Programming with GAP: Discussion

Ten hints for GAP beginners

1. Remember that GAP is case-sensitive! That means that ABC, Abc and abc are three different identifiers. A call to SymmetricGroup(3) works, but Symmetricgroup(3) will cause an error.

2. An error message “Error, Variable: 'FuncName' must have a value” in a function call usually points to a typo in the function name (see the previous hint), or to some package that must be loaded in advance using LoadPackage.

3. Do not hesitate to use longer and more informative variable names where appropriate. For example, x looks perfectly suitable for List([1..10], x -> x^2), while ConClassesReps may be more informative than just x for a list of representatives of conjugacy classes of a group.

4. Use command line editing: scroll the history of commands and navigate within the command line using arrow keys to edit it.

5. Use autocompletion instead of typing names of functions and variables in full. Type the initial part of the identifier and then press <Tab>. It will be completed if its unique completion is possible. If not, you may press <Tab> again to see all possible suggestions.

6. To view help pages, use ? and ?? commands. This will search not only in the GAP manuals, but also in the manuals of all GAP packages available in your GAP installation.

7. Set default help format to HTML. Use SetHelpViewer to view it with your preferred browser.

8. Use **LogTo** to save all GAP input and output into a text file. It should be called before calculations, not after!

9. If the calculation takes too long, press <Control>-C to interrupt it. Then type quit; to leave the break loop.

10. Read A First Session with GAP from the GAP Tutorial.

Writing programs in GAP

• Use direct calculations in the command line for object exploration and prototyping, then think how to organise your code to make it reusable.

• Use LogTo to save inputs and outputs to a logfile which then you can edit in a text editor.

• Save code in text files and use Read to load them. Find some advanced text editor to edit these files.

• Make your code modular and reusable by organising it into functions.

• Write comments in the code – this will help you when you will return to it after a while.

• Understand break loops: you may explore variables on the current break level and use Where to show the last commands before the error occurred.

• Use preferences mechanism to customise GAP, for example, to set help viewer to HTML or to make command line history available after quitting GAP in the next GAP session.

• Understand the theory behind calculations: theoretical improvements could improve the performance much more than highly optimised code which still does a brute-force calculation.

• Implementing algorithms, do not forget corner cases. For example, does the implementation work on a trivial group or an identity element?

• Do not rely that GAP functions return results in a particular order, unless this is documented. For example, dependently on the method, it may be not guaranteed that conjugacy classes or irreducible characters are listed in some particular order, or that the first element in their list is the conjugacy class of an identity element or the trivial character.

• Don’t ask for more than you need as this may have performance implications. For example:

  • If the property is an invariant of a conjugacy class (of elements or of subgroups), you may be interested only at looking at the representatives of conjugacy classes.

  • If you are interested in a list of elements of a collection, without a particular order, use AsList instead of AsSSortedList.

  • Do not calculate conjugacy classes of all subgroups, if e.g. you are interested only in normal or maximal subgroups - there are special methods to compute them.

  • If you are looking for p-subgroups, first you may calculate a Sylow p-subgroup of a group, and then look at its subgroups and their conjugates.

  • Representation matters: it may be worth to convert a group from fp group to an isomorphic pc group or a permutation group to use faster methods.

• See GAP Frequently Asked Questions for further hints.

Stay in touch

• Subscribe to the GAP Forum.

• If you are on Twitter, follow gap_system.

• If you need help, choose among these three options, dependently on the question:

  • ask questions in the GAP Forum

  • send them to the GAP Support

  • post them at the Mathematics Q&A site

Contributing to GAP

• If you think that you’ve found a bug: please create an issue on GitHub or report it by email to GAP Support.

• Please cite GAP if you use it. This helps for the community to grow, and this will help you in return. This page suggests how to cite GAP, and the function Cite will help to generate citation sample for precisely the same version of GAP that is used.

• Consider sharing your GAP developments with others, from sharing your code by available means to organising it into a GAP package, submitted for the redistribution with GAP and optionally for the refereeing.

• Contribute to the further development of this lesson itself.

Tips and tricks

• This is a simples approach to call GAP from a shell script. Create the shell script called check-one-order.sh with the following content:

#!/bin/sh

gap -r -b -q avgord.g << EOI
TestOneOrderEasy( $1 );
quit;
EOI

and make it executable using chmod u+x check-one-order.sh. Now you may call it as follows:

$ ./check-one-order.sh 24

fail

$ ./check-one-order.sh 105

[ 105, 1 ]

• Reading data files

GAP can read any valid GAP input from the code using Read. The contents will be read and evaluated in the main read-evaluate-print loop, but the results will not printed. Sometimes you may want to read the content of the file as a function and return that function - for that purpose you may find ReadAsFunction useful. But what to do if you have some a data file coming from other source, and it is not a valid GAP input? Sometimes you may have control over the tool that exports data, and may be able to tweak it to generate GAP input file. But where to look if this option is not possible?

ReadCSV( filename[, nohead][, separator] ) reads a file in a CSV (comma separated values) format and returns its entries as a list of records (see here). The entries of the first line of the file will be used for the names of the record components (blanks will be translated into underscores). One could also indicate that the first line contains data instead of field names, and also specify a custom separator. Conversely, PrintCSV may be used to output CSV files.

To read arbitrary (binary or text) files as strings, use the StringFile function provided by the GAPDoc package (see here). It will return the content of the file as a string. After that, you may use various string manipulation tools (see Strings and Characters in the GAP reference manual) to process it in the way you need. GAPDoc package also provides the FileString function which writes the content of a string into a file.

If you need to organise reading/writing line by line, instead of reading or writing the whole file/string at once, we suggest to look at the functionality provided by the IO package (see here), in particular at IO_ReadLine and IO_WriteLine.