When polymake is called, its output is read as a string and then processed as follows:
the lines containing upper case letters are found. These are treated as lines containing the keywords. Each of those lines marks the beginning of a block of data.
The string is then cut into a list of blocks (also strings). Each block starts with a line containing the keyword and continues with some lines of data.
for each of the blocks, the appropriate function of ObjectConverters
is called. Here "appropriate" just means, that the keyword of the block coincides with the name of the function.
The output of the conversion function is then added to the known properties of the PolymakeObject
for which Polymake
was called.
The converter functions should take meaningful polymake data into meaningful GAP data. This sometimes means that the (mathematical) representation is changed. Here is an example: polymake writes vectors as augmented affine vectors of the form 1 a1 a2 a3...
which does not go very well with the usual GAP conventions of column vectors and multiplying matrices from the right. So polymaking converts such a vector to [a1,a2,a3,...]
and the user is left with the problem of augmentation and left or right multiplication.
Another area where the GAP object isn't a literal translation from the polymake world is combinatorics. In Polymake, list elements are enumerated starting from 0. GAP enumerates lists starting at 1. So the conversion process adds 1 to the numbers corresponding to vertices in facet lists, for example.
Call for help: As of version 0.7.0, most of the converters were assigned blindly to keywords. So some of them might not be suitable. If you find one of those, please tell me about it. Also, for a lot of keywords and data types, there is no converter. Please tell me if you want a new converter.
The conversion process is done by the following methods:
> ConvertPolymakeOutputToGapNotation ( string ) | ( method ) |
Returns: Record having polymake keywords as entry names and the respective converted polymake output as entries.
Given a the output of the polymake program as a string string, this method first calls SplitPolymakeOutputStringIntoBlocks
(4.1-3). For each of the returned blocks, the name (=first line) of the block is read and the record ObjectConverters
(4.1-4) is looked up for an entry with that name. If such an entry exists, it (being a function!) is called and passed the block. The returned value is then given the name of the block and added to the record returned by ConvertPolymakeOutputToGapNotation
.
> SplitPolymakeOutputStringIntoBlocks ( string ) | ( method ) |
Returns: List of strings -- "blocks"--
The string string is cut at the lines starting with an upper case character and consisting only of upper case letters, numbers and underscore (_) characters. The parts are returned as a list of strings. The initial string string remains unchanged.
> ObjectConverters | ( global variable ) |
The entries of this record are labeled by polymake keywords. Each of the entries is a function which converts a string returned by polymake to GAP format. So far, only a few converters are implemented. To see which, try RecNames(ObjectConverters);
You can define new converters using the basic functions described in section 4.2.
The following functions are used for the functions in ObjectConverters
(4.1-4).
> ConvertPolymakeNumber ( string ) | ( method ) |
The string string is converted to a rational number. Unlike Rat
, it tests, if the number represented by string is a floating point number an converts it correctly. If this is the case, a waring is issued.
> ConvertPolymakeScalarToGAP ( list ) | ( method ) |
If list contains a single string, this string is converted into a number using ConvertPolymakeNumber
(4.2-1).
> ConvertPolymakeMatrixOrListOfSetsToGAP ( list ) | ( method ) |
> ConvertPolymakeMatrixOrListOfSetsToGAPPlusOne ( list ) | ( method ) |
Tries to decide if the list list of strings represents a matrix or a list of sets by testing if they start with "{". It then calls either ConvertPolymakeMatrixToGAP
(4.2-4) or ConvertPolymakeListOfSetsToGAP
(4.2-8). The "PlusOne" version calls ConvertPolymakeListOfSetsToGAPPlusOne
(4.2-8) if list represents a list of sets.
> ConvertPolymakeMatrixToGAP ( list ) | ( method ) |
> ConvertPolymakeMatrixToGAPKillOnes ( list ) | ( method ) |
The list list of strings is interpreted as a list of row vectors and converted into a matrix. The "KillOnes" version removes the leading ones.
> ConvertPolymakeVectorToGAP ( list ) | ( method ) |
> ConvertPolymakeVectorToGAPKillOne ( list ) | ( method ) |
> ConvertPolymakeIntVectorToGAPPlusOne ( list ) | ( method ) |
As the corresponding "Matrix" version. Just for vectors. ConvertPolymakeIntVectorToGAPPlusOne
requires the vector to contain integers. It also adds 1 to every entry.
> ConvertPolymakeBoolToGAP ( list ) | ( method ) |
If list contains a single string, which is either 0,false,1, or true this function returns false
or true
, respectively.
> ConvertPolymakeSetToGAP ( list ) | ( method ) |
Let list be a list containing a single string, which is a list of numbers separated by whitespaces and enclosed by { and } . The returned value is then a set of rational numbers (in the GAP sense).
> ConvertPolymakeListOfSetsToGAP ( list ) | ( method ) |
> ConvertPolymakeListOfSetsToGAPPlusOne ( list ) | ( method ) |
Let list be a list containing several strings representing sets. Then each of these strings is converted to a set of rational numbers and the returned value is the list of all those sets. The "PlusOne" version adds 1 to every entry.
> ConvertPolymakeGraphToGAP ( list ) | ( method ) |
Let list be a list of strings representing sets (that is, a list of integers enclosed by { and }). Then a record is returned containing two sets named .vertices
and .edges
.
> ConvertPolymakeHasseDiagramToGAP ( list ) | ( method ) |
Returns: Record
with entries .hasse
and .faceindices
Takes a list list of strings (the HASSE_DIAGRAM
returned by polymake and split with SplitPolymakeOutputStringIntoBlocks
(4.1-3)) and returns a record containing a list of pairs .hasse
and a list of lists of integers .faceindices
. Every pair in .hasse
is of the form [[vertlist],[upfaces]]
where vertlist is a list of vertices (given as integers) and [upfaces]
is a list of integers encoding the faces in the next (higher) dimension as positions in the list .hasse
. If polymake returns special faces representing the full object or the empty face, they are also included in .hasse
.
.faceindices
is a list of lists of integers corresponding to the indices of faces in each dimension in .hasse
(possibly including the extra faces).
Note that the order in which the faces are listed in .hasse
is the order they are output from polymake. In particular, the list might start with vertices for one object and with facets for another.
gap> square:=CreatePolymakeObjectFromFile(Directory("~/tmp/"),"2cube.poly"); <polymake object. No properties known> gap> # "VERTICES" is converted using ConvertPolymakeMatrixToGAPKillOnes gap> Polymake(square,"VERTICES"); [ [ -1, -1 ], [ 1, -1 ], [ -1, 1 ], [ 1, 1 ] ] gap> Polymake(square,"HASSE_DIAGRAM"); rec( hasse := [ [ [ ], [ 2, 3, 4, 5 ] ], [ [ 1 ], [ 6, 8 ] ], [ [ 2 ], [ 7, 8 ] ], [ [ 3 ], [ 6, 9 ] ], [ [ 4 ], [ 7, 9 ] ], [ [ 1, 3 ], [ 10 ] ], [ [ 2, 4 ], [ 10 ] ], [ [ 1, 2 ], [ 10 ] ], [ [ 3, 4 ], [ 10 ] ], [ [ 1, 2, 3, 4 ], [ ] ] ], faceindices := [ [ 1 ], [ 2, 3, 4, 5 ], [ 6, 7, 8, 9 ], [ 10 ] ] ) gap> simp:=CreatePolymakeObjectFromFile(Directory("~/tmp/"),"tmp.top"); <polymake object. No properties known> gap> Polymake(simp,"FACETS"); ## a very small simplicial complex [ [ 2, 4 ], [ 1, 2, 3 ] ] gap> Polymake(simp,"HASSE_DIAGRAM"); rec( hasse := [ [ [ ], [ ] ], [ [ 1, 2, 3 ], [ 1 ] ], [ [ 2, 4 ], [ 1 ] ], [ [ 1, 2 ], [ 2 ] ], [ [ 1, 3 ], [ 2 ] ], [ [ 2, 3 ], [ 2 ] ], [ [ 2 ], [ 3, 4, 6 ] ], [ [ 4 ], [ 3 ] ], [ [ 1 ], [ 4, 5 ] ], [ [ 3 ], [ 5, 6 ] ], [ [ ], [ 7, 8, 9, 10 ] ] ], faceindices := [ [ 1 ], [ 2 ], [ 3, 4, 5, 6 ], [ 7, 8, 9, 10 ], [ 11 ] ] ) gap> ## ... and this one starts with facets |
generated by GAPDoc2HTML