C[omp]ute

From: andrew cooke <andrew@...>

Date: Fri, 30 May 2014 16:01:01 -0400

I am working on a project that I have been docmuenting in a GitHub wiki, using
markdown.  Now we want to provide the client with a "paper" document.  Rather
than copy and paste everything, I am automating the document generation from
the Wiki.

(Obviously the code below isn't perfect - in particular, there's no real order
to the pages - but it's a good first start and can be extended as required.)


First, the driver script that generates the docs (change WIKI-DIR and WIKI-URL
as required);

  #!/bin/bash

  if [ ! -e WIKI-DIR ]; then
      git clone WIKI-URL
  else
      pushd WIKI-DIR
      git pull
      popd
  fi

  REPORT=report.md
  rm -f $REPORT
  cp header.md $REPORT

  for WIKIPAGE in `ls -1 cats-scripts.wiki | grep -v "~" | grep -iv "home.md" `
  do
  TITLE=`echo $WIKIPAGE | sed 's/-/ /g' | sed 's/.md$//'`
      echo "" >> $REPORT
      echo "# $TITLE" >> $REPORT
      echo "" >> $REPORT
      ./fix-md.py WIKI-DIR/$WIKIPAGE | sed 's/^% .*//' >> $REPORT
  done

  cat footer.md >> $REPORT
  pandoc -s --toc -f markdown $REPORT -o report.pdf


Second, a Python script that youcan use to modify each page as required:

  #!/usr/bin/python

  # convert wiki pages so that they are suitable for inclusion in the report

  # reads from stdin (or args), writes to stdout. i can't remember enough
  # awk...

  from __future__ import print_function
  import fileinput
  import re

  HEADER = re.compile(r'^(#+)\s+(.*)')

  for line in fileinput.input():

      # replace wikilinks by normal links
      line = re.sub(r'\[\[([^]]*)\]\]', r'[\1](\1)', line)

      # headers are increased one level and auto-capped
      header = HEADER.match(line)
      if header:
	  print('#%s ' % header.group(1), end='')
	  for word in header.group(2).split():
	      print(word[0] + word[1:].lower() + ' ', end='')
	  print()

      # anything else is copied to stdout
      else:
	  print(line, end='')


Also, you need to define your own footer.md and header.md.

Andrew