Iterate over files and convert with a new name / directory

Question

I'm pretty new to bash scripting. I want a script that I can run in a directory that contains .md files, runs pandoc on them, and delivers the output to a sibling folder html, i.e. my folder structure is like this:

.
├── html
└── md

Here is what I have working, I just don't know if I'm following good practices or if there's a nicer / more readable way of doing what I'm doing.

#! /bin/bash
# run this from a folder containing markdown files
# it delivers html to the sibling html folder
if hash pandoc 2>/dev/null; then
  for file in ./*.md ; do
    echo Converting $file
    cat "$file" | pandoc -o "$(echo $file | sed 's/^./..\/html/' |  sed 's/.md$/.html/')"
  done
else
  echo "I need pandoc installed"
fi

Answer 1

• UUOC. pandoc takes the input file name as an argument. echo $file | pandoc -o outfile is equivalent to pandoc -o outfile $file. One process invocation down.

• UUOS (sed in this case). bash has very rich string transformation features built in. For example, ${file%md} (see Parameter expansion section of man bash) will strip the md suffix form the filename. So consider

  for file in *.md; do
      pandoc -o ../html/${file%md}html
  done
  

  Two more process invocations down.

• If by any chance an .md filename contains funny characters (like whitespace), the substitution will produce unexpected results. It is safer to use double quotes, as in `"${....}".

PS: It might be beneficial to abandon the shell script altogether in favor of make, with a Makefile having a single stem rule

../html/%.html: %.md
    pandoc -o $@ $<