#!/usr/bin/ruby
#
# This is a ruby adaptation of bashpodder.shell By Linc 10/1/2004
# Adaptation by sgbirch@imsmail.org licenced under the GPL.
#
# This program expects to be run in $HOME/rubypodder.  If you need to
# run it slesewhere please change the definition of RUNDIR below.
#
# If you want to see what the program is doing, run it like this:
#
#  ruby -d rubypodder.rb
#
# Find the latest bashpodder script at:
#
#   http://linc.homeunix.org:8080/scripts/bashpodder
#
# Last bashpodder revision 07/01/2005 - Many Contributers!
# If you use this and have made improvements or have comments
# drop me an email at linc dot fessenden at gmail dot com
# I'd appreciate it!

require 'fileutils'

#
# Constants
#

RUNDIR = ENV['HOME'] + '/rubypodder'
LOGFILE = 'rubypodder.log'
CONFFILE = 'bp.conf'
OUTDIR = 'podder'

puts "RUNDIR: #{RUNDIR}" if $DEBUG
puts "LOGFILE: #{LOGFILE}" if $DEBUG
puts "CONFFILE: #{CONFFILE}" if $DEBUG
puts "OUTDIR: #{OUTDIR}" if $DEBUG

#
# Switch to home dir so crontab friendly
#

puts "Changing working directory to #{RUNDIR}" if $DEBUG
FileUtils.cd RUNDIR

#
# Read the list of files already retrieved from the
# file pointed to by LOGFILE (usually rubypodder.log).
# The list is cleaned up (lf removed) and placed in
# the variable 'retrieved'.
#

retrieved = []
if File.exist? LOGFILE
  retrieved = open(LOGFILE) do |io|
    retrieved = io.readlines
    retrieved.each {|x| x.chomp!}
  end
end

#
# Create a new sequence number used as a prefix for each sound file.
#

seq = retrieved.length
puts "Next sequence number: #{seq}" if $DEBUG

#
# The next section performs the black magic taken from bashpodder.
# Each feed listed in bp.conf is visited and a list of published
# ogg/mp3 files is created in the 'files' variable.
#

files = []
begin
  open(CONFFILE) do |f|
    f.each_line do |podcast|
      podcast.chomp!
      puts "Checking site: #{podcast}" if $DEBUG
      `wget -q #{podcast} -O - >temp.xml`
      `xsltproc parse_enclosure.xsl temp.xml >temp.fil 2>/dev/null`
      if $? != 0
        system %q!tr '\r' '\n' <temp.xml | tr \' \" | ! +
          %q!sed -n 's/.*url="\([^"]*\)".*/\1/p' >temp.fil!
      end
      open("temp.fil") do |io|
        io.readlines.each { |f| files << f.chomp }
      end
    end
  end
ensure
  FileUtils.rm_f 'temp.fil'
  FileUtils.rm_f 'temp.xml'
end

files = files.sort.uniq
print "Candidates: #{files.length}\n" if $DEBUG


#
# Remove from 'files' all files previously downloaded and
# create the list of new sound files to download (fetchfiles).
#

fetchfiles = files.reject { |url| retrieved.member?(url) }
puts "Files to fetch: #{fetchfiles.length}" if $DEBUG


#
# Make sure the output directory exists OUTDIR (podder)
#

if not File.exists?(OUTDIR)
  puts "Creating #{OUTDIR}" if $DEBUG
  FileUtils.mkdir(OUTDIR)
end

#
# Retrieve the sound files. Update the logfile after each wget.
#

fetchfiles.each do |url|
  url.chomp!
  outfilename = 'podder/' + seq.to_s + '-' + url.split("/")[-1]
  begin
    cmd = "wget -q -O #{outfilename} \"#{url}\""
    puts "#{cmd}" if $DEBUG
    print "outfilename: #{outfilename}\n"
    `#{cmd}`

    if $? == 0
      retrieved << url
      seq += 1
      
      # Update of logfile. This is done via
      # a temp file so we don't end up losing data
      # if the program is aborted in the middle.
      
      FileUtils.rm_f LOGFILE + '.tmp'
      open(LOGFILE + '.tmp', 'w') do |io|
        io.write retrieved.join("\n")
        io.write("\n")
      end
      outfilename = ''
      FileUtils.mv(LOGFILE + '.tmp', LOGFILE)
    end
  ensure
    # Make sure the temp file is removed.
    FileUtils.rm_f outfilename
  end
end