[1X11 Internal functions[0X [1X11.1 Matrices as G-generators of a FG-module vector space[0X Both [9XFpGModuleGF[0m (Chapter [14X5[0m) and [9XFpGModuleHomomorphismGF[0m (Chapter [14X6[0m) store a matrix whose rows are G-generators for a module vector space (the module and the homomorphism's image respectively). The internal functions listed here provide common operations for dealing with these matrices. [1X11.1-1 HAPPRIME_ValueOptionMaxFGExpansionSize[0m [2X> HAPPRIME_ValueOptionMaxFGExpansionSize( [0X[3Xfield, group[0X[2X ) __________[0Xoperation [6XReturns:[0X Integer Returns the maximum matrix expansion size. This is read from the [10XMaxFGExpansionSize[0m option from the [5XGAP[0m options stack [14X'Reference: Options Stack'[0m, computed using the [10XMaxFGExpansionMemoryLimit[0m option. [1X11.1-2 HAPPRIME_KernelOfGeneratingRowsDestructive[0m [2X> HAPPRIME_KernelOfGeneratingRowsDestructive( [0X[3Xgens, rowlengths, GA[0X[2X ) [0Xoperation [6XReturns:[0X List Returns a list of generating vectors for the kernel of the FG-module homomorphism defined by the generating rows [3Xgens[0m using the group and action [3XGA[0m. This function computes the kernel recursively by partitioning the generating rows into [ B 0 ] [ C D ] doing column reduction if necessary to get the zero block at the top right. The matrices B> and C are small enough to be expanded, while the kernel of D is calculated by recursion. The argument [3Xrowlengths[0m lists the number of non-zero blocks in each row; the rest of each row is taken to be zero. This allows the partitioning to be more efficiently performed (i.e. column reduction is not always required). The [5XGAP[0m options stack [14X'Reference: Options Stack'[0m variable [10XMaxFGExpansionSize[0m can be used to specify the maximum allowable expanded matrix size. This governs the size of the B and C matrices, and thus the number of recursions before the kernel of D is also computed by recursion. A high value for will allow larger expansions and so faster computation at the cost of more memory. The [10XMaxFGExpansionMemoryLimit[0m option can also be used, which sets the maximum amount of memory that [5XGAP[0m is allowed to use (as a string containing an integer with the suffix [10Xk[0m, [10XM[0m or [10XG[0m to indicate kilobyes, megabytes or gigabytes respectively). In this case, the function looks at the free memory available to [5XGAP[0m and computes an appropriate value for [10XMaxFGExpansionSize[0m. [1X11.1-3 HAPPRIME_GActMatrixColumns[0m [2X> HAPPRIME_GActMatrixColumns( [0X[3Xg, Vt, GA[0X[2X ) _________________________[0Xoperation [2X> HAPPRIME_GActMatrixColumnsOnRight( [0X[3Xg, Vt, GA[0X[2X ) __________________[0Xoperation [6XReturns:[0X Matrix Returns the matrix that results from the applying the group action u=gv (or u=vg in the case of the [10XOnRight[0m version of this function) to each [13Xcolumn[0m vector in the matrix [3XVt[0m. By acting on [13Xcolumns[0m of a matrix (i.e. the transpose of the normal [5XGAP[0m representation), the group action is just a permutation of the rows of the matrix, which is a fast operation. The group and action are passed in [3XGA[0m using the [2XModuleGroupAndAction[0m ([14X5.4-5[0m) record. If the input matrix [3XVt[0m is in a compressed matrix representation, then the returned matrix will also be in compressed matrix representation. [1X11.1-4 HAPPRIME_ExpandGeneratingRow[0m [2X> HAPPRIME_ExpandGeneratingRow( [0X[3Xgen, GA[0X[2X ) _________________________[0Xoperation [2X> HAPPRIME_ExpandGeneratingRows( [0X[3Xgens, GA[0X[2X ) _______________________[0Xoperation [2X> HAPPRIME_ExpandGeneratingRowOnRight( [0X[3Xgen, GA[0X[2X ) __________________[0Xoperation [2X> HAPPRIME_ExpandGeneratingRowsOnRight( [0X[3Xgens, GA[0X[2X ) ________________[0Xoperation [6XReturns:[0X List Returns a list of G-generators for the vector space that corresponds to the of G-generator [3Xgen[0m (or generators [3Xgens[0m). This space is formed by multiplying each generator by each element of [3XG[0m in turn, using the group and action specified in [3XGA[0m (see [2XModuleGroupAndAction[0m ([14X5.4-5[0m)). The returned list is thus |G| times larger than the input. For a list of generators [3Xgens[0m [v_1, v_2, ..., v_n], [2XHAPPRIME_ExpandGeneratingRows[0m returns the list [g_1v_1, g_2v_1, ..., g_1v_2, g_2v_2, ..., g_|G|v_n] In other words, the form of the returned matrix is block-wise, with the expansions of each row given in turn. This function is more efficient than repeated use of [2XHAPPRIME_ExpandGeneratingRow[0m since it uses the efficient [2XHAPPRIME_GActMatrixColumns[0m ([14X11.1-3[0m) to perform the group action on the whole set of generating rows at a time. The function [2XHAPPRIME_ExpandGeneratingRowsOnRight[0m is the same as above, but the group action operates on the right instead. [1X11.1-5 HAPPRIME_AddGeneratingRowToSemiEchelonBasisDestructive[0m [2X> HAPPRIME_AddGeneratingRowToSemiEchelonBasisDestructive( [0X[3Xbasis, gen, GA[0X[2X ) [0Xoperation [6XReturns:[0X Record with elements [10Xvectors[0m and [10Xbasis[0m This function augments a vector space basis with another generator. It returns a record consisting of two elements: [10Xvectors[0m, a set of semi-echelon basis vectors for the vector space spanned by the sum of the input [3Xbasis[0m and all G-multiples of the generating vector [3Xgen[0m; and [10Xheads[0m, a list of the head elements, in the same format as returned by [2XSemiEchelonMat[0m ([14XReference: SemiEchelonMat[0m). The generator [3Xgen[0m is expanded according to the group and action specified in the [3XGA[0m record (see [2XModuleGroupAndAction[0m ([14X5.4-5[0m)). If the input [3Xbasis[0m is not zero, it is also modified by this function, to be the new basis (i.e. the same as the [10Xvectors[0m element of the returned record). [1X11.1-6 HAPPRIME_ReduceVectorDestructive[0m [2X> HAPPRIME_ReduceVectorDestructive( [0X[3Xv, basis, heads[0X[2X ) _____________[0Xoperation [6XReturns:[0X Boolean Reduces the vector [3Xv[0m (in-place) using the semi-echelon set of vectors [3Xbasis[0m with heads [3Xheads[0m (as returned by [2XSemiEchelonMat[0m ([14XReference: SemiEchelonMat[0m)). Returns [9Xtrue[0m if the vector is completely reduced to zero, or [9Xfalse[0m otherwise. [1X11.1-7 HAPPRIME_ReduceGeneratorsOfModuleByXX[0X [2X> HAPPRIME_ReduceGeneratorsOfModuleBySemiEchelon( [0X[3Xgens, GA[0X[2X ) ______[0Xoperation [2X> HAPPRIME_ReduceGeneratorsOfModuleBySemiEchelonDestructive( [0X[3Xgens, GA[0X[2X ) [0Xoperation [2X> HAPPRIME_ReduceGeneratorsOfModuleByLeavingOneOut( [0X[3Xgens, GA[0X[2X ) ____[0Xoperation [2X> HAPPRIME_ReduceGeneratorsOnRightByLeavingOneOut( [0X[3Xgens, GA[0X[2X ) _____[0Xoperation [6XReturns:[0X List of vectors Returns a subset of the module generators [3Xgens[0m over the group with action specified in the [3XGA[0m record (see [2XModuleGroupAndAction[0m ([14X5.4-5[0m)) that will still generate the module. The [10XBySemiEchelon[0m functions gradually expand out the module generators into an F-basis, using that F-basis to reduce the other generators, until the full vector space of the module is spanned. The generators needed to span the space are returned, and should be a small set, although not minimal. The [10XDestructive[0m version of this function will modify the input [3Xgens[0m parameter. The non-destructive version makes a copy first, so leaves the input arguments unchanged, at the expense of more memory. The [10XByLeavingOneOut[0m function is tries repeatedly leaving out generators from the list [3Xgens[0m to find a small subset that still generates the module. If the generators are from the field GF(2), this is guaranteed to be a minimal set of generators. The [10XOnRight[0m version computes a minimal subset which generates the module under group multiplication on the right. [1X11.1-8 HAPPRIME_DisplayGeneratingRows[0m [2X> HAPPRIME_DisplayGeneratingRows( [0X[3Xgens, GA[0X[2X ) ______________________[0Xoperation [6XReturns:[0X nothing Displays a set of G-generating rows a human-readable form. The elements of each generating vector are displayed, with each block marked by a separator (since the group action on a module vector will only permute elements within a block). This function is used by [9XDisplay[0m for both [9XFpGModuleGF[0m and [9XFpGModuleHomomorphismGF[0m. NOTE: This is currently only implemented for GF(2) [4X--------------------------- Example ----------------------------[0X [4Xgap> HAPPRIME_DisplayGeneratingRows([0X [4X> ModuleGenerators(M), HAPPRIME_ModuleGroupAndAction(M));[0X [4X[...1..11|........|.......1|........|........][0X [4X[........|........|........|.1....11|........][0X [4X[........|........|........|........|..1.1.1.][0X [4X[........|.1.1..1.|........|........|........][0X [4X[........|........|......11|........|........][0X [4X[........|........|1......1|........|........][0X [4X------------------------------------------------------------------[0X [1X11.1-9 HAPPRIME_GeneratingRowsBlockStructure[0m [2X> HAPPRIME_GeneratingRowsBlockStructure( [0X[3Xgens, GA[0X[2X ) _______________[0Xoperation [6XReturns:[0X Matrix Returns a matrix detailing the block structure of a set of module generating rows. The group action on a generator permutes the vector in blocks of length [10XGA.actionBlockSize[0m: any block that contains non-zero elements will still contain non-zero elements after the group action; any block that is all zero will remain all zero. This operation returns a matrix giving this block structure: it has a one where the block is non-zero and zero where the block is all zero. [4X--------------------------- Example ----------------------------[0X [4Xgap> b := HAPPRIME_GeneratingRowsBlockStructure([0X [4X> ModuleGenerators(M), ModuleActionBlockSize(M));[0X [4X[ [ 1, 0, 1, 1, 1 ], [ 1, 0, 1, 1, 1 ], [ 0, 1, 1, 1, 1 ], [ 0, 0, 1, 1, 1 ] ][0X [4X------------------------------------------------------------------[0X [1X11.1-10 HAPPRIME_DisplayGeneratingRowsBlocks[0m [2X> HAPPRIME_DisplayGeneratingRowsBlocks( [0X[3Xgens, actionBlockSize[0X[2X ) ___[0Xoperation [6XReturns:[0X nothing Displays a set of G-generating rows a compact human-readable form. Each generating rows can be divided into blocks of length [3XactionBlockSize[0m. The generating rows are displayed in a per-block form: a [9X*[0m where the block is non-zero and [9X.[0m where the block is all zero. This function is used by [2XDisplayBlocks[0m ([14X5.4-10[0m) (for [9XFpGModuleGF[0m) and [2XDisplayBlocks[0m ([14X6.5-4[0m) (for [9XFpGModuleHomomorphismGF[0m). [4X--------------------------- Example ----------------------------[0X [4Xgap> HAPPRIME_DisplayGeneratingRowsBlocks([0X [4X> ModuleGenerators(M), HAPPRIME_ModuleGroupAndAction(M));[0X [4X[*.*..][0X [4X[...*.][0X [4X[....*][0X [4X[.*...][0X [4X[..*..][0X [4X[..*..] [0X [4X------------------------------------------------------------------[0X [1X11.1-11 HAPPRIME_IndependentGeneratingRows[0m [2X> HAPPRIME_IndependentGeneratingRows( [0X[3Xblocks[0X[2X ) ____________________[0Xoperation [6XReturns:[0X List of lists Given a block structure as returned by [2XHAPPRIME_GeneratingRowsBlockStructure[0m ([14X11.1-9[0m), this decomposes a set of generating rows into sets of independent rows. These are returned as a list of row indices, where each set of rows share no blocks with any other set. [4X--------------------------- Example ----------------------------[0X [4Xgap> DisplayBlocks(M);[0X [4XModule over the group ring of Group( [ f1, f2, f3 ] )[0X [4X in characteristic 2 with 6 generators in FG^5.[0X [4X[**...][0X [4X[.*...][0X [4X[.**..][0X [4X[.**..][0X [4X[...*.][0X [4X[....*][0X [4XGenerators are in minimal echelon form.[0X [4Xgap> gens := ModuleGenerators(M);;[0X [4Xgap> G := ModuleGroup(M);;[0X [4Xgap> blocks := HAPPRIME_GeneratingRowsBlockStructure(gens, G);[0X [4X[ [ 1, 1, 0, 0, 0 ], [ 0, 1, 0, 0, 0 ], [ 0, 1, 1, 0, 0 ], [ 0, 1, 1, 0, 0 ],[0X [4X [ 0, 0, 0, 1, 0 ], [ 0, 0, 0, 0, 1 ] ][0X [4Xgap> HAPPRIME_IndependentGeneratingRows(blocks);[0X [4X[ [ 1, 2, 3, 4 ], [ 5 ], [ 6 ] ][0X [4X------------------------------------------------------------------[0X [1X11.1-12 HAPPRIME_GactFGvector[0m [2X> HAPPRIME_GactFGvector( [0X[3Xg, v, MT[0X[2X ) _______________________________[0Xoperation [6XReturns:[0X Vector Returns the vector that is the result of the action u=gv of the group element [3Xg[0m on a module vector [3Xv[0m (according to the group multiplication table [3XMT[0m. This operation is the quickest current method for a single vector. To perform the same action on a set of vectors, it is faster write the vectors as columns of a matrix and use [2XHAPPRIME_GActMatrixColumns[0m ([14X11.1-3[0m) instead. [1X11.1-13 HAPPRIME_CoefficientsOfGeneratingRowsXX[0X [2X> HAPPRIME_CoefficientsOfGeneratingRows( [0X[3Xgens, GA, v[0X[2X ) ____________[0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRows( [0X[3Xgens, GA, coll[0X[2X ) _________[0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsDestructive( [0X[3Xgens, GA, v[0X[2X ) _[0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsDestructive( [0X[3Xgens, GA, coll[0X[2X ) [0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGF( [0X[3Xgens, GA, v[0X[2X ) __________[0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGF( [0X[3Xgens, GA, coll[0X[2X ) _______[0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGFDestructive( [0X[3Xgens, GA, v[0X[2X ) [0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGFDestructive( [0X[3Xgens, GA, coll[0X[2X ) [0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGFDestructive2( [0X[3Xgens, GA, v[0X[2X ) [0Xoperation [2X> HAPPRIME_CoefficientsOfGeneratingRowsGFDestructive2( [0X[3Xgens, GA, coll[0X[2X ) [0Xoperation [6XReturns:[0X Vector, or list of vectors For a single vector [3Xv[0m, this function returns a vector x giving the G-coefficients from [3Xgens[0m needed to generate [3Xv[0m, i.e. the solution to the equation x*A=v, where A is the expansion of [3Xgens[0m. If there is no solution, [9Xfail[0m is returned. If a list of vectors, [3Xcoll[0m, then a vector is returned that lists the solution for each vector (any of which may be [9Xfail[0m). The standard forms of this function use standard linear algebra to solve for the coefficients. The [10XDestructive[0m version will corrupt both [3Xgens[0m and [3Xv[0m. The [10XGF[0m versions use the block structure of the generating rows to expand only the blocks that are needed to find the solution before using linear algebra. If the generators are in echelon form, this can save memory, but is slower. The [10XGFDestructive2[0m functions also assume an echelon form for the generators, but use back-substitution to find a set of coefficients. This can save a lot of memory but is again slower. [1X11.1-14 HAPPRIME_GenerateFromGeneratingRowsCoefficientsXX[0X [2X> HAPPRIME_GenerateFromGeneratingRowsCoefficients( [0X[3Xgens, GA, c[0X[2X ) __[0Xoperation [2X> HAPPRIME_GenerateFromGeneratingRowsCoefficients( [0X[3Xgens, GA, coll[0X[2X ) [0Xoperation [2X> HAPPRIME_GenerateFromGeneratingRowsCoefficientsGF( [0X[3Xgens, GA, c[0X[2X ) [0Xoperation [2X> HAPPRIME_GenerateFromGeneratingRowsCoefficientsGF( [0X[3Xgens, GA, coll[0X[2X ) [0Xoperation [6XReturns:[0X Vector, or list of vectors For a vector [3Xc[0m, returns (as a vector), the module element generated by multiplying [3Xc[0m by the expansion of the generators [3Xgens[0m. For a list of coefficient vectors [3Xcoll[0m, this returns a list of generating vectors. The standard versions of this function use standard linear algebra. The [10XGF[0m versions only performs the expansion of necessary generating rows, and only expands by one group element at a time, so will only need at most twice the amount of memory as that to store [3Xgens[0m, which is a large saving over expanding the generators by every group element at the same time, as in a naive implementation. It may also be faster. [1X11.1-15 HAPPRIME_RemoveZeroBlocks[0m [2X> HAPPRIME_RemoveZeroBlocks( [0X[3Xgens, GA[0X[2X ) ___________________________[0Xoperation [6XReturns:[0X Vector Removes from a set of generating vectors [3Xgens[0m (with [2XModuleGroupAndAction[0m ([14X5.4-5[0m) [3XGA[0m) any blocks that are zero in every generating vector. Removal is done in-place, i.e. the input argument [3Xgens[0m will be modified to remove the zero blocks. Zero blocks are unaffected by any row or expansion operation, and can be removed to save time or memory in those operations. The function returns the original block structure as a vector, and this can be used in the function [2XHAPPRIME_AddZeroBlocks[0m ([14X11.1-16[0m) to reinstate the zero blocks later, if required. See the documentation for that function for more detail of the block structure vector. [1X11.1-16 HAPPRIME_AddZeroBlocks[0m [2X> HAPPRIME_AddZeroBlocks( [0X[3Xgens, blockStructure, GA[0X[2X ) ______________[0Xoperation [6XReturns:[0X List of vectors Adds zero blocks to a set of generating vectors [3Xgens[0m to make it have the block structure given in [3XblockStructure[0m (for a given [2XModuleGroupAndAction[0m ([14X5.4-5[0m) [3XGA[0m). The generators [3Xgens[0m are modified in place, and also returned. The [3XblockStructure[0m parameter is a vector of which is the length of the required output vector and has zeros where zero blocks should be, and is non-zero elsewhere. Typically, an earlier call to [2XHAPPRIME_RemoveZeroBlocks[0m ([14X11.1-15[0m) will have been used to remove the zero blocks, and this function and such a [3XblockStructure[0m vector is returned by this function. [2XHAPPRIME_AddZeroBlocks[0m can be used to reinstate these zero blocks. [1X11.2 FG-modules[0X FG-modules in [5XHAPprime[0m use the datatype [9XFpGModuleGF[0m (Chapter [14X5[0m). Internally, this uses many of the functions listed in Section [14X11.1[0m, and further internal functions are listed below. [1X11.2-1 HAPPRIME_DirectSumForm[0m [2X> HAPPRIME_DirectSumForm( [0X[3Xcurrent, new[0X[2X ) __________________________[0Xoperation [6XReturns:[0X String Returns a string containing the form of the generator matrix if the direct sum is formed between a [9XFpGModuleGF[0m with the form [3Xcurrent[0m and a [9XFpGModuleGF[0m with the form [3Xnew[0m. The direct sum is formed by placing the two module generating matrices in diagonal form. Given the form of the two generating matrices, this allows the form of the direct sum to be stated. See [2XModuleGeneratorsForm[0m ([14X5.5-5[0m) for information about form strings. [1X11.2-2 HAPPRIME_PrintModuleDescription[0m [2X> HAPPRIME_PrintModuleDescription( [0X[3XM, func[0X[2X ) ______________________[0Xoperation [6XReturns:[0X nothing Used by [2XPrintObj[0m ([14XReference: PrintObj[0m), [2XViewObj[0m ([14XReference: ViewObj[0m), [2XDisplay[0m ([14XReference: Display[0m) and [2XDisplayBlocks[0m ([14X5.4-10[0m), this helper function prints a description of the module [3XM[0m. The parameter [3Xfunc[0m can be one of the strings [10X"print"[0m, [10X"view"[0m, [10X"display"[0m or [10X"displayblocks"[0m, corresponding to the print different functions that might be called. [1X11.2-3 HAPPRIME_ModuleGeneratorCoefficients[0X [2X> HAPPRIME_ModuleGeneratorCoefficients( [0X[3XM, elm[0X[2X ) __________________[0Xoperation [2X> HAPPRIME_ModuleGeneratorCoefficientsDestructive( [0X[3XM, elm[0X[2X ) _______[0Xoperation [2X> HAPPRIME_ModuleGeneratorCoefficients( [0X[3XM, coll[0X[2X ) _________________[0Xoperation [2X> HAPPRIME_ModuleGeneratorCoefficientsDestructive( [0X[3XM, coll[0X[2X ) ______[0Xoperation [6XReturns:[0X Vector Returns the coefficients needed to make the module element [3Xelm[0m as a linear and G-combination of the module generators of the [9XFpGModuleGF[0m [3XM[0m. The coefficients are returned in standard vector form, or if there is no solution then [10Xfail[0m is returned. If a list of elements is given, then a list of coefficients (or [10Xfails[0m) is returned. The [10XDestructive[0m form of this function might change the elements of of [3XM[0m or [3Xelm[0m. The non-[10XDestructive[0m version makes copies to ensure that they are not changed. See also [2XHAPPRIME_ModuleElementFromGeneratorCoefficients[0m ([14X11.2-4[0m). [1X11.2-4 HAPPRIME_ModuleElementFromGeneratorCoefficients[0X [2X> HAPPRIME_ModuleElementFromGeneratorCoefficients( [0X[3XM, c[0X[2X ) _________[0Xoperation [2X> HAPPRIME_ModuleElementFromGeneratorCoefficients( [0X[3XM, coll[0X[2X ) ______[0Xoperation [6XReturns:[0X Vector Returns an element from the module [3XM[0m, constructed as a linear and G-sum of the module generators as specified in [3Xc[0m. If a list of coefficient vectors is given, a list of corresponding module elements is returned. See also [2XHAPPRIME_ModuleGeneratorCoefficients[0m ([14X11.2-3[0m) [1X11.2-5 HAPPRIME_MinimalGeneratorsVectorSpaceGeneratingRowsDestructive[0m [2X> HAPPRIME_MinimalGeneratorsVectorSpaceGeneratingRowsDestructive( [0X[3Xvgens, GA[0X[2X ) [0Xoperation [2X> HAPPRIME_MinimalGeneratorsVectorSpaceGeneratingRowsOnRightDestructive( [0X[3Xvgens, GA[0X[2X ) [0Xoperation [6XReturns:[0X FpGModuleGF Returns a module with minimal generators that is equal to the FG-module with [13Xvector space[0m basis [3Xvgens[0m and [2XModuleGroupAndAction[0m ([14X5.4-5[0m) as specified in [3XGA[0m. The solution is computed by the module radical method, which is fast at the expense of memory. This function will corrupt the matrix [3Xgens[0m. This is a helper function for [2XMinimalGeneratorsModuleRadical[0m ([14X5.5-9[0m) that is also used by [2XExtendResolutionPrimePowerGroupRadical[0m ([14XHAPprime: ExtendResolutionPrimePowerGroupRadical[0m) (which knows that its module is already in vector-space form). [1X11.2-6 HAPPRIME_IsGroupAndAction[0m [2X> HAPPRIME_IsGroupAndAction( [0X[3Xobj[0X[2X ) ________________________________[0Xoperation [6XReturns:[0X Boolean Returns [10Xtrue[0m if [3Xobj[0m appears to be a [10XgroupAndAction[0m record (see [2XModuleGroupAndAction[0m ([14X5.4-5[0m)), or [10Xfalse[0m otherwise. [1X11.3 Resolutions[0X For details of the main resolution functions in [5XHAPprime[0m, see Chapter [14X2[0m of this datatypes reference manual, and [14X'HAPprime: Resolutions'[0m in the [5XHAPprime[0m user guide. This section describes the internal helper functions used by the higher-level functions. [1X11.3-1 HAPPRIME_WordToVector[0m [2X> HAPPRIME_WordToVector( [0X[3Xw, dim, orderG[0X[2X ) ____________________________[0Xmethod [6XReturns:[0X [5XHAP[0m word (list of lists) Returns the boundary map vector that corresponds to the [5XHAP[0m word vector [3Xw[0m with module ambient dimension [3Xdim[0m and group order [3XorderG[0m (assumed to be the [10XactionBlockSize[0m). A [5XHAP[0m word vector has the following format: [10X[ [block, elm], [block, elm], ... ][0m where [10Xblock[0m is a block number and [10Xelm[0m is a group element index (see example below). See also [2XHAPPRIME_VectorToWord[0m ([14X11.3-2[0m) [4X--------------------------- Example ----------------------------[0X [4Xgap> G := CyclicGroup(4);;[0X [4Xgap> v := HAPPRIME_WordToVector([ [1,2],[2,3] ], 2, Order(G));[0X [4X<a GF2 vector of length 8>[0X [4Xgap> HAPPRIME_DisplayGeneratingRows([v], CanonicalGroupAndAction(G));[0X [4X[.1..|..1.][0X [4Xgap> HAPPRIME_VectorToWord(v, Order(G));[0X [4X[ [ 1, 2 ], [ 2, 3 ] ][0X [4X------------------------------------------------------------------[0X [1X11.3-2 HAPPRIME_VectorToWord[0m [2X> HAPPRIME_VectorToWord( [0X[3Xvec, orderG[0X[2X ) _____________________________[0Xfunction [6XReturns:[0X Vector The [5XHAP[0m word format vector that corresponds to the boundary vector [3Xvec[0m with [10XactionBlockSize[0m assumed to be [3XorderG[0m. See [2XHAPPRIME_WordToVector[0m ([14X11.3-1[0m) for a few more details and an example. [1X11.3-3 HAPPRIME_BoundaryMatrices[0m [2X> HAPPRIME_BoundaryMatrices( [0X[3XR[0X[2X ) __________________________________[0Xattribute [6XReturns:[0X List of matrices If [3XR[0m is a resolution which stores its boundaries as a list of matrices (e.g. one created by [5XHAPprime[0m, this list is returned. Otherwise, [10Xfail[0m is returned. Note that the first matrix in this list corresponds to the zeroth degree: for resolutions of modules, this is the generators of the module; for resolutions of groups, this is the empty matrix. The second matrix corresponds to the first degree, and so on. [1X11.3-4 HAPPRIME_AddNextResolutionBoundaryMapMatNC[0m [2X> HAPPRIME_AddNextResolutionBoundaryMapMatNC( [0X[3XR, BndMat[0X[2X ) _________[0Xoperation [6XReturns:[0X [10XHapResolution[0m Returns the resolution [3XR[0m extended by one term, where that term is given by the boundary map matrix [3XBndMat[0m. If [3XBndMat[0m is not already in compressed matrix form, it will be converted into this form, and if the boundaries in [3XR[0m are not already in matrix form, they are all converted into this form. [1X11.3-5 HAPPRIME_CreateResolutionWithBoundaryMapMatsNC[0m [2X> HAPPRIME_CreateResolutionWithBoundaryMapMatsNC( [0X[3XG, BndMats[0X[2X ) ____[0Xoperation [6XReturns:[0X [10XHapResolution[0m Returns a [5XHAP[0m resolution object for group [3XG[0m where the module homomorphisms are given by the boundary matrices in the list [3XBndMats[0m. This list is indexed with the boundary matrix for degree [13Xzero[0m as the first element. If the resolution is the resolution of a module, the module's minimal generators are this first boundary matrix, otherwise (for the resolution of a group), this should be set to be the empty matrix [10X[][0m. [1X11.4 Polynomial rings[0X Ths internal functions in this section implement isomorphisms between polynomial rings via maps from one set of indeterminates to another. [1X11.4-1 HAPPRIME_SwitchPolynomialIndeterminates[0m [2X> HAPPRIME_SwitchPolynomialIndeterminates( [0X[3XR, S, poly[0X[2X ) ____________[0Xfunction [6XReturns:[0X Polynomial or List of Polynomials Changes the indeterminates in [3Xpoly[0m, which should be a polynomial or a list of polynomials, substituting the indeterminates of the polynomial ring [3XS[0m one-for-one for those in [3XR[0m (from which all polynomials in [3Xpoly[0m must come). The returned object is either a polynomial or a list of polynomials in the new indeterminates, depending on the input object. See [2XHAPPRIME_MapPolynomialIndeterminates[0m ([14X11.4-2[0m) for a function that can work with more general indeterminate maps. [1X11.4-2 HAPPRIME_MapPolynomialIndeterminates[0m [2X> HAPPRIME_MapPolynomialIndeterminates( [0X[3Xold, new, poly[0X[2X ) ___________[0Xfunction [6XReturns:[0X Polynomial or List of Polynomials Changes the indeterminates in [3Xpoly[0m, which can be a polynomial or a list of polynomials, substituting the polynomials in [3Xold[0m for those in [3Xnew[0m. The returned object is either a polynomial or a list of polynomials in the new indeterminates, depending on the input object. The change of variable arguments, [3Xold[0m and [3Xnew[0m, do not have to be simply indeterminates: they can be can be lists of polynomials which are equivalent in the two different sets of indeterminates. If a polynomial cannot be converted (i.e. if it cannot be generated from the polynomials in [3Xold[0m) then [9Xfail[0m is returned for that polynomial. [1X11.4-3 HAPPRIME_CombineIndeterminateMaps[0m [2X> HAPPRIME_CombineIndeterminateMaps( [0X[3Xcoeff, M, N[0X[2X ) _________________[0Xfunction [6XReturns:[0X List Returns the indeterminate map that results from applying map [3XM[0m followed by map [3XN[0m. An indeterminate map is a list containing two lists, the first of which is a list of polynomials in the original indeterminates, the second the equivalent polynomials in the new ring indeterminates. [1X11.5 Gröbner Bases[0X Ths internal functions in this section provide a transparent way of using Singular's Gröbner basis functions, if they are available. See the [5Xsingular[0m package documentation [14X'singular: singular: the GAP interface to Singular'[0m for further details. [1X11.5-1 HAPPRIME_SingularGroebnerBasis[0m [2X> HAPPRIME_SingularGroebnerBasis( [0X[3Xpols, O[0X[2X ) ________________________[0Xfunction [2X> HAPPRIME_SingularReducedGroebnerBasis( [0X[3Xpols, O[0X[2X ) _________________[0Xfunction [6XReturns:[0X List Returns the Gröbner basis (or reduced Gröbner basis) with respect to the ordering [3XO[0m for the ideal generated by the polynomials [3Xpols[0m. This function uses the Gröbner basis implementation from [5Xsingular[0m, for preference, if available ([2XGroebnerBasis[0m ([14Xsingular: GroebnerBasis[0m)), and if so it also manuipulates the result to fix a bug in [5Xsingular[0m where the returned polynomials are not necessarily returned with a value external representation (see [14X'Reference: The Defining Attributes of Rational Functions'[0m). If the option [10XobeyGBASIS[0m is [9Xtrue[0m, then this function will use whichever algorithm is specified by the [9XGBASIS[0m global variable (see [2XSINGULARGBASIS[0m ([14Xsingular: SINGULARGBASIS[0m)). [1X11.6 Presentations of graded algebras[0X For the main functions dealing with presentations of graded algebras, and details of the datatype see Chapter [14X4[0m. This section details the internal functions used by the higher-level functions. [1X11.6-1 HAPPRIME_HilbertSeries[0m [2X> HAPPRIME_HilbertSeries( [0X[3XP[0X[2X ) _____________________________________[0Xattribute [6XReturns:[0X List The Hilbert-Poincaré series for a graded ring is a polynomial H_P(t) whose coefficients are the dimensions of each degree of the ring. It can be written as H_P(t) = Q(t) / (1-t^a)(1-t^b)...(1-t^n) This function returns the list of coefficients for the polynomial Q(t) in the Hilbert Series for the graded algebra with presentation [3XP[0m. This function simply calls the Singular function [10Xhilb[0m via the [5Xsingular[0m package and returns the result. [1X11.6-2 HAPPRIME_SwitchGradedAlgebraRing[0m [2X> HAPPRIME_SwitchGradedAlgebraRing( [0X[3XA, R[0X[2X ) ________________________[0Xattribute [6XReturns:[0X GradedAlgebraPresentation Returns a new presentation for the graded algebra [3XA[0m which uses the ring [3XR[0m instead of the one in [3XA[0m. [1X11.6-3 HAPPRIME_GradedAlgebraPresentationAvoidingIndeterminates[0m [2X> HAPPRIME_GradedAlgebraPresentationAvoidingIndeterminates( [0X[3XA, avoid[0X[2X ) [0Xattribute [6XReturns:[0X GradedAlgebraPresentation Returns a new presentation for the graded algebra [3XA[0m which avoids the indeterminates listed in [3Xavoid[0m. [1X11.6-4 HAPPRIME_LastLHSBicomplexSize[0m [2X> HAPPRIME_LastLHSBicomplexSize______________________________[0Xglobal variable Stores the last bicomplex size last used by [2XLHSSpectralSequenceLastSheet[0m ([14X4.4-10[0m). Used for testing and generating results for paper. [1X11.6-5 HAPPRIME_LHSSpectralSequence[0m [2X> HAPPRIME_LHSSpectralSequence( [0X[3XG, N, n[0X[2X ) _________________________[0Xoperation [6XReturns:[0X [9XGradedAlgebraPresentation[0m or list This function is called by both [2XLHSSpectralSequence[0m ([14X4.4-10[0m) and [2XLHSSpectralSequenceLastSheet[0m ([14X4.4-10[0m) and does the actual work of computing the Lyndon-Hoschild-Serre spectral sequence for the group extension N -> G -> G/N. See the documentation for those functions for the main details, including options that are passed through to this function. If [3Xn[0m is the string [10X"Einf"[0m then only the limiting sheet is returned (and the other sheets are discarded during computation, which can save time and memory). Otherwise, [3Xn[0m sheets are returned, or enough until convergence is proved, if that is smaller. [1X11.6-6 HAPPRIME_CohomologyRingWithoutResolution[0m [2X> HAPPRIME_CohomologyRingWithoutResolution( [0X[3XG[0X[2X ) ____________________[0Xfunction [6XReturns:[0X [9XGradedAlgebraPresentation[0m or [9Xfail[0m If the mod-p cohomology ring for [3XG[0m can be computed without building a resolution for [3XG[0m then this ring is returned, otherwise this function returns [9Xfail[0m. Current cases where the resolution for [3XG[0m is not needed are -- if [3XG[0m is the group of order two then H^*(G, F) = F[x] where x has degree one -- if [3XG[0m is cyclic of order greater than two then H^*(G, F) = F[x,y]/x^2 where x has degree one and y has degree two -- if [3XG[0m can be expressed as the direct sum of other groups then the cohomology rings for those groups are found (by recursive calls to [2XModPCohomologyRingPresentation[0m ([14XHAPprime: ModPCohomologyRingPresentation (for group)[0m)) and combined using [2XTensorProduct[0m ([14X4.4-1[0m) [1X11.6-7 HAPPRIME_Polynomial2Algebra[0m [2X> HAPPRIME_Polynomial2Algebra( [0X[3XA[, ringA], poly[0X[2X ) __________________[0Xfunction [6XReturns:[0X Algebra element Converts a polynomial [3Xpoly[0m in [3XringA[0m (the mod-p cohomology ring presentation for the cohomology algebra [3XA[0m) into the equivalent element in [3XA[0m. If [3XringA[0m is not provided, it is recovered using the attribute [2XPresentationOfGradedStructureConstantAlgebra[0m ([14X3.3-1[0m). The [3XringA[0m argument is provided to allow cases where the ring and polynomial use different (but isomorphic) indeterminants from those provided by [2XPresentationOfGradedStructureConstantAlgebra[0m ([14X3.3-1[0m) (as is the case in [2XHAPPRIME_LHSSpectralSequence[0m ([14X11.6-5[0m)). [1X11.6-8 HAPPRIME_Algebra2Polynomial[0m [2X> HAPPRIME_Algebra2Polynomial( [0X[3XA, alg[0X[2X ) ____________________________[0Xfunction [6XReturns:[0X Polynomial Converts an algebra element [3Xalg[0m in the mod-p cohomology algebra [3XA[0m (given by [2XModPCohomologyRing[0m ([14XHAP: ModPCohomologyRing[0m)) into the equivalent polynomial from the corresponding ring presentation (given by [2XPresentationOfGradedStructureConstantAlgebra[0m ([14X3.3-1[0m)). [1X11.7 Derivations[0X The kernel of a derivation is efficiently computed by writing the ring R as an S-module where S is the subring of squares (or some larger ring in the kernel). The internal function in this section is used to describe this S-module and allow conversion between elements of R and elements of the S-module. [1X11.7-1 HAPPRIME_SModule[0m [2X> HAPPRIME_SModule( [0X[3XR, exps[0X[2X ) ______________________________________[0Xfunction [6XReturns:[0X Record For a polynomial ring [3XR[0m, k[x_1, x_2, ..., x_n], and list of exponents [3Xexps[0m, [e_1, e_2, ..., e_n], returns a record which represents [3XR[0m as an S-module, where S is the subring of [3XR[0m given by k[x_1^e_1, x_2^e_2, ..., x_n^e_n] The record has the following components: -- [10XRring[0m the original ring [3XR[0m -- [10XSring[0m a ring isomorphic to the subring of powers S -- [10XSpows[0m the exponents [3Xexps[0m -- [10Xgenerators[0m the list of generators of [3XR[0m as an S-module -- [10XFromPoly[0m a function which takes a polynomial in [3XR[0m and returns the corresponding element in the S-module (as a vector) -- [10XToPoly[0m a function which takes an element in the S-module (as a vector) and returns the corresponding polynomial in [3XR[0m [1X11.8 Test functions[0X Internal helper functions for testing [5XHAPprime[0m. [1X11.8-1 HAPPRIME_VersionWithSVN[0m [2X> HAPPRIME_VersionWithSVN( [0X[3X[0X[2X ) ______________________________________[0Xfunction [6XReturns:[0X String Returns a string giving a current version number for the HAPprime installation assuming that it is checked out from a subversion repository. This fetches the current version number from [10XPackageInfo.g[0m and appends the return of the [10Xsvnversion[0m program to this, returning the resulting composite string. [4X--------------------------- Example ----------------------------[0X [4Xgap> HAPPRIME_VersionWithSVN();[0X [4X"0.2.1.302:319M"[0X [4X------------------------------------------------------------------[0X [1X11.8-2 HAPPRIME_Random2Group[0m [2X> HAPPRIME_Random2Group( [0X[3X[orderG][0X[2X ) _______________________________[0Xoperation [2X> HAPPRIME_Random2GroupAndAction( [0X[3X[orderG][0X[2X ) ______________________[0Xoperation [6XReturns:[0X Group or [10XgroupAndAction[0m record Returns a random 2-group, or a [10XgroupAndAction[0m record (see [2XModuleGroupAndAction[0m ([14X5.4-5[0m)) with the canonical action. The order may be specified as an argument, or if not then a group is chosen randomly (from a uniform distribution) over all of the possible groups with order from 2 to 128. [4X--------------------------- Example ----------------------------[0X [4Xgap> HAPPRIME_Random2Group();[0X [4X<pc group of size 8 with 3 generators>[0X [4Xgap> HAPPRIME_Random2Group();[0X [4X<pc group of size 32 with 5 generators>[0X [4X------------------------------------------------------------------[0X [1X11.8-3 HAPPRIME_TestResolutionPrimePowerGroup[0m [2X> HAPPRIME_TestResolutionPrimePowerGroup( [0X[3X[ntests][0X[2X ) ______________[0Xoperation [6XReturns:[0X Boolean Returns [9Xtrue[0m if [2XResolutionPrimePowerGroupGF[0m ([14XHAPprime: ResolutionPrimePowerGroupGF (for group)[0m) and [2XResolutionPrimePowerGroupRadical[0m ([14XHAPprime: ResolutionPrimePowerGroupRadical (for group)[0m) appear to be working correctly, or [9Xfalse[0m otherwise. This repeatedly creates resolutions of length 6 for random 2-groups (up to order 128) using both of the [5XHAPprime[0m resolution algorithms, and compares them both with the original [5XHAP[0m [2XResolutionPrimePowerGroup[0m ([14XHAP: ResolutionPrimePowerGroup[0m) and checks that they are equal. The optional argument [3Xntests[0m specifies how many resolutions to try: the default is 25.