Inspecting manuscripts in LaTeX#

This vignette walks through some utilities to inspect manuscripts compiled from LaTeX (lsys/texCheckmate). Examples of inspection include retrieving text with hardcoded numbers (so that they can checked by hand), finding unreferenced labels, or finding dead links. Much of this is adapted directly from Jonathan Dingel’s project template (jdingel/projecttemplate) which includes a Makefile to inspect (“review”) a manuscript in LaTeX.

Repeated strings#

To find repeated strings (one of the most common example being “… the the …”), use the extended grep

$ egrep "\b([a-zA-Z]+) \1\b" paper.tex -n

# Example output
243:the the

which grabs repeated words and report the line number (-n). This can be adapted to more than just single words (thanks ChatGPT):

$ egrep "\b(\S+) (\S+) \1 \2\b" paper.tex -n

# Example output
246:flag this flag this

A simple shell script for this would look like this:

# inspecting/repeated-strings.sh
#!/bin/bash
SRC_TEX="paper.tex"

echo "Repeated unigrams:"
egrep "\b([a-zA-Z]+) \1\b" "$SRC_TEX" -n

echo "Repeated bigrams:"
egrep "\b(\S+) (\S+) \1 \2\b" "$SRC_TEX" -n

echo "Repeated trigrams:"
egrep "\b(\S+) (\S+) (\S+) \1 \2 \3\b" "$SRC_TEX" -n

and the Make recipe would be:

# Makefile
REPEATED_STRINGS_SRC = ./inspecting/repeated-strings.sh
.PHONY: repeated_strings
repeated_strings: ## Check for repeated words
    @echo "==> $@"
    @echo "Check for repeated words (e.g. 'the the table shows...')"
    -$(REPEATED_STRINGS_SRC)

Regex nuance

[a-zA-Z] and \S are not the same. \S includes digits. So a repeated “Fig 3 Fig 3” is flagged with \S but not with [a-zA-Z].

Duplicated labels#

To check for duplicated labels (TeX does warn for this too), e.g., having used the same \label{tab:summary} more than once in all the *.tex files, grep all such labels

$ grep -o '\\label{[^}]*}' paper.tex | uniq -cd

Unpacking the above:

  • grep -o '\\label{[^}]*}' paper.tex finds all “\label{**}” in the source.

  • sort sorts the labels so that repeats can be detected.

  • uniq (assumes results have been sorted, hence sort) then finds duplicates.

  • c prefix lines by number of occurences.

  • d reports only duplicated lines, one for each group.

A Make recipe can then be defined as:

# Makefile
.PHONY: duplicated_labels
duplicated_labels: ## Check for duplicated labels
    @echo "==> $@"
    @echo "Check for duplicated labels"
    grep -o '\\label{[^}]*}' paper.tex | sort | uniq -cd

Unreferenced labels#

To find unreferenced labels in the tex files, use the diff to compare the labels defined from \label{**} and labels used in all the \*ref*{**}

A simple shell script would look like:

# unreferenced-labels.sh
#!/bin/bash
SRC_TEX="paper.tex"

diff \
  --side-by-side \
  --suppress-common-lines \
  <(grep -o --no-filename 'ref{[A-Za-z0-9:_]*}' "$SRC_TEX" | sed 's/ref//' | sort | uniq) \
  <(grep -o --no-filename 'label{[A-Za-z0-9:_]*}' "$SRC_TEX" | sed 's/label//' | sort | uniq) 

# Example output
	> {fig:un}
	> {sec:intro}
	> {sm:results}

A Make recipe would look like:

# Makefile
.PHONY: unreferenced_labels
unreferenced_labels: ## Check for label referencing
    @echo "==> $@"
    @echo "Check for unreferenced labels"
    ./inspecting/unreferenced-labels.sh

Hardcoded numbers#

One of the things to check with a manuscript is whether hardcoded numbers are correct (e.g., “…sample size of 42…”). A quick way to grab all lines of a tex file that contains some hardcoded number is using grep (again).

For example, a simple shell script would be:

# hardcoded-numbers.sh
#!/bin/bash
do
cat paper.tex | 
  sed -n '/\\begin{document}/,$p' |  # Skip lines until \begin{document}
  sed '/\\begin{table}/,/\\end{table}/d' | # Remove lines within table environments
  sed 's/\\cite{[A-Za-z0-9:,\-]*}//g' | #Drop citations with \cite{} that may contain numbers
  grep '[0-9]'
done

which excludes lines before \begin{document}, lines within table environments, and lines where numbers come only from the \cite{} command. The last line grep '[0-9]' then extracts lines with numbers and prints them.

# Example output
Hello world! The tex also, contains error of \emph{gramar and and spelling}. or use this text too see an few of of the
problems that LanguageTool can detecd. What do you thinks of grammar checkers? Please not that they are not perfect. Style 
ssues get a blue marker: It's 5 P.M. in the afternoon. LanguageTool 3.8 was released on Thursday, 27 June 2017.
Here,I forget to put a space after a comma. I refer to \ref{fig:thisfig1} in the text, but I do not refer to the second one. \\
const seedUrl = "http://127.0.0.1:8080/";
\email{asdfgh.zxcvbnm@123.12345678.ca}

using the example.tex from textidote. Amongst other numbers, I can then, for example, check that the “5 P.M.” is correct or that the LanguageTool version is correct, and if not, make the necessary amendments.

The hardcoded-numbers.sh (fully adapted to omit lines from certain environments at the end) can be called with a Make recipe:

# Makefile
.PHONY: hardcodednumbers
hardcodednumbers: ## Find hardcoded numbers
    @echo "==> $@"
    ./inspecting/hardcoded-numbers.sh

Acronyms#

A shell one-liner to find all the acronyms in the file with a tally of count usage is

$ cat *.tex | grep -wo "[A-Z]\+\{2,10\}" | sort | uniq -c | sort -gr

  • grep -wo "[A-Z]\+\{2,10\}" matches upper case words between 2 to 10 characters long

  • sort | uniq -c sorts the results (like in Duplicated labels) before creating a counter (-c)

  • sort -gr then sorts the results numerically (-g) in descending order (-r)

# Example output
    167 PFAS
    123 PFBS
     59 GUSTO
     30 SD
     ...

This can be adapted to exclude certain parts of a tex file (like Harcoded numbers)

# acronyms.sh
#!/bin/bash
do
cat paper.tex | 
  sed -n '/\\begin{document}/,$p' |  # Skip lines until \begin{document}
  sed '/\\begin{table}/,/\\end{table}/d' | # Remove lines within table environments
  sed 's/\\cite{[A-Za-z0-9:,\-]*}//g' | #Drop citations with \cite{} that may contain numbers
  grep -wo "[A-Z]\+\{2,10\}" | sort | uniq -c | sort -gr
done

which can then be called from the Makefile:

# Makefile
.PHONY: acronyms
acronyms: ## Find and tally acronyms
    @echo "==> $@"
    ./inspecting/acronyms.sh

Resource:

Check URLs#

Another thing to check is that the URLs referenced in the manuscripts are live and working. This can be done by combining the use of grep to find all URLs specified in the \url{} and \href{}{} commands before using wget to check that the link works:

# urls.sh
#!/bin/bash
HREF_LIST=$(grep -o 'href{[A-Za-z0-9:/\._?#=]*}' paper.tex | sed 's/.*href{//' | sed 's/}//;s/\\#/#/')
for URL in $HREF_LIST; 
do
  wget --spider --no-verbose "$URL"
done

to get output that looks like:

2023-09-29 14:42:49 URL: https://en.wikipedia.org/wiki/ 200 OK
2023-09-29 14:42:49 URL: https://developers.google.com/maps/documentation/javascript 200 OK
wget: unable to resolve host address ‘www.this-should-not-work.com’
...
...

fake.com

I tried using www.fake.com as an example of a failure but apparently that’s a real website..

The Make recipe would then be:

# Makefile
.PHONY: checkURLs
checkURLs:
    @echo "==> $@"
    ./inspecting/urls.sh

Wordcount#

To count words, TeXcount comes in handy. The basic TeXcount command-line syntax is:

$ texcount paper.tex [opts]

# Example output
File: paper.tex
Encoding: utf8
Words in text: 4382
Words in headers: 47
Words outside text (captions, etc.): 266
Number of headers: 20
Number of floats/tables/figures: 6
Number of math inlines: 32
Number of math displayed: 3
Subcounts:
  text+headers+captions (#headers/#floats/#inlines/#displayed)
  1+0+3 (1/0/0/0) _top_
  442+1+0 (1/0/0/0) Section: Introduction}\label{sec:intro
  ...
  ...

A simple .sh file can be made to log the wordcount. The -nobib option exclude bibliography and -quiet suppresses error messages.

# wordcount.sh
#!/bin/bash
DETAILED_OUTPUT_FILE="inspecting/wordcount-detailed.txt"
STORE=$(texcount \
  -nobib \
  -quiet \
  paper.tex
)
echo "$STORE" > "$DETAILED_OUTPUT_FILE"

which can then be called from a Makefile:

# Makefile
.PHONY: wordcount
wordcount: ## Wordcount via texcount
    @echo "==> $@"
    @echo "Check word count using texcount"
    ./inspecting/wordcount.sh

Ignore text count from within tex

Within the tex file(s), use %TC:ignore to start ignoring text and end it with %TC:endignore.

Commented text are ignored

Apparently TeXcount intelligently ignores text (including notes to self) that are commented out in the *.tex files.

Resources:

Textidote: Grammar and LaTeX checker#

A helpful tool to check for both LaTeX and language (using LanguageTool) is textidote. textidote can take a pre-specified dictionary of ignores (using --dict path_to_dict) and output a HTML file with detected errors.

$ textidote --check en paper.tex

https://github.com/sylvainhalle/textidote (Source: PatrBal/vim-textidote)

Installation

I had issues installing textidote. But turns out it’s available via brew (although undocumented).

The Make recipe would look something like:

# Makefile
.PHONY: textidote
textidote: ./inspecting/textidote_dict.txt
    @echo "==> $@"
    @echo "Check doc with textidote"
    textidote --check en --dict $< --output ../paper.tex > inspecting/textidote.html

which outputs the results like in the image above to inspecting/textidote.html. It looks like the distinct colors indicate types of error but the colors cannot be configured.

https://sylvainhalle.github.io/textidote/

(Source: https://sylvainhalle.github.io/textidote/)

Color meanings:

  • Red: Spelling errors

  • Orange: Grammar errors

  • Yellow: TeX syntax style suggestions

Resources

Bonus: “Due to”#

All it takes to find bureaucratese like “due to” (Elements of Style by Strunk and White) is:

$ grep -n 'due to' paper.tex

which then spits out lines containing “due to” that can then be nuked.

The corresponding Make recipe is:

# Makefile
.PHONY: dueto
dueto: ## Find "due to"s; Did you mean "because of", "owing to", or "from"?
    @echo "==> $@"
    @echo "Find all the 'due to's in writing"
    grep -n 'due to' paper.tex

Resources

  • Strunk, William, and E. B. White. The Elements of Style. 4th ed., Pearson, 2000.

All together#

All together, the directory containing the tex files, the Makefile, and the shell scripts would look something like

.
├── inspecting
│  ├── logs
│  ├── acronyms.sh
│  ├── ay2numeric.sh
│  ├── hardcoded-numbers.sh
│  ├── linkchecker.sh
│  ├── numeric2ay.sh
│  ├── repeated-strings.sh
│  ├── textidote_dict.txt
│  ├── unreferenced-labels.sh
│  ├── wordcount.sh
├── Makefile
├── paper.tex
├── paper.pdf
└── references.bib

./inspecting/logs contains the log files for reference. ay2numeric.sh and numeric2ay.sh swaps citations from author-year to numeric, and vice versa.

Click to expand and see the full .sh scripts (lsys/texCheckmate) that logs the output:

acronyms.sh
#!/bin/bash

SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"
LOG_FILE="$LOG_DIR/acronyms.log"

# Ensure the log directory exists
mkdir -p "$LOG_DIR"

# Overwrite the log file at the start
> "$LOG_FILE"

# Function to process files, renamed to "acronyms"
acronyms() {
  local file="$1"
  
  # Check if file exists
  if [[ ! -f "$file" ]]; then
    echo "Error: $file not found" | tee -a "$LOG_FILE"
    return 1
  fi

  echo "Checking file: $file" | tee -a "$LOG_FILE"
  
  # Output file marker
  echo '===' "$file" '===' | tee -a "$LOG_FILE"
  
  cat "$file" | 
  # Process the file content and log results
  # =======================================================================
  # Ignore lines relating to environment
  sed -n '/\\begin{document}/,$p' |  # Skip lines until \begin{document}
  grep -Ev '^\\newcommand' | # Ignore lines starting with \newcommand (e.g., \newcommand{\nc}{} )
  grep -Ev '^\\setcounter' | # Ignore lines starting with \setcounter
  grep -Ev '^\\vspace' | 
  grep -Ev '^\\hspace' | 
  grep -Ev '^\\usepackage' | 
  grep -Ev '^\\setlength' | 
  # =======================================================================
  # Ignore math, tables, figures
  sed '/begin{equation}/,/end{equation}/d' | sed '/begin{equation\*}/,/end{equation\*}/d' | sed '/begin{align}/,/end{align}/d' | sed '/begin{align\*}/,/end{align\*}/d' | # Remove equation environments
  sed 's/\\input{[A-Za-z0-9_\/\.]*}//g' | # Drop input files that might contain numbers
  grep -v 'includegraphics' | # Drop lines that are graphics filepaths or numbers setting the figure size
  sed 's/[0-9\.]*\\textwidth//g' |
  sed 's/\$[A-Za-z0-9+=_{}\ ]*\$//g' |  # Drop inline equations that may contain numbers
  sed '/\\begin{figure}/,/\\end{figure}/d' | # Remove lines within figure environments
  sed '/\\begin{table}/,/\\end{table}/d' | # Remove lines within table environments
  # =======================================================================
  # Ignore commented-out lines
  sed '/\\iffalse/,/\\fi/d' | # Remove lines between \iffalse and \fi
  sed '/^%/d' | # Remove lines beginning with %
  # =======================================================================
  # Ignore citations
  sed 's/\\cite{[A-Za-z0-9:,\-]*}//g' | # Drop citations with \cite{} that may contain numbers
  sed 's/\\citealt{[^}]*}//g' |  # Drop citations with \citealt{} that may contain numbers
  sed 's/\\citep{[^}]*}//g' |  # This seems to work for \citep and not line 16..
  sed 's/\\citet{[^}]*}//g' |  # This seems to work for \citep and not line 16..
  # =======================================================================
  grep -wo "[A-Z]\+\{2,20\}" | sort | uniq -c | sort -gr |
  tee -a "$LOG_FILE"  # Log results to file
}

# Main loop to find and process all files named $SRC_TEX
find . -name "$SRC_TEX" | while IFS= read -r file; do
  acronyms "$file"
done
hardcoded-numbers.sh
#!/bin/bash
#This script allow you to review hardcoded numbers appearing in the manuscript
#From https://github.com/jdingel/projecttemplate/blob/master/paper/reviewing/hardcodednumbers.sh
SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"
LOG_FILE="$LOG_DIR/hardcoded-numbers.log"

for file in $(find -name $SRC_TEX)
do
    echo '===' ${file} '==='
    cat ${file} | 
    # =======================================================================
    # Ignore lines relating to environment
    sed -n '/\\begin{document}/,$p' |  # Skip lines until \begin{document}
    grep -Ev '^\\newcommand' | # Ignore lines starting with \newcommand (e.g., \newcommand{\nc}{})
    grep -Ev '^\\setcounter' | # Ignore lines starting with \setcounter
    grep -Ev '^\\vspace' | 
    grep -Ev '^\\hspace' | 
    grep -Ev '^\\usepackage' | 
    grep -Ev '^\\setlength' | 
    # =======================================================================
    # Ignore math, tables, figures
    sed '/begin{equation}/,/end{equation}/d' | sed '/begin{equation\*}/,/end{equation\*}/d' | sed '/begin{align}/,/end{align}/d' | sed '/begin{align\*}/,/end{align\*}/d' | #Remove equation environments
    sed 's/\\input{[A-Za-z0-9_\/\.]*}//g' | # Drop input files that might contain numbers
    grep -v 'includegraphics' | #Drop lines that are graphics filepaths or numbers setting the figure size
    sed 's/[0-9\.]*\\textwidth//g' |
    sed 's/\$[A-Za-z0-9+=_{}\ ]*\$//g' |  #Drop inline equations that may contain numbers
    sed '/\\begin{figure}/,/\\end{figure}/d' | # Remove lines within figure environments
    sed '/\\begin{table}/,/\\end{table}/d' | # Remove lines within table environments
    # =======================================================================
    # Ignore commented-out lines
    sed '/\\iffalse/,/\\fi/d' | # Remove lines between \iffalse and \fi
    sed '/^%/d' | # Remove lines beginning with %
    # =======================================================================
    # Ignore internal links
    sed 's/\\cref{[^}]*}//g' |  # Ignore numbers from \cref{}
    sed 's/\\ref{[^}]*}//g' |  # Ignore numbers from \ref{}
    sed 's/\\nameref{[^}]*}//g' |  # Ignore numbers from \nameref{}
    sed 's/\\crefrange{[^}]*}//g' |  # Ignore numbers from \nameref{}
    # =======================================================================
    # Ignore citations
    sed 's/\\cite{[A-Za-z0-9:,\-]*}//g' | #Drop citations with \cite{} that may contain numbers
    sed 's/\\citealt{[^}]*}//g' |  #Drop citations with \citealt{} that may contain numbers
    sed 's/\\citep{[^}]*}//g' |  # This seems to work for \citep and not line 16..
    sed 's/\\citet{[^}]*}//g' |  # This seems to work for \citep and not line 16..
    sed 's/\\cite{[^}]*}//g' |  
    grep '[0-9]' -n 
done > $LOG_FILE

cat $LOG_FILE
linkchecker.sh
#!/bin/bash
SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"
LOG_FILE="$LOG_DIR/linkchecker.log"

# Ensure the log directory exists
mkdir -p "$LOG_DIR"

# Overwrite the log file at the beginning
> "$LOG_FILE"

# Log and output both stdout and stderr
{
    echo "---> Retrieving URL(s) in \href{}{}:"
    # This one reverses the escaping of \# in the URL
    HREF_LIST=$(grep -o 'href{[A-Za-z0-9:/\._?#=]*}' "$SRC_TEX" | sed 's/.*href{//' | sed 's/}//;s/\\#/#/')

    echo "---> Retrieving URL(s) in \url{}:"
    URL_LIST=$(grep -o '\\url{[^}]*}' "$SRC_TEX" | sed 's/\\url{//;s/}//;s/\\#/#/g')

    echo "---> Retrieving all other URLs"
    GLOBAL_LIST=$(cat "$SRC_TEX" | egrep -o "(http|https)://[a-zA-Z0-9./?=_%:-]*" | sort -u)

    ALL_URLS=$(echo "$HREF_LIST $URL_LIST $GLOBAL_LIST" | tr ' ' '\n' | sort -u)
    for URL in $ALL_URLS; 
    do
        echo "Checking URL: $URL"
        wget --spider --no-verbose "$URL"
    done
} 2>&1 | tee -a "$LOG_FILE"
repeated-strings.sh
#!/bin/bash
SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"
LOG_FILE="$LOG_DIR/repeated-strings.log"

# Overwrite the log file at the beginning
> "$LOG_FILE"

{
    echo "Repeated unigrams:"
    egrep "\b([a-zA-Z]+) \1\b" "$SRC_TEX" -n

    echo "Repeated bigrams:"
    egrep "\b(\S+) (\S+) \1 \2\b" "$SRC_TEX" -n

    echo "Repeated trigrams:"
    egrep "\b(\S+) (\S+) (\S+) \1 \2 \3\b" "$SRC_TEX" -n
} 2>&1 | tee -a "$LOG_FILE"
unreferenced-labels.sh
#!/bin/bash
# From https://github.com/jdingel/projecttemplate/blob/master/paper/reviewing/Makefile
SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"
LOG_FILE="$LOG_DIR/unreferenced-labels.log"

diff \
  --side-by-side \
  --suppress-common-lines \
  <(grep -o --no-filename 'ref{[A-Za-z0-9:_]*}' "$SRC_TEX" | sed 's/ref//' | sort | uniq) \
  <(grep -o --no-filename 'label{[A-Za-z0-9:_]*}' "$SRC_TEX" | sed 's/label//' | sort | uniq) |
  tee "$LOG_FILE"
wordcount.sh
#!/bin/bash
SRC_TEX="paper.tex"
LOG_DIR="./inspecting/logs"

# total: Do not give sums per file, only total sum.
# nobib: Do not include bibliography in count (default).
# sum Make sum of all word and equation counts. May also use
             # -sum=#[,#] with up to 7 numbers to indicate how each of the
             # counts (text words, header words, caption words, #headers,
             # #floats, #inlined formulae, #displayed formulae) are summed.
             # The default sum (if only -sum is used) is the same as
             # -sum=1,1,1,0,0,1,1.
# sub: Generate subcounts. Option values are none, part, chapter,
             # section or subsection. Default (-sub) is set to subsection,
             # whereas unset is none. (Alternative option name is -subcount.)
# inc: Parse included TeX files (as separate file).
# quiet: Quiet mode, no error messages. Use is discouraged!
TEXCOUNT_OUTPUT=$(texcount \
    -total \
    -nobib \
    -sum \
    -sub \
    -inc \
    -quiet \
    "$SRC_TEX"
)

echo "Word count report for $SRC_TEX: $TEXCOUNT_OUTPUT"

# Store total word count
OUTPUT_FILE="$LOG_DIR/wordcount.txt"

STORE=$(texcount \
    -1 \
    -nobib \
    -nosum \
    -quiet \
    "$SRC_TEX")

echo "$STORE" > "$OUTPUT_FILE"

# Store detailed word count
DETAILED_OUTPUT_FILE="$LOG_DIR/wordcount-detailed.txt"

STORE=$(texcount \
    -nobib \
    -quiet \
    "$SRC_TEX")

echo "$STORE" > "$DETAILED_OUTPUT_FILE"

The Makefile with all the recipes would be

Makefile
# .SUFFIXES: .pdf .tex
.DEFAULT_GOAL := help

#==============================================================================
# Inspecting
#==============================================================================
WORDCOUNT_SCR := ./inspecting/wordcount.sh
.PHONY: wordcount
wordcount: ## Wordcount via texcount
    @echo "==> $@"
    @echo "Check word count using texcount"
    dos2unix $(WORDCOUNT_SCR)
    chmod +x $(WORDCOUNT_SCR)
    $(WORDCOUNT_SCR)

ACRONYMS_SRC = ./inspecting/acronyms.sh
.PHONY: acronyms
acronyms: ## Find and tally acronyms
    @echo "==> $@"
    @echo "Check for acronyms"
    dos2unix $(ACRONYMS_SRC)
    chmod +x $(ACRONYMS_SRC)
    $(ACRONYMS_SRC)

.PHONY: dueto
dueto: ## Find "due to"s; Did you mean "because of", "owing to", or "from"?
    @echo "==> $@"
    @echo "Find all the 'due to's in writing"
    grep -n "due to" $(SRC_TEX) > inspecting/logs/duetos.log

.PHONY: duplicated_labels
duplicated_labels: ## Check for duplicated labels
    @echo "==> $@"
    @echo "Check for duplicated labels"
    grep -o '\\label{[^}]*}' $(SRC_TEX) | sort | uniq -cd | tee ./inspecting/logs/duplicated-labels.log

HARDCODEDNUMBERS_SRC = ./inspecting/hardcoded-numbers.sh
.PHONY: hardcodednumbers
hardcodednumbers: ## Find hardcoded numbers
    @echo "==> $@"
    @echo "Check for hardcoded numbers"
    dos2unix $(HARDCODEDNUMBERS_SRC)
    chmod +x $(HARDCODEDNUMBERS_SRC)
    $(HARDCODEDNUMBERS_SRC)

LINKCHECKER_SRC = ./inspecting/linkchecker.sh
.PHONY: linkchecker
linkchecker: ## Check URLs
    @echo "==> $@"
    @echo "Check that URLs work"
    dos2unix $(LINKCHECKER_SRC)
    chmod +x $(LINKCHECKER_SRC)
    $(LINKCHECKER_SRC)

REPEATED_STRINGS_SRC = ./inspecting/repeated-strings.sh
.PHONY: repeated_strings
repeated_strings: ## Check for repeated words
    @echo "==> $@"
    @echo "Check for repeated words (e.g. 'the the table shows...')"
    dos2unix $(REPEATED_STRINGS_SRC)
    chmod +x $(REPEATED_STRINGS_SRC)
    -$(REPEATED_STRINGS_SRC)

.PHONY: textidote
textidote: ## Check with textidote
textidote: ./inspecting/textidote_dict.txt
    @echo "==> $@"
    @echo "Check doc with textidote"
    -textidote --check en --dict $< --output html $(SRC_TEX) > inspecting/logs/textidote.html

UNREFERENCED_LABELS_SRC = ./inspecting/unreferenced-labels.sh
.PHONY: unreferenced_labels
unreferenced_labels: ## Check for label referencing
    @echo "==> $@"
    @echo "Check for unreferenced labels"
    dos2unix $(UNREFERENCED_LABELS_SRC)
    chmod +x $(UNREFERENCED_LABELS_SRC)
    -$(UNREFERENCED_LABELS_SRC)

.PHONY: inspect
inspect: ## Do all inspections of manuscript
inspect: duplicated_labels \
         repeated_strings \
         unreferenced_labels \
         hardcodednumbers \
         acronyms \
         linkchecker \
         textidote \
         dueto \
         wordcount


#==============================================================================
# Additional utilities
#==============================================================================
AY2NUMERIC_SRC = ./inspecting/ay2numeric.sh
.PHONY: aynumeric
aynumeric: ## Change author-year to numeric citation
    @echo "==> $@"
    @echo "Change author-year to numeric citations"
    dos2unix $(AY2NUMERIC_SRC)
    chmod +x $(AY2NUMERIC_SRC)
    $(AY2NUMERIC_SRC)


NUMERICAY_SRC = ./inspecting/numeric2ay.sh
.PHONY: numericay
numericay: ## Change author-year to numeric citation
    @echo "==> $@"
    @echo "Change numeric to author-year citations"
    dos2unix $(NUMERICAY_SRC)
    chmod +x $(NUMERICAY_SRC)
    $(NUMERICAY_SRC)


#==============================================================================
# Help
#==============================================================================
.PHONY: help
help: ## Show this help message and exit    
    @grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | sort | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-16s\033[0m %s\n", $$1, $$2}'

Then, to run all inspections, just do:

$ make inspect

Some shell scripts like the diff to check unreferenced labels can return non-0 status codes and might (annoyingly) turn up as a Make error (make: *** [Makefile: ... Error 1) even if there are no errors.

# Example output
...
...
==> unreferenced_labels
Check for unreferenced labels
dos2unix ./inspecting/unreferenced-labels.sh
dos2unix: converting file ./inspecting/unreferenced-labels.sh to Unix format...
./inspecting/unreferenced-labels.sh
    ...
    ...
make: *** [Makefile:93: unreferenced_labels] Error 1

To make Make ignore this, do a make inspect -k (where the -k prevents Make from stopping).

$ make inspect -k

Or, add a - before the command that would return non-zero status codes (e.g., -./inspecting/unreferenced-labels.sh instead of ./inspecting/unreferenced-labels.sh).

Summary#

And this, in summary, is how to procrastinate on submitting a manuscript.


Home Back to homepage.

Notes See more notes.