"""RelPathRedux

update to relpath
"""

__author__ = "Josh English (english@spiritone.com)"
__version__ = "$Revision: 0.2 $"
__date__ = "$Date: 2003/08/85 16:33:42 $"
__copyright__ = "Copyright (c) 2003 Josh English"
__license__ = "Python"

__history__ = """
0.1 2003-08-05: Initial release 
0.2 2004-07-24: Updated code to use urlparse, added ability to detect fragments
"""


import urlparse

class RelPathError(Exception): pass
class NotFullPathError(RelPathError): pass

def relPathTo(Here,There):
	"""relPathTo(Here,There)
	Returns a relative path from one page to another. Both Here and There
	must be absolute URL's
	"""	
	herelist = urlparse.urlsplit(Here)
	therelist = urlparse.urlsplit(There)
	res = ''
	
	#print herelist
	#print therelist
	if not herelist[0]: raise NotFullPathError, "%s is not a full path" % Here
	if not therelist[0]: raise NotFullPathError, "%s is not a full path" % There
	
	
	#schemes are different, so return There as a full url
	if herelist[0].lower() != therelist[0].lower(): return urlparse.urlunsplit(therelist)
	
	if herelist[1].lower() != therelist[1].lower(): return urlparse.urlunsplit(therelist)
	else: ### same netlocations
		herepath = herelist[2].split('/')[1:]
		therepath = therelist[2].split('/')[1:]

		if herepath[0].startswith('~') and therepath[0].startswith('~'):
			if herepath[0].lower() != therepath[0].lower(): 
				return urlparse.urlunsplit(therelist)

		#remove all elements that are the same
		while herepath[0]==therepath[0]:
			d =herepath.pop(0)
			therepath.pop(0)
			if not herepath or not therepath: break

			
		if len(herepath) < len(therepath):
			res = '/'.join(therepath)
			#print res	
		else:
			gotolist=  ['..']*(len(herepath)-1)
			gotolist.extend(therepath)
			res = '/'.join(gotolist)
		
		
		if res and therelist[4]:
			res = '%s#%s' % (res,therelist[4])
		elif not res and therelist[4]:
			res = '#%s' % therelist[4]
		elif not res:
			res = '#'
		
		return res
	
	
def relPathFrom(Here,There):
	return relPathTo(There,here)


if __name__=='__main__':
	# Simple Test Cases. For a more rigorous test see relpathtest.py
	hpath = "http://www.s.com/~english/index.html"
	tpath = "http://www.s.com/~josh/index.html"
	ipath = "https://www.s.com/~english/index.html"
	ypath = "http://www.s.com/~english/index.html#blog"
	epath = "http://www.s.com/~english/education.html"
	mpath = "http://www.s.com/~english/math/index.html"
	zpath = "http://www.s.com/~english/math/zach.html"
	cpath = "http://www.s.com/~english/code.html"
	fpath = "http://www.s.com/~english/code/fraction.html"
	wpath = "http://www.s.com/~english/code/webmenu/webmenu.html"
	spath = "http://www.s.com/~english/code/wcms/wcms.html"
	npath = "http://www.c.com/index.html"
	opath = "http://www.c.com/java/magic.html"
	print '-'*24
	_debug = 0
	print relPathTo(hpath,tpath)