##
## addhrefs.py
## by Peter Bengtsson, 2004-2005, mail@peterbe.com
##
## License: ZPL (http://www.zope.org/Resources/ZPL)
##
__doc__='''A little function that puts HTML links into text.'''
__version__='0.9.3'


__changes__ = '''
0.9.3         Fixed a bug when text contains "m." but wasnt followed but a a-z.

0.9.2         Added supports for URLs starting with m., mobile. and www2.
              Added 1 new unit test

0.9.1         Fixed broken link parsing containing {curly brackets}
              Added 15 new unit tests

0.9           Better support for strings already containing <a href.
              New unittest called testAddhrefs.py

0.8           Improvements to emaillinkfunction and urllinkfunction
              
0.7           addhrefs() checks that emaillinkfunction and urllinkfunction if passed
              are callable.
    
0.6           three new parameters:
                 return_everything=0 - returns (text, urlinfo, emailinfo)
                                       where urlinfo and emailinfo are lists
                 emaillinkfunction=None - callback function for making an email
                                          into a HTML link the way you want it
                 urllinkfunction=None - callback function for making an URL into
                                        a HTML link the way you want it
                                        
0.5           "foo https bar" created one link

0.4           Python2.1 compatible with custom True,False

0.3           Only one replace which solved the problem with
              "www.p.com bla bla http://www.p.com"
              
0.2           Lots of fine tuning. Created unit tests.

0.1           Project started. First working draft
'''

__credits__='''
David Otton,
"flump cakes"
'''

                     
import re, sys


_end_dropouts = list(')>.;:,"')
_start_dropouts = list('(<')
def _massageURL(url):
    while url[-1] in _end_dropouts:
        url = url[:-1]
    if url[0] in _start_dropouts:
        url = url[1:]
  
    return url

def _improveURL(url):
    # ok_middle_name_starts looks something like this:
    #  ('ftp','http','www.','mobile.','m.','www2.')
    # If our url here starts with any of those that end in a .
    # then add http:// to it
    for each in ok_middle_name_starts:
        if each.endswith('.') and url.startswith(each):
            return 'http://'+url
    return url


def _makeLink(url):
    return '<a href="%s">%s</a>'%(_improveURL(url), url)

def _makeMailLink(url):
    return '<a href="mailto:%s">%s</a>'%(_improveURL(url), url)

def _rejectEmail(email, start):
    if email.startswith("mailto:"):
        email = email[7:]
    if email.find(':') > -1:
        return True
    return False

_bad_in_url = list('!()<>')
_dont_start_url = list('@')
def _rejectURL(url, start):
    """ return true if the URL can't be a URL """
    if url.lower()=='https':
        return True
    for each in _bad_in_url:
        if url.find(each) > -1:
            return True
    whereat = url.find('@')
    if whereat > -1:
        for each in "http:// ftp:// https://".split():
            url = url.replace(each, '')
        if not -1 < url.find(':') < whereat:
            return True
    if start in _dont_start_url:
        return True
    return False


def _make_regexp(regexp):
    _whitespace = "[\s\({}<>\)]"
    #_not_whitespace  = "[^\s\({}<>\)]"
    _not_whitespace  = "[^\s{}<>]"
    ## don't allow url to end in ( or < but fine with ) or >
#    _not_whitespace  = "[^\s<>\)]" 
    regexp = regexp.replace("\s", _whitespace)
    regexp = regexp.replace("\S", _not_whitespace)
    regexp = re.compile(regexp)
    return regexp

ok_middle_name_starts = ('ftp','http','www.','mobile.','m.','www2.')
ok = {'start': ('^','\(','{','>','<','@','\s',''),
      'middle':('ftp\S+', 'http\S+', 'www\.\w\S+', 'mobile\.\w\S+', 'm\.\w\S+',),
      'end':('\)','}','>','\s','$'),
      }
      
#_url_regex = _make_regexp('((^|\(|<|@|\s|)(ftp\S+|http\S+|www\.\S+)(\)|>|\s|$))')
_or = lambda some_list: "|".join(some_list)
_url_regex = _make_regexp('((%s)(%s)(%s))'%(_or(ok['start']), _or(ok['middle']), _or(ok['end'])))
    
#_mailto_regex = re.compile('((^|\(|<|\s|)(\S+@\S+\.\S+)(\)|>|\s|$))')
_mailto_regex = _make_regexp('((%s)(\S+@\S+\.\S+)(%s))' % (_or(ok['start']), _or(ok['end'])))

                                          
def addhrefs(text, return_everything=0, 
             emaillinkfunction=_makeMailLink,
             urllinkfunction=_makeLink):
    
    if not callable(emaillinkfunction):
        if emaillinkfunction is not None:
            _msg = "%r is not callable email link function"
            print >>sys.stderr, _msg%emaillinkfunction
        emaillinkfunction = _makeMailLink
    
    if not callable(urllinkfunction):
        if urllinkfunction is not None:
            _msg = "%r is not callable URL link function"
            print >>sys.stderr, _msg%urllinkfunction
        urllinkfunction = _makeLink
    
    info_emails = []
    info_urls = []
    
    urls = _url_regex.findall(text)
    for each in urls:
        whole, start, url, end = each
        if whole.endswith('">'):
            # reject it because it looks like it's taken out of a tag
            continue
        if whole.endswith('<'):
            # the next thing is a tag, if that tag is a </a>
            # the chicken out!
            pos = text.find(whole)
            if text[pos+len(whole)-1:pos+4+len(whole)] == '</a>':
                continue
        #print each

        url = _massageURL(url)
        
        if _rejectURL(url, start):
            continue
        link = urllinkfunction(url)
        if return_everything:
            info_urls.append((url, link))
        better = whole.replace(url, link)
        text = text.replace(whole, better, 1)
        
    mails = _mailto_regex.findall(text)
    for each in mails:
#        print each
        whole, start, url, end = each
        url = _massageURL(url)
        if _rejectEmail(url, start):
            continue
        if url.find(':') > -1:
            link = urllinkfunction(url)
            if return_everything:
                info_urls.append((url, link))
            better = whole.replace(url, link)
        else:
            link = emaillinkfunction(url)
            info_emails.append((url, link))
            better = whole.replace(url, link)
        text = text.replace(whole, better)        

    if return_everything:
        return text, info_urls, info_emails
    else:
        return text


def test():
    raise "TODO", "Move these slowly into testAddhrefs.py"

    t="this some text http://www.peterbe.com/ with links www.peterbe.com in it"
    
    t='''this <a href="http://www.google.com">some</a> text http://www.peterbe.com/
    with links www.peterbe.com in it <a href="http://www.example.com">Example</a>'''
    

    

    t2='this <a href="http://www.google.com">some</a> text http://www.peterbe.com/ '\
       'with links www.peterbe.com in it '\
      '<a href="http://www.example.com">Example</a>'
    print addhrefs(t)

    t3='''this <a href="http://www.google.com">some</a> text http://www.peterbe.com/
with links www.peterbe.com in it <a href="http://www.example.com">Example</a>
www,peterbe.com and <a href="www.peterbe.com">www.peterbe.com</a>
    '''

    t4='''https://www.imdb.com (www.peterbe.com/?a=e) asd tra la www.google.com'''
    
    t = 'word (www.peterbe.com) word'
    
    t = 'word <http://www.peterbe.com?poage=2314> and so on'
    
    t = 'Go to: http://www.peterbe.com. There youll find'
    
    t = 'Go to: http://www.peterbe.com:'
    t = '''https://www.imdb.com.'''
    t = 'Hello mail@peterbe.com to you'
    t = 'Hello <mail@peterbe.com> and <www.google.com> to you'
    #t = open('sample-htmlfree.txt').read()
    t = '<a href="link1.html">Link1</a> link www.2.com'
    t = "<a href='www.google.com'>Link1</a> link www.2.com"
    t = '''1. http://www.peterbe.com
2. www.peterbe.com
3. <http://www.peterbe.com>
4. mail@foobar.com
5. "Name <mail@foobar.com>"'''
    t = 'xxx mail@peterbe.com peter@grenna.net'
    t += ' xxx www.peterbe.com www.google.com xxx'

    t = 'mail@peterbe.com 123@a.com or www2.ibm.com or www.ibm.com?asda=ewr&amp;gr:int=34.'
    
    t = 'peter@grenna.net 123@a.com ftp://ftp.uk.linux.org/'
    t = 'http://david:otton@www.something.com david:otton@www.something.com'
    t = ''' <http://www.google.com/?asd=q32> xxx
    <http://www.peterbe.com> abc <mail@peterbe.com>'''
    t='''www.msn.co.uk

http://msn.co.uk
http://www.msn.co.uk

ftp:/google.com
'''
    t = 'At http://localhost/ I have apache and at http://localhost:8080 '\
        ' I have Zope David used http://enchanter or http://enchanter/'
    t = 'See http://www.something.com/page?this=that#001'
    t = 'Bla bla https bla bla and http bla'
    t = '<p>mail@peterbe.com</p>\n\n<p>www.something.com</p>'
    t = '<p>http://something.com</p>\n\n<p>mail@peterbe.com</p>'
    t = '<p>http://example.com</p>\n\n<p>kilobug@freesurf.fr</p>'
    t += '\n\nhttp://www.dil(bert.com'
    
    



        
    
    
if __name__=='__main__':
    test()