/lab/wgettsv

#! /usr/bin/ruby -w

#================================================================
# wgettsv
# Collect WWW resources and generate TSV data
#================================================================

require 'open-uri'
require 'iconv'
require 'kconv'
require 'date'
require 'time'
require 'cgi'

def main
  seeds = []
  hist = {}
  filters = []
  max = 1 << 30
  lim = 1 << 20
  wait = 0
  ndf = false
  i = 0
  while i < ARGV.length
    if seeds.length < 1 && ARGV[i] =~ /^-/
      if ARGV[i] == '-allow'
        usage if (i += 1) >= ARGV.length
        regex = Regexp::new(ARGV[i])
        filters.push([true, regex]) if regex
      elsif ARGV[i] == '-deny'
        usage if (i += 1) >= ARGV.length
        regex = Regexp::new(ARGV[i])
        filters.push([false, regex]) if regex
      elsif ARGV[i] == '-max'
        usage if (i += 1) >= ARGV.length
        max = ARGV[i].to_i
      elsif ARGV[i] == '-lim'
        usage if (i += 1) >= ARGV.length
        lim = ARGV[i].to_i
      elsif ARGV[i] == '-wait'
        usage if (i += 1) >= ARGV.length
        wait = ARGV[i].to_f
      elsif ARGV[i] == '-ndf'
        ndf = true
      else
        usage
      end
    else
      if ARGV[i] =~ /^http:\/\//i
        seeds.push(ARGV[i])
        hist[ARGV[i]] = true
      else
        usage
      end
    end
    i += 1
  end
  usage if seeds.length < 1
  if !ndf
    filters.push([false, /\.(txt|text|asc|c|cc|cxx|cpp|h|hxx|hpp|in)$/i])
    filters.push([false, /\.(css|js|csv|tsv|log|md5|crc|conf|ini|inf|lnk|sys|tmp|bak)$/i])
    filters.push([false, /\.(xml|xsl|xslt|rdf|rss|dtd|sgml|sgm)$/i])
    filters.push([false, /\.(pgp|sig|cer|csr|pem|key|b64|uu|uue|[0-9])$/i])
    filters.push([false, /\.(rtf|pdf|ps|eps|ai|doc|xls|ppt|sxw|sxc|sxi|xdw|jtd|oas|swf)$/i])
    filters.push([false, /\.(zip|tar|tgz|gz|bz2|tbz2|z|lha|lzh)?$/i])
    filters.push([false, /\.(7z|lzo|lzma|cpio|shar|cab|rar|sit|ace|hqx)?$/i])
    filters.push([false, /\.(bin|o|a|so|exe|dll|lib|obj|ocx|class|jar|war)?$/i])
    filters.push([false, /\.(rpm|deb|qdb|qdb|dbx|dbf|dat|msi|bat|com|iso)?$/i])
    filters.push([false, /\.(png|gif|jpg|jpeg|tif|tiff|bmp|ico|pbm|pgm|ppm|xbm|xpm|dvi)$/i])
    filters.push([false, /\.(au|snd|mid|midi|kar|smf|mp2|mp3|m3u|wav|wma|wmp|asx|at3|aif)$/i])
    filters.push([false, /\.(mpg|mpeg|qt|mov|avi|wmv|wvx|asf|ram|rm)$/i])
    filters.push([false, /\.(tch|tdb|tdf|tct)$/i])
    filters.push([false, /\.idx\./i])
    filters.push([false, /(core|casket|Makefile|README|NEWS|COPYING|LISENCE)($|\/)/i])
  end
  return proc(seeds, hist, filters, max, lim, wait) ? 0 : 1
end

def usage
  STDERR.printf("%s: collect WWW resources and generate TSV data\n", $progname)
  STDERR.printf("\n")
  STDERR.printf("usage:\n")
  STDERR.printf("  %s [-allow regex] [-deny regex] [-max num] [-lim num] [-wait num]" +
                " url ...\n", $progname)
  STDERR.printf("\n")
  exit(1)
end

def proc(seeds, hist, filters, max, lim, wait)
  cnt = 0
  while (url = seeds.shift) && cnt < max
    STDERR.printf("%d: getting: %s\n", cnt + 1, url)
    begin
      opts = {}
      OpenURI.open_uri(url, 0, 0, opts) do |sio|
        baseuri = sio.base_uri
        if baseuri && baseuri.to_s != url
          url = baseuri.to_s
          hist[url] = true
        end
        size = sio.size
        raise "invalid size" if size > lim || size < 3
        type = sio.content_type
        type = "text/plain" if !type
        str = sio.read
        head = str[0,2048]
        if (head[0] == 0xfe && head[1] == 0xff) || (head[0] == 0xff && head[1] == 0xfe)
          str = Kconv::kconv(str, Kconv::UTF8, Kconv::UTF16)
          charset = "UTF-8"
        elsif str.include?(0)
          raise "binary data"
        end
        raise "not HTML" if type != "text/html" && head !~ /<html/i && head !~ /<body/i
        if !charset && head =~ /<\?xml.*encoding=("|')?[-_a-zA-Z0-9]+("|')/im
          charset = head.gsub(/.*<\?xml.*encoding=["']?([-_a-zA-Z0-9]+)["']?.*/im, '\1')
        end
        if !charset && head =~ /<meta.*content-type.*charset=[-_a-zA-Z0-9]+/im
          charset = head.gsub(/.*<meta.*content-type.*charset=([-_a-zA-Z0-9]+).*/im, '\1')
        end
        charset = sio.charset if !charset || charset.length < 1
        if charset && charset.length > 0
          if charset !~ /^UTF-?8$/i
            begin
              nstr = Iconv.conv("UTF-8", charset, str)
              str = nstr if nstr && nstr.length > 0
            rescue
              str = str.toutf8
            end
          end
        else
          str = str.toutf8
        end
        body = str.gsub(/.*<body[^>]*>/im, "")
        body = body.gsub(/<\/body>.*/im, "")
        body = htmltotext(body)
        if str =~ /<title[^>]*>[^<]*<\/title>/im
          title = str.gsub(/.*<title[^>]*>([^<]*)<\/title>.*/im, '\1')
          title = htmltotext(title)
        end
        title = "" if !title
        title = title[0,128] if title.length > 128
        mtime = sio.last_modified
        if mtime
          mtime = Time::parse(mtime.to_s).to_i
        else
          mtime = 0
        end
        printf("%d", cnt + 1)
        printf("\turl\t%s", url)
        printf("\tsize\t%s", size) if size > 0
        printf("\tmtime\t%s", mtime) if mtime > 0
        printf("\ttitle\t%s", title) if title.length > 0
        printf("\tbody\t%s", body) if body.length > 0
        printf("\n")
        str.gsub(/<a[^>]*>/im) do |tag|
          if tag =~ /href=["']?[^"'>]+["']?/
            href = tag.gsub(/.*href=["']?([^"'>]+)["']?.*/, '\1')
            href = URI::join(url, href).to_s
            href = href.gsub(/#.*/, "")
            if !hist[href] && checkurl(href, filters)
              seeds.push(href)
              hist[href] = true
            end
          end
        end
      end
      cnt += 1
    rescue
      STDERR.printf("%d: failed: %s: %s\n", cnt + 1, url, $!)
    end
    sleep(wait) if wait > 0
  end
  return 0
end

def htmltotext(str)
  str = str.gsub(/<style[^>]*>.*?<\/style>/im, " ")
  str = str.gsub(/<script[^>]*>.*?<\/script>/im, " ")
  str = str.gsub(/<\/?(p|br|div|h1|h2|h3|h4|h5|h6|ul|ol|dl|li|dd|dt|td|th|pre)[^>]*>/im, " ")
  str = str.gsub(/<[^>]*>/, "")
  str = str.gsub(/&(nbsp|#160|#0160|#xa0|#x00a0);/i, " ")
  hexrx = Regexp::new("^&#x[0-9a-zA-Z]+;")
  decrx = Regexp::new("^&#[0-9]+;")
  str = str.gsub(/&#?[A-Za-z0-9]+;/) do |pat|
    case pat
    when "&lt;"
      pat = '<'
    when "&gt;"
      pat = '>'
    when "&quot;"
      pat = '"'
    when "&apos;"
      pat = "'"
    when "&nbsp;"
      pat = " "
    else
      begin
        if pat =~ hexrx
          pat = [ pat.gsub(/&#x([A-Za-z0-9]+);/i, '\1').hex ].pack("n")
          pat = Iconv.conv("UTF-8", "UTF-16BE", pat)
        elsif pat =~ decrx
          pat = [ pat.gsub(/&#([A-Za-z0-9]+);/i, '\1').to_i ].pack("n")
          pat = Iconv.conv("UTF-8", "UTF-16BE", pat)
        else
          pat = " "
        end
      rescue
        pat = ""
      end
    end
    pat
  end
  str = str.gsub(/[\x00-\x20]/, " ")
  str = str.gsub(/\xe3\x80\x80/, " ")
  str = str.gsub(/ +/, " ")
  str = str.gsub(/^ */, "")
  str = str.gsub(/ *$/, "")
  return str
end

def checkurl(url, filters)
  return false if url !~ /^http:\/\//i;
  return true if filters.length < 1
  ok = !filters[0][0]
  filters.each do |filter|
    ok = filter[0] if url =~ filter[1]
  end
  return ok
end

STDOUT.sync = true
$progname = $0.dup
$progname.gsub!(/.*\//, "")
srand
exit(main)