Skip to content

Latest commit

 

History

History
86 lines (51 loc) · 5.28 KB

HOWTO.markdown

File metadata and controls

86 lines (51 loc) · 5.28 KB
Raw

Most important things

  • pittetd is a submodule. When cloning the repo, do git clone --recurse-submodules https://github.com/berquist/dissertation.git (see here for help).
  • Tex.SE is your friend!

Compilation

  • Don't run a bunch of pdflatex/bibtex commands or use some crazy Makefile. latexmk -pdf etd.tex does everything. By crazy Makefile I mean that thing Adam uses for everything. Don't use it. See how my Makefile cleans some other files, but otherwise just calls latexmk.

  • For latexmk, if you want to use xetex/xelatex, pass the -xelatex flag.

  • If you're using minted to display source code, you must have the pymgmentize Python package installed and pass -shell-escape to latexmk.

  • How to properly 'make' a latex project?: https://tex.stackexchange.com/q/40738/94717

General

Answer to both: yes.

For structure:

I and others use \input to avoid any page issues, and apparently it's much faster. For separate compilation, I used the standalone (rather than subfiles) package.

One of the most common problems is the dreaded overfull \hbox. This usually happens because TeX can't figure out how to optimally flow everything inside of whatever display box it's working on, and usually results in something sticking out of the right-hand side of the text. One common cause of this for chemists is that TeX can't figure out how to hyphenate (and therefore line break) on non-standard (chemistry) words. Rather than add discretionary hyphens with \- everywhere, it's better to set hyphenation rules.

Labels. For figures and tables that have captions, the label must appear inside or after the caption. For equations inside of an align (yes, you can have multiple labels inside align), the label must be on the same row as the equation, so pay attention to the \\.

Bibliography info

pittetd works just fine with biblatex/biber. See https://tex.stackexchange.com/q/25701/94717 for the differences between all the available options.

Dealing with messy libraries:

I mostly used JabRef, sed, and a text editor to do stuff by hand.

  • JabRef will automatically convert bibtex -> biblatex, format many journal abbreviations, help with name formatting, ...

  • sed to delete abstract, keywords, mendeley-tags; sed -i '/mendeley-tags/d' library.bib; this may cause compilation failure if fields are not on a single line but JabRef and trying to compile the document can point you to the line number that has junk, clean it by hand.

  • Special integration of biber with your favorite editor: https://tex.stackexchange.com/q/154751/94717

siunitx

Used for proper typesetting of units and numbers, either combined or separately.

This is specific to the pittetd class: you cannot put \si, \SI, or any other siunitx command in captions, because there will be complaints about undefined commands when making the list of figures/tables. The solution is to use a short caption (in []) that doesn't contain siunitx commands. You will probably need to do this anyway so the captions don't go over a line on the lof/lot pages.

How to do...?

Including source code

The best package to use is probably minted. Here is a comparison of the most common options. listings is useful for plain text and looks good, but minted is probably what you're thinking of in terms of text editor syntax highlighting.

Including Jupyter Notebooks

The summary is that rather than doing

jupyter nbconvert --to latex foo.ipynb

you should do

jupyter nbconvert --to markdown foo.md
pandoc --listings -f markdown -t latex foo.md -o foo.tex

Metadata

  • See here for how metadata can be generated using the Creative Commons license chooser.
  • See the hyperxmp package for how hyperref can be used to generate metadata instead.