Pandoc can be installed in anaconda by

conda install pandoc

or in brew. For brew, switch to brew environment first.

swt_brew # custom command in .bash_profile

Then go to the folder where you have .tex file to convert. Run the following command:

pandoc -s FeSH-JGR.tex -F pandoc-crossref  --citeproc -f latex -o FeSH-JGR.docx --bibliography=FeSH.bib

The command works only with brew version of pandoc for me. Change the file names appropriately.

Note that the order betweeen pandoc-crossref and citeproc should not change. Run multiples of time (unconfirmed).

pandoc knows how to deal with bibunits and \input{}. Note that line number referencing and citing do not work with pandoc.

If you have track changes, the command can be removed by:

python ../acceptchanges.py -c -n --infile=0-main.tex --outfile=0-main-no-track.tex

Note that acceptchanges.py works only with python v2.7. The acceptchanges.py can be found under TrackChanges folder or it comes with the latex package.

acceptchanges.py does not automatically deal with \input{} or \include{} files. So I have to process with acceptchanges.py all associated files individually.

For the command above to work, copy entire latex folder under /trackchanges-0.7.0/PythonPackage/ folder.

Under base environment, I installed arxiv_latex_cleaner (https://github.com/google-research/arxiv-latex-cleaner).

arxiv_latex_cleaner Hydrogen_FeS_paper/ --im_size 500

This will create organized package in a new folder, Hydrogen_FeS_paper_arXiv.

arxiv_latex_cleaner seems to know how to deal with \include{} and \input{}. It also deals well with bibunits.

pandoc -t latex -f docx in.docx -o out.tex

pandoc text-main.tex -f latex -t html -s -o text-main.html --mathjax --bibliograph=B30.bib --citeproc

For better outcome to html:

• convert all pdf figures to png figures
• \multicolumn is not compatible with pandoc