysalo
Re: [sage-support] Random connected poset on n elements
On Thu, 30 Nov 2017, Christian Stump wrote: How big is your n? not very big, I aim for the biggest n for which I can loop through all permutations of n and compute some numbers. I expect this to be between 10 and 14. "Almost all" finite posets are connected, so uniform distribution of all posets would work too for bigger n. how would I get uniform distribution on all posets? I meant that it won't help you to discard unconnected ones. There is a code for generating posets, see attachment at https://trac.sagemath.org/ticket/14110 , but unfortunately it has not been integrated to Sage. I just tested and it takes about 2,2 seconds to generate 11-element posets (there are 46749427 of those) and 38 seconds for 12-element posets. I guess you could use that up to 14 elements. -- Jori Mäntysalo
Re: [sage-support] Random connected poset on n elements
On Thu, 30 Nov 2017, Christian Stump wrote: Is there a way to obtain a random connected poset on n unlabelled elements in sage? How big is your n? "Almost all" finite posets are connected, so uniform distribution of all posets would work too for bigger n. -- Jori Mäntysalo
Re: [sage-support] Generation of graphs where all maximal cliques have the same size
On Thu, 19 Oct 2017, Jori Mantysalo wrote: Might be worth trying to use from_graph6 directly. Tested. Don't bother, does not give really speedup that would mean something. -- Jori Mäntysalo
Re: [sage-support] Generation of graphs where all maximal cliques have the same size
On Thu, 19 Oct 2017, Christian Stump wrote: But this seems to be too slow, already because it takes too long for this to turn the graph6 string from nauty into a sage graph. Might be worth trying to use from_graph6 directly. Currenty nauty_geng() says G = graph.Graph(s[:-1], format='graph6'), and the function for that first does some checks and then does if format == 'graph6': if weighted is None: weighted = False self.allow_loops(loops if loops else False, check=False) self.allow_multiple_edges(multiedges if multiedges else False, check=False) from .graph_input import from_graph6 from_graph6(self, data) I think that doing import in a loop might take some time in Python. This won't help much, but maybe a little. -- Jori Mäntysalo
Re: [sage-support] Well covered graph
On Mon, 18 Sep 2017, Selvaraja S wrote: Definition: A graph $G$ is called well-covered if all minimal vertex covers of $G$ have the same number of elements. What is the coding to check well-covered graph? Maybe you can reformulate the question using independent sets and use PairwiseCompatibleSubsets, see the last example in page http://doc.sagemath.org/html/en/reference/combinat/sage/combinat/subsets_hereditary.html -- Jori Mäntysalo
[sage-support] Data structure for partition refinement
Sage has a static data structure SetPartition, and a dynamic structure DisjointSet that can *combine* blocks. Is there a "dual" structure of latter, something that would start from one big block and would have functions to *split* block? -- Jori Mäntysalo
Re: [sage-support] Re: Installing GAP packages & SageNB
Solved, kind of, but please read... On Thu, 26 Jan 2017, Dima Pasechnik wrote: But from the SageNB I got #I MakeReadOnlyGlobal: GradedAlgebraPresentationType no value bound true This looks OK But why the error message? GroupHomology(G,99); (I tried this with contents "gap" - set by the 4th switch above the cells) Me too. Forgotten ';' or something like this? No. I tried GroupHomology(G,99x); to test and got an error RuntimeError: Gap terminated unexpectedly while reading in a large line: Gap produced error output Error, Variable: '99x' must have a value I also tried F:=FreeGroup(2);; x:=F.1;; y:=F.2;; G:=F/[x^2,y^201,(x*y)^2];; G:=Image(IsomorphismPermGroup(G)); and got So, I tried foo:=GroupHomology(G,99); and a cell with only foo and got [ 2, 3, 67 ] So what the hell... I do not get output from the last line as I should? SageNB and some strang error? -- Jori Mäntysalo
[sage-support] Installing GAP packages & SageNB
I was asked to install the optional package HAP to the GAP inside
SageMath.
I run ./sage -i gap_packages and then followed instructions at
https://wiki.sagemath.org/InstallingGapPackages . Now it works from
command line, i.e. I can say
gap_console()
and then
LoadPackage("HAP");
F:=FreeGroup(2);; x:=F.1;; y:=F.2;;
G:=F/[x^2,y^201,(x*y)^2];; G:=Image(IsomorphismPermGroup(G));;
GroupHomology(G,99);
and got
[ 2, 3, 67 ]
But from the SageNB I got
#I MakeReadOnlyGlobal: GradedAlgebraPresentationType no value bound
true
from LoadPackage() and nothing from GroupHomology(G,99).
What to do?
--
Jori Mäntysalo
Re: [sage-support] Re: How to limit heavy computations
On Sun, 27 Nov 2016, Jeroen Demeyer wrote: In general it is not easy to limit resource usage in Linux. It's not easy, that's true. But that's independent of the fact that ulimit within SageNB simply doesn't work due to some SageNB bug. Yes, bugs are one thing. But the other is limiting resources for some users. And that is much harder when the resource use comes through SageNB or some similar system. I won't except to see good solutions to this in near future. -- Jori Mäntysalo
Re: [sage-support] Re: How to limit heavy computations
On Sun, 27 Nov 2016, Enrique Artal wrote: Thanks, As you say, it would be better something more direct, but your approach is a strong improvement for my needs. By the way, I changed in our experimental notebook 7.4 -> 7.3 and the limits work: they stop the process and the notebook is still running. OK, so there is a degeneration somewhere. Hopefully this is read by some of developers who can do something for it. In general it is not easy to limit resource usage in Linux. That's due to subprocesses, overcommit and the way programs are usually done. And of course having something like SageNB above process level does this even more complicated. -- Jori Mäntysalo
Re: [sage-support] Re: How to limit heavy computations
On Sat, 26 Nov 2016, Enrique Artal wrote: By the way, using server_pool, is there a way to know which user of the notebook is using a user of the server_pool? AFAIK not really. I can list files in /tmp and run fuser for them, i.e. fuser /tmp/* and see, for example /tmp/tmpwIR2_j: 7589c and in that file I can see a line like DATA = . . . /sage_notebook.sagenb/home/jm58660 and hence user jm58660 is behind the process #7589. This could be scripted of course, but I would like to have better NB made as admin viewpoint in mind. -- Jori Mäntysalo
Re: [sage-support] Randomness test
On Thu, 3 Nov 2016, chandra chowdhury wrote:
I have sequence like
{57, 44, 40, 57, 50, 53, 17, 24, 20, 3, 8, 51, 62, 56, 2, ..},
where each value is between 0 to 63. I want to check its randomness. Is
there any function in Sage for this?
How do you define "random"? Sometimes a perfect random number generator
will generate the sequence 1,2,3, Only definition of "random" I can
invent is "Can not be compressed." So can we do that?
I think that no. That would need some kind of heuristic, as the numbers
might have infinite number of generating functions. For example this is
not random at all: 92, 65, 35, 89, 79, 32, 38, 46: I got it from
str(pi.n(digits=30))[10:].
Suppose that your list has 1000 integers. Now we make a program "print 57,
44, 40". That program is, say, 1010 characters long. Is there any program
of at most 1009 characters that prints those numbers? Theoretically we can
just generate all strings up to 1009 characters, run them, and see what
happens. But if some program of 453 characters has been run for a day, has
produced first ten numbers of your list, and is still running, we do not
know if it will eventually produce other 990 numbers.
So, do you have some kind of guess about those numbers, something that
they might be produced from a
--
Jori Mäntysalo
Re: [sage-support] Re: Element vs UniqueRepresentation
On Sat, 8 Oct 2016, Nils Bruin wrote: What are all of the drawbacks? The ones I am aware of: Also random testing for a hypotheses is hard. I can generate infinitely many random graphs in a loop and try to found a counterexample, but the same does not apply to posets. -- Jori Mäntysalo
Re: [sage-support] Is it possible to add methods to an existing class at runtime ?
On Sat, 8 Oct 2016, Emmanuel Charpentier wrote: I wonder if it is possible to patch a new method in an existing class. TypeError: can't set attributes of built-in/extension type 'sage.symbolic.ex pression.Expression' AFAIK it is impossible due to limitations of Python. You can patch classes not defined in C, like type(Graph()).is_eulerian_and_hamiltonian = lambda self: self.is_eulerian() and self.is_hamiltonian() but that won't work for example to Sage integers. -- Jori Mäntysalo
Re: [sage-support] Re: plotting ln(x) graph on sagemath
On Sun, 18 Sep 2016, jack wrote:
Ubuntu16.04.
P=plot(log((1+x)/(1-x)), (x, -1,1))
show(P)
gives a lengthy error message which ends with
ImportError: cannot import name scimath
I installed sage at /home/jack/Tools
One clue might be the initial message I get on initiating sage in a
terminal. It reads:
sys:1: RuntimeWarning: not adding directory '' to sys.path since everybody
can write to it.
I have no real clue, but I would start with
strace -e file ./sage 2> log-f
then write
plot(sin)
or something similar and then quit Sage. After that
fgrep tmp_ log-f
shows
open(".../.sage/temp/j-op7010/12072/tmp_n3Sn3X.png", ...) = 11
etc. and
fgrep EACCESS log-f
finds nothing.
--
Jori Mäntysalo
Re: [sage-support] Paper citing sagemath
On Tue, 30 Aug 2016, Montgomery-Smith, Stephen wrote: I just submitted a paper: https://arxiv.org/abs/1608.06314 that cites sagemath. Would you guys like to add it to your webpage? http://www.sagemath.org/library-publications.html Is that web page maintained at all? If it is: Me too! See http://arxiv.org/abs/1403.5389 (publication information http://dl.acm.org/citation.cfm?id=2801897&CFID=662122917&CFTOKEN=12870255) -- Jori Mäntysalo
Re: [sage-support] How should I determine if two strongly regular graphs are isomorphic?
On Tue, 23 Aug 2016, Paul Leopardi wrote:
Do you have the package bliss or nauty installed?
sage: is_package_installed('bliss')
True
That explains it. canonical_label() uses Bliss by default if it is
installed. Setting algorithm='sage' will override this (and slow down
computation). So we should just expose more Bliss and Nauty features to
Sage.
--
Jori Mäntysalo
Re: [sage-support] Meta question: How do I best specify graphs in sage-support?
On Tue, 23 Aug 2016, Paul Leopardi wrote: Is there a particular reason why you prefer DiGraph(g).dig6_string() to g.graph6_string() even though it likely to give a string that is twice as long? No. I just didn't know that. (TODO: Add seealso-links to documentation.) -- Jori Mäntysalo
Re: [sage-support] Meta question: How do I best specify graphs in sage-support?
On Mon, 22 Aug 2016, Paul Leopardi wrote:
I am seeing a problem where "g.is_isomorphic(h)" is taking an inordinately
long time, where g and h are graphs, even though
"g.automorphism_group().order() == h.automorphism_group().order()", and
"g.canonical_label() == h.canonical_label()" both take a reasonable time.
Do you have Bliss installed? If so, then canonical_label() uses it, but
is_isomorphic() does not.
My main question for the moment is how do I specify graphs here in
sage-support so that others can reproduce the problem?
dig6_string() is only available do directed graphs, but a Graph made from
DiGraph is just a graph with directions forgotten. So you can say
DiGraph(g).dig6_string()
and on the other direction
Graph(DiGraph('DOOOW?'))
will give an undirected graph. We can then try to relabel your graph (like
g.relabel(lambda x: x+1) ) and check for isomorphism.
--
Jori Mäntysalo
Re: [sage-support] Re: Installing the optional python3 package seems to break Sage
On Wed, 10 Aug 2016, William Stein wrote: I was really surprised when I started Sage that people would often try to **install all optional packages**, then report anything that went wrong.I bet people still try to do this... Isn't it supposed to work... And on the servers I would like to have good coverage of features. This is not sage-specific: I can guess that some users here will use R and some might use Octave, so why not install both of them? (Well, nowadays I can just forward them to our Sage-server.) Of course I will not install all packages for, say, Ubuntu Linux, but Sage has not that many available. -- Jori Mäntysalo
Re: [sage-support] how to display graph legend with colored digraph
Quoting Ben :
That is great. Now, is there someway to permanently associate those two
parameters to the graph? i.e., G.show() has the colors and title?
AFAIK no.
You can set positions for vertices with set_pos() and you can attach
an arbitrary value to a vertex like
G.set_vertex(2, 'red')
but there seems to be no way to otherwise set default options. There is
G.graphics_array_defaults
that says among other things "'graph_border': True", but it seems to
have no effect. Hopefully someone can tell what for it is.
* * *
New version of Sage will make it possible to say
G.show(vertex_color='red', vertex_colors={'blue': [0, 1], 'green': [2, 3, 4]})
so that you can have a default color for vertices not listed in
vertex_colors. Same will be available for edge colors.
--
Jori Mäntysalo
--
You received this message because you are subscribed to the Google Groups
"sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-support+unsubscr...@googlegroups.com.
To post to this group, send email to sage-support@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-support.
For more options, visit https://groups.google.com/d/optout.
Re: [sage-support] how to display graph legend with colored digraph
On Tue, 26 Jul 2016, Ben wrote:
.graphplot() or .show() to do this. However, I'm having trouble displaying
some kind a legend to go along with the graph to explain which colors are
which properties.
I'd like to display with the graph something like
color1 = property A
color2 = property B
Good question. You can say
G = DiGraph({1:[2,3], 2:[3]})
G.show(vertex_colors={'red': G.sinks(), 'blue': G.sources()}, title="Sinks in
red.\nSources in blue.")
but I don't know how to add text to a plot. For example
G = DiGraph({1:[2,3], 2:[3]})
gr = G.plot(vertex_colors={'red': G.sinks(), 'blue': G.sources()})
graphics_array([gr, text("Some colors\nto show special vertices.", (0,0))])
does something that you don't want. I guess that LaTeX (with, say, Tikz)
is needed for production quality pictures. Sage's graphics works for
demonstration in classroom.
--
Jori Mäntysalo
[sage-support] Matrix over SR, reduced row echelon form
What is the logic behind this:
var('a b c d')
M = Matrix(SR, [[a, b], [c, d]])
M.rref()
-->
[1 0]
[0 1]
?
--
Jori Mäntysalo
[sage-support] Default output form for AlgebraicNumber
Matrix(QQ, [[1,2],[3,4]]).eigenvalues() --> [-0.3722813232690144?, 5.372281323269015?] Not good for teaching (and I just got a complain for that). Of course I can do map(lambda x: x.radical_expression(), M.eigenvalues()) or something similar. Or begin with Matrix(SR, ...). But is there a default setting to change output type of AlgebraicNumber? -- Jori Mäntysalo
Re: [sage-support] Re: Weber point
On Thu, 7 Apr 2016, kcrisman wrote:
It's very easy to install a new R package from within Sage, even in
sagenb, though. r.install_packages('orloca')
True. Took about two seconds to download, compile and install. But I have
self-compiled Sage - does it also work in ready-made binary Sage?
Installation ends with message "Please restart Sage in order to use
'orloca'.". However, even restarting worksheet was not needed.
--
Jori Mäntysalo
Re: [sage-support] Re: Weber point
On Thu, 7 Apr 2016, Dominique Laurain wrote:
Trying just now in sagemath cloud worksheet
r.library('orloca')
and it seems no package orlorca (or depth)
This is actually more general problem that comes from the "building a car"
-principle. SageMath contains at least R and GAP, and they both have their
own subpackage system. I think I tried semigroup-package of GAP some time
ago, and there was some complication when installing it inside Sage.
* * *
OFF-TOPIC:
R is extremely easy to install. It has ready-made binaries for Windows and
Mac. For Linux at least for both Ubuntu and Fedora have it in standard
repository.
So I suggest you to install local R.
--
Jori Mäntysalo
[sage-support] Weber point
[ Not important, I am just curious. ] Does SageMath have a function that takes a set of points in a plane as an argument and return an approximation of the point that has shortest distance to given points in average? The point is called the Weber point if I have understood this right. r.eval(...something...) maybe? -- Jori Mäntysalo
Re: [sage-support] graph plotting: heights option
On Fri, 11 Mar 2016, Pierre wrote:
This is from the documentation of the plot method of the Digraph class:
sage: T = list(graphs.trees(7))
sage: t = T[3]
sage: t.plot(heights={0:[0], 1:[4,5,1], 2:[2], 3:[3,6]})
The result (on SMC, and I think it used to be the same on my local sage) is
not what I expected at all: perhaps I misunderstand the 'heights' option, bu
t I thought that the vertices 3 and 6, for example, would be on the same hor
izontal line, as would be 4, 5, 1. Not the case for me at all !
I can confirm this. I know that this may depend on dot2text, installing it
may make difference.
For example after t = Posets.PentagonPoset().hasse_diagram() the
height-option makes nothing. Except that it does, if you use option
layout='acyclic'.
I am afraid that you must directly use .set_pos() on the graph.
--
Jori Mäntysalo
Re: [sage-support] Re: %latex and no output
On Thu, 10 Mar 2016, William Stein wrote: The real bug is that we have no guide to installing Sage server. https://wiki.sagemath.org/SageServer I stand corrected. I guess I will go throught that after Ubuntu 16.04 LTS is out. -- Jori Mäntysalo
Re: [sage-support] Re: %latex and no output
On Thu, 10 Mar 2016, kcrisman wrote: I found the reason. File permissions. In your specific case, or in general? If in general, we can open an issue to at least keep track of it and suggest workarounds. Sage only has notebook(..., server_pool=['someone@somewhere'], ...). It is up to administrator to make the account running Sage GUI to be able to log in as someone@somewhere. The real bug is that we have no guide to installing Sage server. But... If I say plot(sin) it works: the process running GUI as uid 'sagegui' makes a file to /tmp, the process running computation as uid 'sagecalc' reads it and makes a picture file that 'sagegui' can read. But for %latex it does not work: file written by 'sagecalc' gets permissions that do not allow 'sagegui' to read it. Why so? I think you'd have to go through the code that parses the various percent directives in sagenb and see how they get used; sometimes I think they do start up new processes, whereas plot() is a "Sage native" process. OK, so there is some difference coming from system() vs. some other functin. Duh. I will look this at monday, when we have a service break. -- Jori Mäntysalo
Re: [sage-support] Re: %latex and no output
On Mon, 29 Feb 2016, kcrisman wrote: Got it. As I say, I can't see this in any current Sage I have, except a server with no LaTeX, so it would be hard to debug. I found the reason. File permissions. But... If I say plot(sin) it works: the process running GUI as uid 'sagegui' makes a file to /tmp, the process running computation as uid 'sagecalc' reads it and makes a picture file that 'sagegui' can read. But for %latex it does not work: file written by 'sagecalc' gets permissions that do not allow 'sagegui' to read it. Why so? -- Jori Mäntysalo
Re: [sage-support] Re: %latex and no output
On Thu, 25 Feb 2016, Henri Girard wrote: In sagenb you don't need to write %latex ? Just click on the box latex to mark it ? What box? But now I tested "Typeset" box. It works, so LaTeX installation is OK. And for example %timeit works, so %-string works. Strange. -- Jori Mäntysalo
Re: [sage-support] Re: %latex and no output
On Thu, 25 Feb 2016, kcrisman wrote: %latex x does not provide any output. Where to start debugging? Is this in sagenb? Yes. I can confirm this works for me in 7.1.beta3, though I do have LaTeX installed. In the server I have access to with 6.9 I get this. 6.5 on my computer works fine, so I *suspect* it is related to the LaTeX on the server somehow. But maybe there was something that was wrong and fixed in short order in Sage itself? Can it relate to server_pool -option? With a slight misconfiguration one can get SageNB where sin(0) works but plot(sin) does not. -- Jori Mäntysalo
[sage-support] %latex and no output
A cell containing only
%latex
x
does not provide any output. Where to start debugging?
This works on my local computer with newest beta. On server where this
does not work we use latest stable release. For example
%latex
x^2
gives
An error occurred.
This is pdfTeX, Version 3.1415926-2.5-1.40.14 (TeX Live 2013/Debian)
(format=pdflatex 2016.1.4) 25 FEB 2016 15:40
entering extended mode
restricted \write18 enabled.
. . .
(/usr/share/texlive/texmf-dist/tex/latex/latexconfig/epstopdf-sys.cfg
File: epstopdf-sys.cfg 2010/07/13 v1.3 Configuration of (r)epstopdf for
TeX Liv
e
))
! Missing $ inserted.
$
l.37 x^
2
I've inserted a begin-math/end-math symbol since I think
you left one out. Proceed, with fingers crossed.
LaTeX Font Info:Try loading font information for U+msa on input line
37.
(/usr/share/texlive/texmf-dist/tex/latex/amsfonts/umsa.fd
File: umsa.fd 2013/01/14 v3.01 AMS symbols A
. . .
I've inserted a begin-math/end-math symbol since I think
you left one out. Proceed, with fingers crossed.
[1
{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}] (./sage65.aux) )
Here is how much of TeX's memory you used:
2798 strings out of 495028
--
Jori Mäntysalo
Re: [sage-support] Writing 'n' variables in Sagemath linux ubuntu
If I understood this, isn't var(['a'+str(i) for i in range(1,4)]) what is wanted? It will make a1, a2 and a3 to variables. -- Jori Mäntysalo
Re: [sage-support] redirecting input and output to Sage
Quoting Luis Finotti stdbuf --output=L ./sage tmp.sage > tmp.out I was unfamiliar with stdbuf... Then maybe you don't know time, timeout, ulimit etc. either. And in some (rare) cases even "at" would be nice for some users. Actually, maybe I should write a short text for Linux commands maybe useful for advanced Sage users. -- Jori Mäntysalo -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To post to this group, send email to sage-support@googlegroups.com. Visit this group at https://groups.google.com/group/sage-support. For more options, visit https://groups.google.com/d/optout.
Re: [sage-support] redirecting input and output to Sage
Quoting Luis Finotti : Could someone explain how redirection works with Sage (or point me in the right direction)? I suggest to not do anything sage- or python-specific to general problem like this. Instead rely on usual tools, stdbuf in this example: stdbuf --output=L ./sage tmp.sage > tmp.out Also check if the command "tee" fits your purpose. -- Jori Mäntysalo -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To post to this group, send email to sage-support@googlegroups.com. Visit this group at https://groups.google.com/group/sage-support. For more options, visit https://groups.google.com/d/optout.
Re: [sage-support] Packages needed for self-compiled Sage on Ubuntu
On Wed, 6 Jan 2016, Jeroen Demeyer wrote: Is there a list somewhere on Ubuntu packages that are needed to install fully working Sage from source? The hard part here is defining "fully working". "Everything that is supposed to work without optional packages (installed with sage -i) really works."? -- Jori Mäntysalo
[sage-support] Packages needed for self-compiled Sage on Ubuntu
Is there a list somewhere on Ubuntu packages that are needed to install fully working Sage from source? For example now I have a machine where view() does not work as LaTeX has no tikz. Instead of every admin to do the same job it would be nice to have a copy-and-paste help. -- Jori Mäntysalo
Re: [sage-support] For loop difference between sage in the cloud and local instalation
On Sat, 26 Dec 2015, William Stein wrote: quite sure that I was able to do something like for x in SomeObjects(n): if x.foo() and not x.bar(): x.show() and see pictures coming as the loop is running. Yep. But there are no print statements above. OK. But what has changed, as now the pictures does not come one-by-one? -- Jori Mäntysalo
Re: [sage-support] For loop difference between sage in the cloud and local instalation
On Sat, 26 Dec 2015, William Stein wrote: print "Original:" P.show() print "Reversed:" P.dual().show() It has never worked that way. ?? Then my main memory is corrupted. Got to reboot for this evening. What about outputting graphics as the computation goes on? I am quite sure that I was able to do something like for x in SomeObjects(n): if x.foo() and not x.bar(): x.show() and see pictures coming as the loop is running. -- Jori Mäntysalo
Re: [sage-support] For loop difference between sage in the cloud and local instalation
On Wed, 23 Dec 2015, Kazimierz Kurz wrote:
On my own instance of sagemath ( ver. 6.8) there is complete different!
The same code produces first list of 36 elements of G[k][0],G[k][1] and them
36 plots.
Is this normal?
Yes, it is. And I don't like that either.
In some previous version of Sage I was able to say
P = Poset({1: [2]})
print "Original:"
P.show()
print "Reversed:"
P.dual().show()
in the notebook and got expected result. It just does not work that way
anymore.
--
Jori Mäntysalo
[sage-support] Using graphic object as a sub-figure
Is it possible to use a plot of some object as a subfigure in given position? Suppose for example that g is a graph. What if I want to plot g, an arrow, and g with some edge deleted? There is graphic_array, but it is not quite flexible. I would like to have something like G = Graphics() G.add(g.plot(), position=(2,3)) G.add_primitive(Arrow(...)) . . . -- Jori Mäntysalo
Re: [sage-support] Relabel vertices
On Mon, 7 Dec 2015, Selva Raja S wrote:
Suppose I want to draw the graph. Coding is
Can i change the relabel vertices?
g = Graph({1:[2,3],2:[3]})
g.relabel(lambda x: chr(ord('a')+x), inplace=False).show()
--
Jori Mäntysalo
Re: [sage-support] How to hide output when opening a sagemath worksheet
On Tue, 17 Nov 2015, Gabriel Cardona wrote: I am sorry but cannot find that setting... where is the Action menu? Sorry, on SageNB. I.e. not in cloud version(?) of Sage GUI, I guess. -- Jori Mäntysalo
Re: [sage-support] Re: Server and certificate chain
On Tue, 17 Nov 2015, kcrisman wrote: Jori, are you using Jonathan Gutow's branch for proxying? No. I rearranged things. We have now a) redirection http://sage.our.unit -> https://sage.our.unit and b) documentation online at http://sage-doc.our.unit. This seems to now work. Apache is used as a proxy to locally running sagenb. -- Jori Mäntysalo
Re: [sage-support] How to hide output when opening a sagemath worksheet
On Tue, 17 Nov 2015, Gabriel Cardona wrote: I evaluate it and hide the output, but when I open the worksheet again the '2' is not hidden (I want my students to open the worksheet and evaluate the expression themselves). Can you just use Action / Delete All Output? -- Jori Mäntysalo
Re: [sage-support] Re: Server and certificate chain
On Fri, 13 Nov 2015, Jori Mantysalo wrote:
RewriteRule ^(.*)$ http://localhost:1234/$1 [P]
Now I was able to delete this and everything almost work.
I have static documentation available without login at port 80. Server is
at 443 so that Sage listens port 8000 and Apache configuration have
ProxyPass http://localhost:8000/
ProxyPassReverse http://localhost:8000/
However, login sent user to http root, not to https root, and so does new
worksheet. So I have on non-sssl configuration
RewriteEngine On
RewriteCond %{HTTPS} off
RewriteCond %{REQUEST_URI} !^/documentation
RewriteCond %{REQUEST_URI} !^/$
RewriteCond %{REQUEST_URI} !^/index.html$
RewriteRule (.*) https://%{HTTP_HOST}%{REQUEST_URI}
But basically we need SITE_ROOT as a configuration for the notebook.
* * *
Btw, see http://httpd.apache.org/docs/2.4/mod/mod_proxy.html#proxypass
Context:server config, virtual host, directory
vs.
Note: This directive cannot be used within a context.
--
Jori Mäntysalo
Re: [sage-support] Re: Server and certificate chain
On Thu, 12 Nov 2015, Gutow, Jonathan H wrote: I strongly recommend using a frontend like ngix or Apache. The efficiency of the Sagenb server when doing encryption is low. Let an industrial strength web server handle that and the encryption. Speed is not a bottleneck for me, but anyways, I'll try Apache. Now, do anyone have ready-to-go configuration for lazy IT admins? I tried basically just RewriteRule ^(.*)$ http://localhost:1234/$1 [P] (where 1234 is the port that Sage listens), and it has at least the error with logout link. -- Jori Mäntysalo
Re: [sage-support] Re: Server and certificate chain
On Thu, 12 Nov 2015, Karl-Dieter Crisman wrote: Without proxying on a local machine sagenb just launches. The only difference is that it is now at http://localhost:8080/sage/. Would the same be true for a non-proxied launch to the internet? (Assuming one did such a thing.) Yes. I have done that. (On port 443 with authbind.) -- Jori Mäntysalo
Re: [sage-support] Re: Server and certificate chain
On Thu, 12 Nov 2015, Michael Orlitzky wrote: Can Sagenb send certificate chain instead of just certificate of the server itself? If yes, how? You can probably just concatenate the certificate files. Something like, $ cat www.example.com.crt intermediate.pem > combined.crt Tested, but it seems that Sage sends only first certificate. -- Jori Mäntysalo
[sage-support] Re: Server and certificate chain
On Wed, 11 Nov 2015, Jori Mantysalo wrote: If not, what do you use as a frontend? Apache? Apache seems to work, I was able to log in and saw my worksheet and plot() worked. However, logout sent me to localhost. So not all links work -- maybe there are more that do not work. Suggestions welcome. -- Jori Mäntysalo
[sage-support] Server and certificate chain
Can Sagenb send certificate chain instead of just certificate of the server itself? If yes, how? If not, what do you use as a frontend? Apache? -- Jori Mäntysalo
Re: [sage-support] Same Matrix Construction gives two different results
On Sun, 8 Nov 2015, Nils Bruin wrote: Ah, that's a very important detail: that indicates that something that the preprocessor does makes a difference. Todo to someone: add something like http://doc.sagemath.org/html/en/faq/faq-usage.html#what-exactly-does-sage-do-when-i-type-0-6-2 to FAQ so that in future we could just give a link as an answer. -- Jori Mäntysalo
Re: [sage-support] Same Matrix Construction gives two different results
On Sun, 8 Nov 2015, Sihuang Hu wrote: I just used the following commands to construct an matrix: q = 2 r = 2 s = 2 n = s*(r+1) d = 3 P = matrix(ZZ, r+2, r+2, lambda u, k: Krawtchouk(r+1, q, k, u)) But sage gives me two different results at different times: I was not able to reproduce the error. I tried in two system, about 500,000 tests together. If there is a real heisenbug, it must depend on architechture etc. So Linux or Mac, what processor...? -- Jori Mäntysalo
Re: [sage-support] Installation using VMare Workstation on Windows 7 64-bit
On Sun, 8 Nov 2015, anga wrote: I did install--multiple times. First time, everything worked: the VM launched and displayed Sage interface. Couldn't install VMware tools. So, deleted VM and reinstalled: VM starts and displays the following: Network boot from AMD CLIENT MAC ADDR: xx xx xx xx xx xx GUID: PXE-E53: No boot filename received OK. For this I can not help. Sorry. Maybe someone else. -- Jori Mäntysalo
[sage-support] "Factoring" cartesian product of posets
Can Sage "factor" a poset, i.e. given poset P compute P1 and P2 such that
P is isomorphic to P1.product(P2)?
There is a module for that on undirected graphs:
http://doc.sagemath.org/html/en/reference/graphs/sage/graphs/graph_decompositions/graph_products.html
. However, how to differentiate situations where P1 might be Poset({0:[1,
2]}) or Poset({0: [1], 1: [2]})? They have isomorphic cover relations
graph.
--
Jori Mäntysalo
Re: [sage-support] Installation using VMare Workstation on Windows 7 64-bit
On Sat, 7 Nov 2015, anga wrote: Is there an guide for installing Sagemath 6.9 using VMware workstation 12 on Windows 7 64-bit? Did you try? .ova -file should work with VMware products; it is not a VirtualBox specific but a common format to export virtual machines. See https://en.wikipedia.org/wiki/Open_Virtualization_Format . 2. How to access the command prompt? See http://wiki.sagemath.org/SageAppliance#Using_the_Sage_shell . You need to know what is your "host key" on virtual machine manager. -- Jori Mäntysalo
Re: [sage-support] \leq converted to <=, can't we use unicode?
On Tue, 6 Oct 2015, Dima Pasechnik wrote: What we assume the user to have on command line? IMHO we can assume a UTF-8 capable terminal. OK. And as this example showed, even real Linux console has quite good support for unicode. (But it was a little surprise to have \cap but not \cup.) (Perhaps not for input, but for output for sure) Can this be a problem? I guess no, but theoretically one might want information about Möbius function without being able to write it. -- Jori Mäntysalo
Re: [sage-support] Re: Where is possible to view the code of one command? for example: lagrange_polynomial?
On Tue, 6 Oct 2015, Simon King wrote: Also this does not work with all (longer?) functions in the notebook. Try g = Graph() g.plot?? for an example of one kind of a bug. Works for me (on the command line at least). Yes, it works on command line and on jyputer notebook. But with older Sage notebook it does not. Same happens if you type @cached_method def f(): return 1+2 and then try f?? you see how it was redefined when you computed the matrix. You can not know that before trying or reading source code for few .py files. That's not a bug, because you see the source code. Not a bug from a technical viewpoint. But unexpected for a new user, and worth mentioning. -- Jori Mäntysalo
Re: [sage-support] \leq converted to <=, can't we use unicode?
On Mon, 5 Oct 2015, Dima Pasechnik wrote: In docstrings we have \leq converted to <=, can't we use unicode, and get ≤? More importantly, can we do similar things to \cap (∩) and \cup (∪), and perhaps even more of this? What we assume the user to have on command line? I just read your email in real Linux console, Ubuntu 14.04 LTS. \cap was OK, \cup was shown as a kind of diamond. In this gnome-terminal on X that I normally use for email everything is fine. Personally I don't mind the change, and I am sure that nobody whose computer support I am will complain. * * * Btw, I have corrected M\"obius function to Möbius function in docs. I hope that it is OK for everyone. (Linux has used US keyboard layout as a default from 0.03 I think. Before it was finnish layout. :=) ) -- Jori Mäntysalo
Re: [sage-support] Re: Where is possible to view the code of one command? for example: lagrange_polynomial?
On Mon, 5 Oct 2015, kcrisman wrote: sage: lagrange_polynomial?? # should be code Note that the latter doesn't always work (for instance Python built-ins, or some (all?) Cythonized functions) or so it seems to me. Also this does not work with all (longer?) functions in the notebook. Try g = Graph() g.plot?? for an example of one kind of a bug. * * * And of course it might be tricky to follow the code. For example P = Poset() P.mobius_function?? basically shows that you must look from the code of Hasse diagram. So you continue with P._hasse_diagram.mobius_function?? and see the function? Yes and no. If you say P = Poset() P.mobius_function_matrix() P.mobius_function?? you see how it was redefined when you computed the matrix. You can not know that before trying or reading source code for few .py files. -- Jori Mäntysalo
Re: [sage-support] Building Sage on Ubuntu 14.04 LTS 3
On Sun, 4 Oct 2015, Dmitrij Moreinis wrote: unfortunatelly the command: make doesnt work. I have added the install.log. As it says: "You cannot build Sage as root" -- Jori Mäntysalo
Re: [sage-support] CGI through sagemath
On Thu, 24 Sep 2015, avi kaur wrote: How to use sage through CGI? CGI in year 2015... I am calling sage inside CGI script but it is not executing. What I want to do is to get data through CGI and run sage. Permissions problem? Maybe try with "bash sage" instead of just "sage"? -- Jori Mäntysalo
Re: [sage-support] Re: correct derivative command returns syntax error
On Mon, 21 Sep 2015, Simon King wrote: What computer algebra systems (or pocket calculators) would understand that 5x means 5*x? Mathematica understand at least "5 x". Or at least understood in version 4. -- Jori Mäntysalo
Re: [sage-support] The sage notebook does not show results
On Tue, 1 Sep 2015, Oscar Lazo wrote: I am setting up a sage notebook server at my research institute to showcase software I have been writing in atomic physics. So I followed the instructions in http://wiki.sagemath.org/SageServer Point 6 says "I don't thing the following sudo syntax works anymore. I had to do this by hand now." Did you do that? 2+2 the notebook waits a little while and then shows no result. 2015-09-01 14:20:49-0500 [HTTPChannel,8,127.0.0.1] Permission denied (publickey,password). What happened on point 7 of instructions? If it worked, have you tried it with accounts sage1, sage2 and so on? I would also, for a start, change server_pool=['sage%d@localhost'%i for i in range(10)] to just server_pool=['sage0@localhost'] -- Jori Mäntysalo
Re: [sage-support] Plotting a q analogue function as a challenge?
On Thu, 13 Aug 2015, saad khalid wrote: I'm currently trying to get support from my professors in order for our school to move from Mathematica to Sage Math. One of them challenged me to - - What should we give as an exhange? Wasn't there some discussion about speed of gamma function on sage-devel, maybe a year ago? Or some graph theory question that is easy to do in Sage? -- Jori Mäntysalo
Re: [sage-support] Re: List of values for vertex_shape
On Fri, 29 May 2015, Nathann Cohen wrote: I just looked at the code and it seems that "vertex_shape" is an alias for the "marker" parameter from matplotlib. Thus, the possible shapes are there: http://matplotlib.org/api/markers_api.html We should add this information in the documentation - - Done this. (#18541) -- Jori Mäntysalo
[sage-support] List of values for vertex_shape
Plotting a graph has option vertex_shape. Where is the list of possible values? -- Jori Mäntysalo
Re: [sage-support] Re: graphics_array() and titles
On Tue, 26 May 2015, kcrisman wrote: But as has been pointed out on some relatively recent ticket, graphics_array needs an overhaul in any case and thus needs someone with the time and expertise to do that. Should you have that time I would be very happy to review! Uh, no. I know about nothing about generating graphics in general. Same with sage internals about graphics. -- Jori Mäntysalo
Re: [sage-support] Re: graphics_array() and titles
On Fri, 22 May 2015, kcrisman wrote: This prints "Another funcion" but not "Some funcion". Is this a bug or a feature? Almost certainly related to http://trac.sagemath.org/ticket/10657 Thanks. This is where I actually noticed this: http://trac.sagemath.org/ticket/18471 In general: should graphics_array work something like decorator pattern, so that it would have .plot() function and one could even combine arrays to bigger ones? -- Jori Mäntysalo
[sage-support] graphics_array() and titles
s=plot(sin(x), (0, 2*pi), title="Some funcion")
c=plot(cos(x), (0, 2*pi), title="Another funcion")
graphics_array([s, c])
This prints "Another funcion" but not "Some funcion". Is this a bug or a
feature?
* * *
How to print something like
TRIG FUNCTIONS
[fig-1-here] [fig-2-here]
[fig-1-here] [fig-2-here]
[fig-1-here] [fig-2-here]
sin cos
?
--
Jori Mäntysalo
Re: [sage-support] find_root for systems
On Thu, 7 May 2015, Dima Pasechnik wrote: nobody really knows how to solve systems of (non-polynomial) equations in general. There are heuristics implemented in various systems... What of those are built-in in Sage? -- Jori Mäntysalo
[sage-support] Simplifying binomials
I was asked where f(n,k)=(binomial(2*n + 2, n + 1) / 2 - binomial(2*n, n) - sum(binomial(n, k) * binomial(n, k-1), k,1,n)) f.full_simplify() got magic constant sqrt(pi). Right answer is zero. -- Jori Mäntysalo
Re: [sage-support] find_root for systems
On Tue, 5 May 2015, Paul Royik wrote:
How can this be applied to systems?
What kind of systems? Let us define f(x,y):
f(\sqrt{2}, \sqrt[3}) = 0
f(x,y) = x^2+y^2+1 if (x is not \sqrt{2}) or (y is not \sqrt[3})
Now this clearly has a root, but no numerical method can find it. So we
must have some assumptions.
--
Jori Mäntysalo
Re: [sage-support] Re: Default charset problem?
On Tue, 28 Apr 2015, Volker Braun wrote: We might be able to hack around this is the displayhook, but I'd rather not. It just is how Python 2.x works. Switching to Python 3 is the proper fix for this issue. OK. It's not that important feature. (Much more important is handling of len(), looping over a string and so on. But they are another story.) -- Jori Mäntysalo
[sage-support] Default charset problem?
On sagenb worksheet, Sage version 6.6 print "Direct translation of 'Mäntysalo' is 'Pine forest'." works as expected, but x="Direct translation of 'Mäntysalo' is 'Pine forest'"; x outputs "Direct translation of 'M\xc3\xa4ntysalo' is 'Pine forest'" Can this be changed? -- Jori Mäntysalo
