This repository contains the source code for an autochecklist, a bash script to edit LaTeX documents according to the group writing checklist. This project is currenly in its early stages, and may see revision to its usage and form.
This bash script takes input of a LaTeX source directory, and processes it according to the ARFC writing checklist found at https://arfc.github.io/manual/guides/writing/checklist/. The original directory isn't changed, but two copies are created: an edit and a diff. The edit has a series of changes applied to it, based on the checklist items which can be automatically executed. The diff breaks down, using the program latexdiff, what these changes were. The diff also highlights, using the latex package xcolor, parts of the document which the checklist says should be changed, but either can't be automatically changed or can't be confidently identified.
To use the script, clone the repository or download the files w_check.sh, elements.csv, and irregular_passive_verbs.txt into a directory.
Then, make a subdirectory at the same level named input, and put the LaTeX source code you want to inspect into that directory. If you made a clone of the entire repository, there should already be an input directory, with an example document inside it. Remove the example if you want, and add your files to the input directory.
Note: To view the highlights, ensure the package xcolor is used in the document, such as by declaring the following at the start of the main tex file:
\usepackage{xcolor}
When the target files are in the input directory, navigate to the main directory and run the script:
bash w_check.sh
The script will generate two copies of the LaTeX source, in the subdirectories output/edit/ and output/diff. The edit is a copy of the source with all possible changes applied to it, to be used or previewed. The diff is a copy of the source which uses latexdiff to show each change that was made. The diff also shows highlighted suggestions for more modifications, which couldn't automatically be performed.
In a LaTeX repository, some utility files may not be intended for show,like an acros.tex file for the acro package. The script can be told to ignore such files by including a .checkignore file with the names of files to ignore. For example, in the above situation, the file might look like the following:
#Files to ignore go below. Lines starting with # are treated as comments.
acros.tex
The script would then avoid editing or highlighting any file named acros.tex in the repository.
The file .checkignore should be placed in the first level of the input directory. E.g., if using the default input of ./input/, the script will check ./input/.checkignore for files to ignore. If a directory my_thesis were given (bash w_check.sh 'my_thesis/'), it would check for the respective ./my_thesis/.checkignore
The repository contains the script and a basic input example.
input is a directory where files are put to analyze. In the main repository, it currently contains an example of an input.
w_check.sh is the core file of the script, a bash script which makes copies of the contents of input into a new directory output.
elements.csv and irregular_passive_verbs.txt are text files with data about elements and a list of exceptions to passive verb patterns. In the future, these may be consolidated into w_check.sh and no longer necessary.
The script's operations are organized to match the checklist itself, for best usability and maintenance. To accomplish this, a couple of utility methods are defined:
The function "insed" takes a sed command and performs it on each tex file in the directory. Example:
insed "s/I'm/I am/gI"
insed -C "s/can't/can not/g"
These lines expand the contractions "I'm" and "can't", respectively. Note that the both give the full sed command as the argument, but don't specify a file or whether to write the original file or print to terminal. On each of these calls, the script would effectively run sed -i -e <sed command> <file> on each file ending in '.tex'. Flags like -C modify how this is performed, with the flag -C in this case saying to modify the command to account for capitalization. (Some of these flags, like -C, are redundant because they could be made in the command, but the flag makes it quicker to write and more readable afterwards)
The function "hladd" adds a pattern to a list of things to highlight, and after finishing edits and generating diff files, searches for and highlights these patterns in the diff files. For example, hladd "red" "\\b[Vv]ery\\b" would add a pattern searching for the word 'very' to the list, with a note to highlight it using the color named "red," changing very to \colorbox{red}{very}
Thus, the script takes the form:
- Basic checks
- Define utility functions
- Edit or prepare highlights to the documents for each checklist item
- Generate diff files between the original and edit using latexdiff
- Apply the prepared list of highlights to the diff files
Author: Owen Strong
Inspired by Matt Might's shell scripts