librarian/picture.py

   1
   2 from dcparser import (as_person, as_date, Field, WorkInfo, DCNS)
   3 from librarian import (RDFNS, ValidationError, NoDublinCore, ParseError, WLURI)
   4 from xml.parsers.expat import ExpatError
   5 from os import path
   6 from StringIO import StringIO
   7 from lxml import etree
   8 from lxml.etree import (XMLSyntaxError, XSLTApplyError)
   9 import re
  10
  11
  12 class WLPictureURI(WLURI):
  13     _re_wl_uri = re.compile('http://wolnelektury.pl/katalog/obraz/'
  14             '(?P<slug>[-a-z0-9]+)(/(?P<lang>[a-z]{3}))?/?$')
  15
  16     def __init__(self, *args, **kw):
  17         super(WLPictureURI, self).__init__(*args, **kw)
  18
  19     @classmethod
  20     def from_slug_and_lang(cls, slug, lang):
  21         uri = 'http://wolnelektury.pl/katalog/obraz/%s/' % slug
  22         return cls(uri)
  23
  24     def filename_stem(self):
  25         return self.slug
  26
  27
  28 class PictureInfo(WorkInfo):
  29     """
  30     Dublin core metadata for a picture
  31     """
  32     FIELDS = (
  33         Field(DCNS('language'), 'language', required=False),
  34         Field(DCNS('subject.period'), 'epochs', salias='epoch', multiple=True),
  35         Field(DCNS('subject.type'), 'kinds', salias='kind', multiple=True),
  36
  37         Field(DCNS('format.dimensions'), 'dimensions', required=False),
  38         Field(DCNS('description.medium'), 'medium', required=False),
  39         Field(DCNS('description.dimensions'), 'original_dimensions', required=False),
  40         Field(DCNS('format'), 'mime_type', required=False),
  41         Field(DCNS('identifier.url'), 'url', WLPictureURI),
  42         )
  43
  44     def validate(self):
  45         """
  46         WorkInfo has a language validation code only, which we do not need.
  47         """
  48         pass
  49
  50
  51 class ImageStore(object):
  52     EXT = ['gif', 'jpeg', 'png', 'swf', 'psd', 'bmp'
  53             'tiff', 'tiff', 'jpc', 'jp2', 'jpf', 'jb2', 'swc',
  54             'aiff', 'wbmp', 'xbm']
  55     MIME = ['image/gif', 'image/jpeg', 'image/png',
  56             'application/x-shockwave-flash', 'image/psd', 'image/bmp',
  57             'image/tiff', 'image/tiff', 'application/octet-stream',
  58             'image/jp2', 'application/octet-stream', 'application/octet-stream',
  59             'application/x-shockwave-flash', 'image/iff', 'image/vnd.wap.wbmp', 'image/xbm']
  60
  61     def __init__(self, dir_):
  62         self.dir = dir_
  63         return super(ImageStore, self).__init__()
  64
  65     def path(self, slug, mime_type):
  66         """
  67         Finds file by slug and mime type in our iamge store.
  68         Returns a file objects (perhaps should return a filename?)
  69         """
  70         try:
  71             i = self.MIME.index(mime_type)
  72         except ValueError:
  73             err = ValueError("Picture %s has unknown mime type: %s" % (slug, mime_type))
  74             err.slug = slug
  75             err.mime_type = mime_type
  76             raise err
  77         ext = self.EXT[i]
  78         # add some common extensions tiff->tif, jpeg->jpg
  79         return path.join(self.dir, slug + '.' + ext)
  80
  81
  82 class WLPicture(object):
  83     def __init__(self, edoc, parse_dublincore=True, image_store=None):
  84         self.edoc = edoc
  85         self.image_store = image_store
  86
  87         root_elem = edoc.getroot()
  88
  89         dc_path = './/' + RDFNS('RDF')
  90
  91         if root_elem.tag != 'picture':
  92             raise ValidationError("Invalid root element. Found '%s', should be 'picture'" % root_elem.tag)
  93
  94         if parse_dublincore:
  95             self.rdf_elem = root_elem.find(dc_path)
  96
  97             if self.rdf_elem is None:
  98                 raise NoDublinCore('Document has no DublinCore - which is required.')
  99
 100             self.picture_info = PictureInfo.from_element(self.rdf_elem)
 101         else:
 102             self.picture_info = None
 103
 104     @classmethod
 105     def from_string(cls, xml, *args, **kwargs):
 106         return cls.from_file(StringIO(xml), *args, **kwargs)
 107
 108     @classmethod
 109     def from_file(cls, xmlfile, parse_dublincore=True, image_store=None):
 110
 111         # first, prepare for parsing
 112         if isinstance(xmlfile, basestring):
 113             file = open(xmlfile, 'rb')
 114             try:
 115                 data = file.read()
 116             finally:
 117                 file.close()
 118         else:
 119             data = xmlfile.read()
 120
 121         if not isinstance(data, unicode):
 122             data = data.decode('utf-8')
 123
 124         data = data.replace(u'\ufeff', '')
 125
 126         # assume images are in the same directory
 127         if image_store is None and xmlfile.name is not None:
 128             image_store = ImageStore(path.dirname(xmlfile.name))
 129
 130         try:
 131             parser = etree.XMLParser(remove_blank_text=False)
 132             tree = etree.parse(StringIO(data.encode('utf-8')), parser)
 133
 134             return cls(tree, parse_dublincore=parse_dublincore, image_store=image_store)
 135         except (ExpatError, XMLSyntaxError, XSLTApplyError), e:
 136             raise ParseError(e)
 137
 138     @property
 139     def mime_type(self):
 140         if self.picture_info is None:
 141             raise ValueError('DC is not loaded, hence we don\'t know the image type')
 142         return self.picture_info.mime_type
 143
 144     @property
 145     def slug(self):
 146         return self.picture_info.url.slug
 147
 148     @property
 149     def image_path(self):
 150         if self.image_store is None:
 151             raise ValueError("No image store associated with whis WLPicture.")
 152         return self.image_store.path(self.slug, self.mime_type)
 153
 154     def image_file(self, *args, **kwargs):
 155         return open(self.image_path, *args, **kwargs)
 156
 157     def partiter(self):
 158         """
 159         Iterates the parts of this picture and returns them and their metadata
 160         """
 161         for part in self.edoc.iter("div"):
 162             pd = {}
 163             pd['type'] = part.get('type')
 164             if pd['type'] == 'area':
 165                 pd['coords'] = ((int(part.get('x1')), int(part.get('y1'))),
 166                                 (int(part.get('x2')), int(part.get('y2'))))
 167
 168             pd['themes'] = []
 169             pd['object'] = None
 170             parent = part
 171             while True:
 172                 parent = parent.getparent()
 173                 if parent is None:
 174                     break
 175                 if parent.tag == 'sem':
 176                     if parent.get('type') == 'theme':
 177                         pd['themes'] += map(unicode.strip, unicode(parent.get('theme')).split(','))
 178                     elif parent.get('type') == 'object' and not pd['object']:
 179                         pd['object'] = parent.get('name')
 180             yield pd