Overview

The rbbt command line tool comes in the rbbt-util gem. To update the basic rbbt-util package and the rbbt command do

gem update rbbt-util

The rbbt command is basically a wrapper that finds a particular script and executes it. Typing rbbt will show the list of first level commands. Some of these are scripts, some of them are directories containing more scripts. You may append subcommands until you reach a script; which will execute it with the rest of the parameters. The rbbt command line also consumes a couple of command line parameters, such as the –log, which specifies the level of logging to use (0 for debug, 4 for default level, 7 or more no logs at all).

Rbbt does not enforce any particular practice on the command scripts, it just loads them. If a command is anything but trivial, it commonly uses Rbbt’s rbbt/util/simpleopt module. Among other things, this means most command-line parameters are assigned a shorthand version. This shorthand version is determined automatically for workflow tasks. You may find these in the command-line documentation with -h.

Some of the most important commands are the following:

• rbbt workflow
  • rbbt workflow list: See all installed workflows
  • rbbt workflow install: Install new workflows from github
  • rbbt workflow remote:
    • rbbt workflow remote {add|list|remove}: Add/list/remove remote workflows
  • rbbt workflow task: Examine, execute and monitor workflow tasks
  • rbbt workflow server: Start a http server with a REST interface that can be contacted through the browser or as a remote workflow
• rbbt tsv
  • rbbt tsv info: Display some general information about a tsv file (field names, number of entries, ect)
  • rbbt tsv change_id: Change columns between formats

Each command is free to implement its functionality however it wants. Unless the command is very simple, it usually displays help with the ‘-h’ parameter. If they are badly documented you can always examine the code using the –locate_file to find the path to the actual script file.

rbbt workflow task --locate_file

When commands get to long and are tedious to write, you can make use of aliases. These work as follows

rbbt alias <alias_name> <commands_and_parameters>

for example:

rbbt alias gene_name workflow task Translation translate -f "Associated Gene Name"

Which will transform:

rbbt gene_name -g ensembl.txt

into:

rbbt workflow task Translation translate -f "Associated Gene Name" -g ensembl.txt

Practical examples

The most common command you will probably be using will be rbbt workflow task. Typing rbbt workflow task <workflowname> will display all the tasks that this workflow makes available. Typing rbbt workflow task <workflowname> <taskname> -h will display a summary of the task parameters. Each parameter is associated with a task input.

Executed tasks are, by default, ran synchronously, with persistence, and the result file is printed to STDOUT. Alternatively, one can list and recover task files that where produced during the execution, re-run the job by using --clean or --recursive_clean to re-run also its dependencies, run the job in the background and monitor its status, etc.

Remote workflow can also be interfaced transparently through the rbbt command, but some functionalities may not be available.

A few interesting workflow tasks are:

• rbbt workflow task Translation translate: Translate a list of genes ids to another format (see a post about this)
• rbbt workflow task Genomics names: Takes a tsv file and translates all identified entities to a their human-friendly names