build

#!/bin/bash -evx

# Support generation of HTML, PDF, and other versions of a single master CV,
# which is kept under source control in Markdown format.
#
# Inspired by Mark Szepieniec's project, "The Markdown Resume":
#
# https://mszep.github.io/pandoc_resume/

# Ensure that the current working directory is the script parent directory (the
# directory in which this Bash script resides); this is necessary for proper
# file access in the Docker container used for the conversions.

cd "$(dirname "${BASH_SOURCE[0]}")"

# Name of the directory in which to place generated files, relative to the
# script parent directory.

output_dir='output'

# Name of the directory containing generic content for inclusion in generated
# files, relative to the script parent directory.

include_dir='include'

# Filename prefix of master input file.

input='cv'

# Master input file, in Markdown format.

input_md="$input.md"

# Output file consisting of master input file with full contact information
# added, if available, in Markdown format.

output_md="$output_dir/$input_md"

# Microsoft Word, HTML, PDF, RTF, and LaTeX output documents.

output_docx="$output_dir/$input.docx"
output_html="$output_dir/$input.html"
output_pdf="$output_dir/$input.pdf"
output_rtf="$output_dir/$input.rtf"
output_tex="$output_dir/$input.tex"

# Pattern in the CV to be replaced with full contact information, if available.

contact_search='^<https://www[.]linkedin[.]com/.*>$'

# Optional file containing full contact information.

contact_replace="$include_dir/contact.md"

# String of sed commands for transforming the ASCII representations " - " and
# "--" of en and em dashes, respectively, into the "--" and "---" expected by
# LaTeX.  Note that word boundaries ("\b") behave differently in the BSD
# (MacOs) and GNU (Linux) versions of sed, so we match non-hyphen characters
# instead.

dash_munge='s|\([^-]\)--\([^-]\)|\1---\2|g;s|\([^-]\) - \([^-]\)|\1--\2|g'

# Intermediate file for dash transformation.

munged_md="$output_dir/$input.munged.md"

# Command to run Pandoc in a Docker container, mounting a Docker volume in
# order to allow access to files in the current directory and its
# subdirectories.  The output file is a standalone document, not a fragment,
# and "smart" processing of apostrophes, quotation marks, hyphens, and dashes
# is enabled.  The MSYS_NO_PATHCONV variable is set to allow proper operation
# in the Windows Git Bash environment.

pandoc="MSYS_NO_PATHCONV=1 docker run --volume='$PWD:/source' \
jagregory/pandoc --standalone --smart"

# Generic HTML header specifying formatting options like the default typeface
# style.

header_html="$include_dir/header.html"

# Generic LaTeX header specifying formatting options like like the margins and
# the default typeface style.

header_tex="$include_dir/header.tex"

# Generic Microsoft Word reference document specifying formatting options like
# the default typeface style.

reference_docx="$include_dir/reference.docx"

# Create the output directory if it doesn't already exist.

mkdir -p "$output_dir"

# Remove output and intermediate files, if present, to ensure a clean build.

rm -f "$munged_md" "$output_docx" "$output_html" "$output_md" "$output_pdf" \
    "$output_rtf" "$output_tex"

# Generate final Markdown file with full contact information, if available.

if [ -r "$contact_replace" ]; then
    sed "\|$contact_search|{r $contact_replace
        d;}" "$input_md" > "$output_md"
else
    cp "$input_md" "$output_md"
fi

# Generate intermediate Markdown file for processing into LaTeX document,
# transforming ASCII representations of dashes.

sed "$dash_munge" "$output_md" > "$munged_md"

# Generate Microsoft Word document intermediate Markdown file and generic
# reference document.

eval $pandoc --reference-doc="$reference_docx"  \
    --from='markdown' --to='docx'  --output="$output_docx" "$munged_md"

# Generate HTML document intermediate Markdown file and generic HTML header.

eval $pandoc --include-in-header="$header_html" \
    --from='markdown' --to='html5' --output="$output_html" "$munged_md"

# Generate PDF document from intermediate Markdown file and generic LaTeX
# header; as documented here, the output format must be specified as "latex"
# instead of "pdf":
#
# https://hub.docker.com/r/jagregory/pandoc/

eval $pandoc --include-in-header="$header_tex"  \
    --from='markdown' --to='latex' --output="$output_pdf"  "$munged_md"

# Generate RTF document from intermediate Markdown file.

eval $pandoc                                    \
    --from='markdown' --to='rtf'   --output="$output_rtf"  "$munged_md"

# Generate LaTeX document from intermediate Markdown file.

eval $pandoc                                    \
    --from='markdown' --to='latex' --output="$output_tex"  "$munged_md"

# Remove intermediate file.

rm -f "$munged_md"