When requesting help on a specific pipeline, only the arguments for that pipeline are displayed. There is a set of arguments that can be used regardless of the pipeline being used and those are all listed at the end of this page. The majority of these arguments will not be used by most users, so the most important arguments are identified and given a brief description here.
The arguments --all-pipelines (-api)
, --categories (-cat)
, --gkno-arguments (-ga)
and --help (-h)
are all used for displaying help information. These arguments are discussed in the help
Disable automatic execution
is executed, (one) or more makefiles are generated (see the overview
for a more detailed description of how gkno
works). By default, if only a single makefile is generated, gkno
will immediately begin execution of the pipeline. There are number of occasions in which this is not desired behaviour and so the --do-not-execute (-dne)
argument can be used to disable execution. If execution was disabled in order to inspect the makefile, it can then be executed using the following command:
- make -k -j N -f <MAKEFILE>
In the above, -k will cause the pipeline to keep going for as long as possible after an error. If the makefile is performing many tasks, even if one task fails, other unrelated tasks can still be successfully run prior to terminating execution. The -j N argument specifies the maximum number of parallel jobs that can be run. Finally, the -f defines the name of the makefile to be executed.
Parameter sets are used to minimise the number of parameters that need to be included on the pipeline by collecting commonly used parameters into named parameter sets. These are discussed in more detail here
. The command --export-parameter-set (-ep)
allows a parameter set to be created, --parameter-set (-ps)
invokes a parameter set for a pipeline, and --remove-parameter-set (-rp)
removes a previously defined parameter set (only user-defined parameter sets can be removed).
Generate multiple makefiles (cluster usage)
In normal usage, on execution gkno
will generate a single makefile and begin execution. The execution can be parallelized (see below), but there are times where it is preferable to generate multiple makefiles and distribute execution of these makefiles across different resources, e.g. a cluster. See the dedicated tutorial
for more information on performing this kind of execution, but this begins by setting the --multiple-makefiles (-m)
Depending on the pipeline and arguments supplied to it, the pipeline can consist of multiple tasks that are performed independently of each other. If this is the case, there is no reason to execute all the tasks sequentially. Instead, we can use the --number-jobs (-nj) to specify the number of tasks that can be executed in parallel.
List of all gkno arguments
- Print all available pipelines to screen.
- Display all available help categories.
- Define the path to the directory in which the configuration file(s) resides.
- Display debugging information.
- Flag to stop execution of the Makefile after gkno is run.
- Do not log this run of gkno.
- Flag to disable reordering of values supplied to linked arguments.
- Name of the file to which a full visual depiction of the pipeline graph will be output.
- A full and reduced version of the pipeline graph will be output to files beginning with this value and appended with '-full' and '-reduced'.
- Name of the file to which a reduced visual depiction of the pipeline graph will be output. This is the form of the pipeline shown on the webpage for the pipeline.
- Export a parameter set to a configuration file of this name.
- Print help information for gkno arguments.
- Print requested help message.
- Path for input files if not defined. Default: current directory.
- Append this text to all makefiles generated in this execution of gkno.
- If any of the long form arguments are supplied with a list of valuess causing one or more tools to require multiple iterations, if this is set to true, multiple makefiles are generated. If set to false [default], a single makefile will be generated.
- The number of parallel jobs to be used (only valid in conjunction with the --internal-loop option).
- Path for output files if not defined. Default: current directory.
- Generate a stdout and stderr for each task.
- Remove a user-generated parameter set.
- Write out json files for use with the web page.