hfmark / tektonika_copyedit

Scripts for automated processing of .docx files into a LaTeX template for Tektonika

docx parsing for Tektonika

dependencies:

• python 3.n (preferably 3.8+)
• numpy
• biblib

A conda environment is a nice way to set this up.

You will also need to have pandoc (pandoc.org/) installed for the initial conversion of the .docx file.

general steps for the conversion process

• convert .docx article file to latex

  • pandoc file.docx -f docx -t latex --wrap=none -s -o file_pandoc.tex

• copy-paste bibliography from docx into anystyle.io (or run the anystyle gem locally if you're into that) and output as bibtex, save as a .bib file

  • [anystyle.io -> file_anystyle.bib]

• fix anystyle bibtex file year fields and keys, make a new .bib file

  • (set input filenames manually in the script)
  • fix_bibtex.py -> file_init.bib

• manually correct any non-ascii keys in bib file, if there are any

  • (these will be printed to stdout so we know they need to be fixed, usually for non-ascii characters)
  • (feels like there should be a way around this but I don't know it)

• parse the pandoc output tex file to a better tex format

  • (set the input filenames manually in the script)
  • parse_pandoc_file.py -> file_init.tex

• run bibtex and pdflatex, look at the output and figure out what needs fixing

  • pdflatex file_init.tex -> file_init.pdf
  • bibtex file_init.aux
  • pdflatex ''
  • pdflatex ''
  • (running at least twice gives inline references a chance to sort themselves out)

• manually link figure files at the right sizes, adjust placement of automated \includegraphics as needed

  • pandoc does not extract image files from word so they will need to be uploaded separately

• manually adjust for extra bits of inline citations (in red), in line citations for multiple papers by the same authors (hopefully in red), and year-only citations (in red)

• add extra hyphenation rules for words latex doesn't know if columns are overfull

• manually add authors, affilitations, short title, other header metadata with default placeholders

• look at junk file and manually reformat/place tables in text where they belong (because I do not understand longtable)

TODO:

• make sure catch for supplemental figures/tables works for in-text references
• figure out parsing for author metadata
• figure out longtable/table parsing?
• parse extra bits of citations, like 'e.g.,' wherever possible
• more user-friendly startup (ie input filenames, rather than editing scripts)
  • related: complete workflow that runs all scripts in sequence automatically
  • and maybe make this all install as a package?