bash scripts

Some useful shell scripts I wrote over the years

If you want your shell to cleverly autocomplete its arguments to only those files which have the right extensions, link (or put a copy of) the content of the bash-completions folder in /usr/share/bash-completion/completions or /etc/bash_completion.d/ e.g. by calling

sudo ln --relative -s bash-completions/* /usr/share/bash-completion/completions/

or equivalently

stow --target=/usr/share/bash-completion/completions bash-completions

Rmarkdown, pandoc and Jekyll tools

• rmarkdown takes an .Rnw, .Rmd or .R file as its argument and builds/knits them using the appropriate R command/library (see this blog post

LaTeX and bibtex management

• bib2bibclean is a simple wrapper around bib2bib that removes any personal bibtex information and annotations such as Mendeley notes and tags, owner, timestamps, keywords, comments, summary or review fields. Handy for whenever you want to pass (some of your) richly annotated bibtex library on to other people.
• extract-references takes a LaTeX .aux file, extracts all references cited in the LaTeX document and puts them in a new references.bib file, stripping off annotations with bib2bibclean above. Practical when you only want to submit the relevant entries of a central large .bib file when making a submission in .tex/.bib format.
• mybib2html uses bib2bib to get all your own publications (filtered by author name) out of a .bib file and prints them out on two minimalist HTML files (a nicely decorated bibliography with links to the other file containing copy-paste-able bibtex entries) using bibtex2html. I use it to generate the publication list over at my website
• wclatex is a python script that gives the wordcount of a LaTeX document according to the regulations of the standard University of Edinburgh MSc Dissertation format (footnotes, tables, captions etc excluded from wordcount). This code formed the basis for my LaTeX word count plugin for SublimeText

PDF processing

• cutpdf extracts the specified page range from a pdf using ghostscript
• mergepdf merge pages together (uses qpdf)
• deprecated (because based on pdftk): makebooklet tiles the pages of an input pdf so that it can be folded into a small booklet when printed duplex. Requires pdftk and pdfjam.

Audio processing scripts

• rip rips audio tracks off a CD and converts them to flac format, including metadata retrieval. It requires tools from the debian packages cdparanoia, libcddb-get-perl and flac.
• tomp3 converts a flac file into an mp3 while preserving all tags. Requires flac and lame.
• dirtomp3 uses tomp3 to convert a directory of flac files, putting the output in a parallel directory.