Docgen, a silly personal script to handle pre and post processing of Pandoc files

gistfile1.py

#! /usr/bin/env python
import re
import subprocess
import tempfile
import sys
import os

class Processor(object):
  def __init__(self):
    pass

  def get_file_lines(self, file_name):
    file_handle = open(file_name, 'r')
    lines = file_handle.readlines()
    file_handle.close()

    return lines

  def write_temp_file(self, lines, file_name='docgen'):
    f = open(file_name, 'w')
    f.writelines(lines)
    f.close()

    return f.name

class LaTeXProcessor(Processor):
  def output(self, input_file_name):
    input_file_name = sys.argv[1]
    jobname = os.path.splitext(input_file_name)[0]
    formatted_input = self.format_citations(self.get_file_lines(input_file_name))
    pandoc_output = self.get_pandoc_output(formatted_input)

    # Output a temp file, even though this is a pain, it seems to be
    # the only way to make pdflatex see the output correctly.
    # Having the output written out is useful, anyway, so I'm not losing
    # sleep over it.
    temp_file_name = self.write_temp_file(pandoc_output, file_name=jobname + '.tex')

    self.run_pdflatex(temp_file_name, jobname=jobname)

    self.run_bibtex(jobname=jobname)

    # Rerun pdflatex again to fix bibliography citations, once
    # for it to notice the citations, once for it to get the labelling right
    for i in range(0,2):
      print self.run_pdflatex(temp_file_name, jobname=jobname)

  def format_citations(self, lines):
    """Return lines with Pandoc citations translated to LaTeX ones.

    Note: The extra slashes are due to Python's interpreter, this is the correct
    output for just a single \cite command.

    >>> format_citations(['foo [@bar]'])
    ['foo \\\cite{bar}']

    >>> format_citations(['foo', '[@bar] baz'])
    ['foo', '\\\cite{bar} baz']

    It should also work with more complicated citations.

    >>> format_citations(['[@foo:bar]'])
    ['\\\cite{foo:bar}']
    """
    return [re.sub('\[@(\S+)\]', '\cite{\g<1>}', l) for l in lines]

  def run_text_processor(self, execution_method, stdin=None):
    processor = subprocess.Popen(execution_method, stdin=subprocess.PIPE, 
                                 stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    
    if stdin:
      (output, errors) = processor.communicate(input=''.join(stdin))
    else:
      (output, errors) = processor.communicate()

    if errors:
      print "Text processor error: " + errors
    
    return output

  def get_pandoc_output(self, lines):
    return self.run_text_processor(['pandoc',
                                    '-s', '-t', 'latex', 
                                    '--template=/Users/cflewis/Library/Pandoc/acm.tex'], 
                                    stdin=lines)

  def run_pdflatex(self, file_name, jobname='docgen'):
    return self.run_text_processor(['pdflatex', 
                                    '--interaction=batchmode',
                                    '--jobname=' + jobname,
                                    file_name])

  def run_bibtex(self, jobname='docgen'):
    try:
      print subprocess.check_output(['bibtex', jobname + '.aux'])
    except subprocess.CalledProcessError as e:
      print "BibTeX error: " + e

latex = LaTeXProcessor()
latex.output(sys.argv[1])