Skip to content
Author your academic documents in markdown
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
bin
csl
docx_templates
examples
pandoc-wrapfig @ 34cf73a
tex_templates
tex_utilities
.gitmodules
R01_template.md
README.md
mediabuilder.make
pandoc_functions.sh

README.md

mediabuilder

This repository helps you write your academic grant, paper, biosketch, or CV in markdown format. It provides templates, style files, and helper scripts that are useful for building media output as PDF from markdown files. This repository bring us closer to the goal of authoring scientific documents in markdown to completely separate content from style.

Examples of input/output

Here are some examples of the output that you can produce with mediabuilder:

Building documents with mediabuilder

  1. Install software prerequisites:

    • Install pandoc to convert markdown to PDF.
    • Install inkscape to convert SVG to PDF.
    • Install libreoffice (optional) for some recipes that read xls or docx files.
    • Install ghostscript (optional) if you need to merge PDFs (should be standard on linux).
  2. Clone and configure mediabuilder:

    • Clone nsheff/mediabuilder (this repository, cloned with --recursive to get the nsheff/pandoc-wrapfig submodule)
    • Configure mediabuilder. The examples use an environment variable $CODEBASE, to where you will store this repo:
    export CODEBASE=`pwd`/
    git clone git@github.com:nsheff/mediabuilder.git --recursive
    
  3. Assemble your BibTeX database (optional).

    If you want to produce a media type that includes citations, you will also need a bibtex file with your references. My favorite BibTeX management software is JabRef, because it is free, actively developed, and uses BibTeX as its native file format. The default makefile (mediabuilder.make) will use an ${BIBTEXDB} environment variable to look for your bibtex database. You can set it like this:

    export BIBTEXDB=path/to/db.bib
    
  4. Produce your content in markdown format.

    There are working examples of different media types in the examples folder. There you can find a grant, manuscript, and CV (pending). For each example there is a basic Makefile, which provides examples of recipes for building different media types. For example, this will render the example manuscript:

    cd examples/manuscript
    make manuscript
    

    This will render the .md manuscript source into a PDF

With these items, just follow some of the recipes below to use the mediabuilder assets to help build your output.

Docker containers

I've also produced docker containers for pandoc, inkscape, libreoffice that make this easier if you use docker:

docker pull nsheff/pandocker
docker pull nsheff/inkscape-docker
docker pull nsheff/libre

Description of mediabuilder repository

This repository contains:

  • tex_templates - a collection of tex templates for, e.g. NIH grants.
  • tex_utilities - some bonus tex software to do things like add page numbers to an existing PDF
  • bin - Small scripted utilities to build figures, extract bibliographies, suppress page numbers, merge PDFs, select versions of grant source files, etc. Documented in /bin/README.md.
  • csl - Contains some citation styles, which are derived from the citationstyles project repository, or from Zotero collection (https://github.com/citation-style-language/styles); it's the set of styles I use frequently, with some possible additions or adjustments for particular things I need. You can use any style from that repository or define your own
  • docx_templates - Word document templates for different grant agencies. To be used in YAML RMarkdown header as
output:
  word_document:
    reference_docx: styles.doc/NSF_grant_style.docx

Recipes

Converting an .xls file to .pdf with libreoffice:

Running libreoffice on the command-line like this will silently fail if you already have libreoffice running. If you use a containerized version, you can get around that issue. Use my libre docker container so you can run it while the real one is open.

budget:
	echo "Make sure libreoffice isn't already open"
	libreoffice --headless --invisible --convert-to pdf \
	--outdir output \
	src/budget_worksheet.xlsx

To set page printing limits in libreoffice calc:

  • go to View > Page Break
  • now select the area to print
  • choose: Format > Print Ranges > Define

Converting a .docx to .pdf with libreoffice:

pdf:
	soffice --convert-to pdf output/toc.docx \
	--outdir output

Merging PDFs with ghostscript:

merge:
	$(mbin)/mergepdf output/merged.pdf \
	output/title_page.pdf \
	output/research_proposal.pdf \
	output/assembly_plan.pdf

Adding page numbers

To add page numbers:

  1. put the PDF document into the tex_utilities/addpages.tex file.
  2. Run: pdflatex addpages.tex
  3. Page numbers are added at addpages.pdf!

Figures

You can refer to figures by label instead of by number, which makes reordering figures within documents easy. It also makes it possible to move figures from one document to another without renumbering.

![\label{abstract}Fig. \ref{abstract}: Example figure](fig/example_figure.png) 

Refer to figures with \ref{label}.

Separate citation lists (how to separate bibliography into its own file)

By default, pandoc will include your references cited just right at the end of the document. That works for some grants, but others want a separate reference document. To accomodate this, we need to do 2 things: 1) make a bibliography-only file; 2) suppress the bibliography in the main file.

  1. make a bibliography-only file

I wrote a script that does this: bin/getrefs

You can run getrefs on your markdown files, and pipe the results pandoc:

getrefs document.md | pandoc ...

However, this will also include any commented references, which is probably not what you want. So, there's a more complicated recipe in the mediabuilder.make Makefile called refs that will do this for you. It looks like this:

# Requires pandoc 2 with --strip-comments implemented
refs_nocomment:
	pandoc --strip-comments -t markdown `$(mbin)/ver src/specific_aims` \
	`$(mbin)/ver src/significance_innovation` \
	`$(mbin)/ver src/aim1` `$(mbin)/ver src/aim2` `$(mbin)/ver src/aim3` | \
	$(mbin)/getrefs | \
	pandoc -o output/references.pdf $(PANDOC_FLAGS)
  1. Suppress the bibliography in the main .md files.

One way to do this is to use a csl file that doesn't have a style for a bibliography. I don't like that, though, because it requires mucking around with style files, and you may want to produce a document with the bibliography some time. The better alternative is this: In the file itself, put in a yaml header:

---
suppress-bibliography: True	
---

Or, I also wrote a little helper script that will do this for you on-the-fly: bin/nobib. Use it like getrefs:

$(mbin)/nobib file.md | \
pandoc \
-o output/approach.pdf \

That will suppress the bibliography in the output. Done!

You can’t perform that action at this time.