# -*- coding:  iso-8859-1 -*-
"""
xmppic.py:  an XMP Image parser.
date:       2006-07-12
Roberto Zoia (roberto@zoia.org)

Parses JPEG images with XMP metadata (More information
about XMP can be found at http://www.adobe.com/products/xmp/).

xmppic.py is a minor hack of the rdfpic.py (v1.1) plugin
for PyBlosxom.  (rdfpic.py was developed by Roberto De Almeida,
and parses JPEG images with metadata inserted using rdfpic, 
http://jigsaw.w3.org/rdfpic).

xmppic.py has been used to extract metadata inserted by PixVUE
(http://www.pixvue.com) in more than 4000 jpeg images with good
results.


Usage:
    import xmppic

    if xmppic.verify_installation():
        data = xmppic.parse("testimage.jpg")
        print data
    
"""


def verify_installation():
    # Try to import libxml2
    try:
        import libxml2
        
    except ImportError:
        print "I could not find the libxml2 module."
        print
        return 0

    # Is PIL installed?
    try:
        import Image

        # Let's check if it was patched.
        import JpegImagePlugin
        if not hasattr(JpegImagePlugin, 'COM'):
            print "You need to patch your Image (Python Imaging Library) module."
            print "Take a look at:"
            print 
            print "    http://groups.yahoo.com/group/python-list/message/110144"
            print
            return 0
        
    except ImportError:
        print "I could not find the Image (Python Imaging Library) module."
        print
        return 0

    return 1


import time
import os
import libxml2
import Image


def purgerdf(rdf):
    """
    Purga los caracteres marcianos del texto.
    A parserdf no le gustan los \x00, y es lo
    que está en la primera de la data de la imagen.
    """
    rdf.replace("\x00", "")  
    rdf0 = rdf[rdf.find('<?'):]
    return rdf0.strip()
    
def parserdf(rdf):
    xml = libxml2.parseDoc(rdf)
    
    ctxt = xml.xpathNewContext()
    ctxt.xpathRegisterNs('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#')
    ctxt.xpathRegisterNs('technical', 'http://www.w3.org/2000/PhotoRDF/technical-1-0#')
    ctxt.xpathRegisterNs('dc', 'http://www.w3.org/2000/PhotoRDF/dc-1-0#')
    ctxt.xpathRegisterNs('content', 'http://sophia.inria.fr/~enerbonn/rdfpiclang#')

    metadata = {}
    for item in ctxt.xpathEval('//rdf:Description/*'):
        # ¿quién me garantiza que los tags están en minúsculas?
        metadata[item.name.lower()] = item.content

    return metadata


def parse(filename):
    """
    Parses the image for XMP metadata.  Returns the data
    as a dictionary if data is found, None otherwise.
    """
    im = Image.open(filename)
    
    # Read metadata.
    try:
        rdf = im.app['APP1']   # APP1 as described by Adobe for XMP
    except (KeyError, AttributeError):
        return None   # no tags, return an empty dictionary
    
    rdf = purgerdf(rdf)
    if not rdf.isspace() and len(rdf)>1:
        metadata = parserdf(rdf)
        for k in metadata.keys():
            metadata[k] = metadata[k].replace("\n","").strip()
        return metadata
    else:
        return None