#!/usr/bin/env python
"""
metagen.py - Generate Metadata Information
Author: Sean B. Palmer, inamidst.com
"""

import cgitb; cgitb.enable()
import sys, os, re, urlparse

# @@ Specify this on the command line
meta = './meta'
if os.environ.has_key('DOCUMENT_ROOT'): 
   metadir = os.path.join(os.environ['DOCUMENT_ROOT'], './meta')
   if os.path.isdir(metadir): 
      meta = metadir

t_template = '(?ms)(<!-- *Template: *([a-z]+) *-->).+?(<!-- */Template *-->)'
r_template = re.compile(t_template)

def escape(html): 
   html = html.replace('&', '&amp;')
   return html.replace('<', '&lt;')

def lstrip(start, text): 
   if text.startswith(start): 
      return text[len(start):]
   return text

def urijoin(site, *parts): 
   path = os.path.join(*parts)
   normpath = os.path.normpath(path)
   if path.endswith('/'): normpath += '/'
   return urlparse.urljoin(site, path)

def urisplit(uri): 
   scheme, host, path, query, fragid = urlparse.urlsplit(uri)
   if not scheme and not host: 
      return '/', path
   return '%s://%s/' % (scheme, host), path

def fn(path): 
   if path.endswith('/'): 
      path += 'index'

   if path.startswith('/'): 
      path = '.' + path
   elif not path.startswith('.'): 
      path = './' + path

   return os.path.join(meta, path)

def exists(path): 
   filename = fn(path)
   return os.path.isfile(filename)

class Metadata(dict): 
   def __init__(self, filename=None, path=None): 
      dict.__init__(self)
      if path is not None: 
         filename = fn(path)
      self.filename = filename
      self.parse()

   def parse(self): 
      f = open(self.filename)
      for line in f: 
         line = line.rstrip('\r\n')
         if (not line.startswith('#')) and line: 
            key, value = line.split(': ', 1)
            if key == 'keywords': 
               value = frozenset(value.split(' '))
            self.__setitem__(key, value)
         else: continue
      f.close()

def fileinfo(path): 
   metadata = {}
   f = open(path)
   for line in f: 
      line = line.rstrip('\r\n')
      if (not line.startswith('#')) and line: 
         metadata.__setitem__(*line.split(': ', 1))
      else: continue
   f.close()
   return metadata

def makelink(uri, metadata, highlight=True): 
   if metadata.has_key('keywords'): 
      keywords = frozenset(metadata['keywords'].split(' '))
   else: keywords = frozenset([])

   if highlight and ('highlight' in keywords): 
      link = '<a href="%s">%s</a><sup class="highlight">*</sup>'
   else: link = '<a href="%s">%s</a>'

   result = link % (uri, escape(metadata.get('title')))
   if ('cgi' in keywords) or ('svc' in keywords): 
      site, path = urisplit(uri)
      srcuri = urlparse.urljoin(site, 'inside' + path)
      result += ' (<a href="%s">src</a>)' % srcuri
   return result + ' - ' + escape(metadata.get('description'))

def getclass(metadata): 
   if metadata.has_key('keywords'): 
      keywords = frozenset(metadata['keywords'].split(' '))
      if 'act' in keywords: return ' class="act"'
      if 'code' in keywords: return ' class="code"'
      if 'doc' in keywords: return ' class="doc"'
      if 'pub' in keywords: return ' class="pub"'
      if 'svc' in keywords: return ' class="svc"'
   return ''

def bykeyword(keyword=None): 
   if keyword is None: 
      result = set()
   else: result = []

   for root, dirs, files in os.walk(meta): 
      if root == meta: 
         del dirs[dirs.index('meta')]
         files.remove('.htaccess')
         files.remove('index.cgi')

      dirname = '/' + root[len(meta):].lstrip('/')
      for filename in files: 
         metafn = os.path.join(root, filename)
         metadata = fileinfo(metafn)
         path = os.path.join(dirname, filename)
         if path.endswith('/index'): 
            path = path[:-5]

         if metadata.has_key('keywords'): 
            if keyword is not None: 
               keywords = frozenset(metadata['keywords'].split(' '))
               if keyword in keywords: 
                  result.append((path, metafn, metadata))
            else: result |= set(metadata['keywords'].split(' '))
   return sorted(result)

def sitemap(): 
   smap = {}
   result = []
   for root, dirs, files in os.walk(meta): 
      if root == meta: 
         del dirs[dirs.index('meta')]
         files.remove('.htaccess')
         files.remove('index.cgi')

      root = lstrip(meta, root).lstrip('/')
      if '/' in root: 
         base, rest = root.split('/', 1)
      else: base, rest = root, ''

      for file in files: # @@ was sorted
         if file == 'index': file = ''
         if not rest: parts = (file,)
         else: parts = (rest, file)

         try: smap[base].append(parts)
         except KeyError: smap[base] = [parts]

   result.append('<ul class="sitemap">')
   for key in sorted(smap.iterkeys()): 
      link = urijoin('http://inamidst.com/', key + '/')
      title = os.path.normpath('/' + key + '/').rstrip('/') + '/'

      subfiles = []
      for parts in sorted(smap[key]): 
         linkuri = urijoin('http://inamidst.com/', *(key,) + parts)
         metafn = os.path.join(*(meta, key) + parts)
         if metafn.endswith('/'): metafn += 'index'
         subfiles.append((linkuri, metafn))

      about = '<strong>in <a href="%s">%s</a></strong>: '
      if subfiles[0][0] == link: 
         metadata = fileinfo(subfiles[0][1])

         result.append('   <li%s>' % getclass(metadata))
         result.append('   ' + about % (link, title))
         result.append('   ' + makelink(link, metadata))
         subfiles = subfiles[1:]
      else: result.append(('   <li>' + about) % (link, title))

      if subfiles: 
         result.append('     <ul>')
         for (linkuri, metafn) in subfiles: 
            metadata = fileinfo(metafn)
            result.append('       <li%s>' % getclass(metadata))
            result.append('       %s</li>' % makelink(linkuri, metadata))
         result.append('     </ul>')
      result.append('   </li>')
   result.append('</ul>')
   return '\n'.join(result) + '\n'

def keywordmap(files): 
   result = ['<ul>']
   for path, metafn, metadata in files: 
      tag = '<li>' # @@ ? % getclass(metadata)
      result.append(tag + makelink(path, metadata) + '</li>')
   result.append('</ul>')
   return '\n'.join(result)

def template(m): 
   opener, name, closer = tuple(m.groups())
   if name != 'sitemap': 
      files = bykeyword(name)
      return opener + '\n' + keywordmap(files) + closer
   else: return opener + '\n' + sitemap() + closer

def main(): 
   f = open(sys.argv[1])
   data = f.read()
   f.close()
   print "Read %s bytes..." % len(data)

   data = r_template.sub(template, data)
   f = open(sys.argv[1], 'w')
   f.write(data)
   f.close()
   print "Wrote %s bytes back to %s" % (len(data), sys.argv[1])

if __name__=="__main__": 
   main()