bash-utils

Some of the scripts here are to aid data inspection in the command line, avoiding the need to fire up R or Python. Also useful are pick (for selecting/combining/transforming/filtering tabular data) and transpose from github.com/micans/reaper for fast and low-memory transposition of (large) tabular data.

Highly useful command line data wrangling: GNU datamash.

Command line directory bookmarks

Apparix used to live here but has its own repository now.

Unix file/stream column and row manipulation using column names

pick used to live here but has its own repository . It is a concise command-line query/programming tool to manipulate streamed data columns and rows. It can be thought of as (unix) cut on steroids, augmented with aspects of R, awk, and SQL.

Unix terminal histograms and bar charts

hissyfit can be used at the end of Unix pipes (or read from file) to draw terminal histograms and bar charts for quickly gauging numerical data and count data. Hissyfit documentation and examples.

Miscellaneous code

• merge-files-col.sh This merges columns of files using transpose from the reaper distribution. It is quite a bit faster and much more memory efficient than a straightforward Python or perl implementation.

• peach.c PArEnthesis CHecker. It's not smart about anything and will complain about things like /* my little list 1) foo 2) bar */ and "->)(<-". Still I've found it helpful over the years. It checks {}, () and []. This is the first C program I wrote, so please be gentle.

• wordmer.pl generate all words of length k over some alphabet.

• tallyho.sh tally the first N sequences of a fastq file, for example to look at index reads. This uses tally from the reaper distribution.

• bubba Bsub/wrapper LSF submission script to take some pain away. It prints the constructed bsub command and submits it. Several options including dry run (-T).

• pp Prettyprint - tiny wrapper around csvtk pretty

• nsort Name sort - tiny wrapper around csvtk sort

• ct Tiny wrapper around unix column (align columns)

• ctpartition Group ct-ed data by first column, split groups by separating line; assume header line

• ctp This is ct followed by ctpartition; pretty print tables with headers grouped by first colum

• sr Tiny wrapper around screen, screen -R without argument, otherwise screen -r <arg>

• dmtp Tiny wrapper expanding to datamash transpose

• ph, preserve_header preserve header, process rest of file, e.g. ph sort -r < data.txt

bash-workutils

Space/time bash functions in .bash-workutils (see further below). Most of these will lead to a lot of disk access, use with care. A bunch of other miscellaneous functions have been added. Two worth picking out are

• funcfile NAME find in which file function NAME is defined.
• ls_func FILE list all functions defined in FILE.

The list of functions in .bash-workutils:

achoo bj bjl colcount colnames ffn funcfile gimme_sum grab groupify
howoldami lines ls_bigold ls_count_files ls_file_spread ls_func
ls_lastfile ls_ls ls_misc ls_mouldy ls_size_any ls_size_suffix myman
nchar procli public silent tailafter ungroupify

Space time functions in more detail:

--- ls_bigold
  List directories up to a certain depth, ordered by disk usage,
  with the number of days since last modified.
  Argument: directory depth.
  Example:
  ls_bigold 2
  NOTE: in a project/team root directory this may take some time and
  tax the file system. Perhaps best to save the output in a file.
  CAVEAT subdirectories of a directory may have changed. Use as guide!
  USEFUL order the output by the third column to group directories together,
    e.g. ls_bigold 2 > out.bigold; sort -k 3 out.bigold

--- ls_mouldy
  Find directories left untouched for longer than first argument (in days)
  up to a depth of second argument.
  Example:
  ls_mouldy 183 3
  CAVEAT subdirectories of a directory may have changed. Use as guide!

--- ls_size_any
  List all regular files recursively and sort by human-readable size.
  First optional argument:   lower bound e.g. 10M or 16k, or 0k
  Second optional argument:  upper bound e.g. 4k (useful for small files)
  Example:
  ls_size_any 10M     # find files larger than 10M
  ls_size_any 0k 4k   # find small files

--- ls_size_suffix
  Find files ending with suffix recursively, sort by human-readable size.
  First argument: suffix, e.g. .cram or .fastq.gz
  Second (optional) argument: a lower bound for size, e.g. 10M or 64k.
  Example:
  ls_size_suffix .fastq.gz
  ls_size_suffix .cram 500M
  ls_size_suffix .cram 1G

--- ls_file_spread
  For each directory count the number of files in it, recursively.
  The output is sorted by count, with a total tally added.
  Useful to check if applications are well-behaved and do not
  crush the file system with large numbers of files in a single directory.
  Modified from code by Glenn Jackmann on stackoverflow.

bash-myutils

More small functions that I use in .bash-myutils. These are slightly more random than the ones in .bash-workutils.

bash_max bash_min countcram cpuhours debug_bash decolon ffncp ffnl fqfa
helpme nflogcmd set_farm_mem themax themax2 themin themin2 theminmax