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('format.checksum.sha1'), 'sha1', required=True),
  39         Field(DCNS('description.medium'), 'medium', required=False),
  40         Field(DCNS('description.dimensions'), 'original_dimensions', required=False),
  41         Field(DCNS('format'), 'mime_type', required=False),
  42         Field(DCNS('identifier.url'), 'url', WLPictureURI),
  43         )
  44
  45     def validate(self):
  46         """
  47         WorkInfo has a language validation code only, which we do not need.
  48         """
  49         pass
  50
  51
  52 class ImageStore(object):
  53     EXT = ['gif', 'jpeg', 'png', 'swf', 'psd', 'bmp'
  54             'tiff', 'tiff', 'jpc', 'jp2', 'jpf', 'jb2', 'swc',
  55             'aiff', 'wbmp', 'xbm']
  56     MIME = ['image/gif', 'image/jpeg', 'image/png',
  57             'application/x-shockwave-flash', 'image/psd', 'image/bmp',
  58             'image/tiff', 'image/tiff', 'application/octet-stream',
  59             'image/jp2', 'application/octet-stream', 'application/octet-stream',
  60             'application/x-shockwave-flash', 'image/iff', 'image/vnd.wap.wbmp', 'image/xbm']
  61
  62     def __init__(self, dir_):
  63         self.dir = dir_
  64         return super(ImageStore, self).__init__()
  65
  66     def path(self, slug, mime_type):
  67         """
  68         Finds file by slug and mime type in our iamge store.
  69         Returns a file objects (perhaps should return a filename?)
  70         """
  71         try:
  72             i = self.MIME.index(mime_type)
  73         except ValueError:
  74             err = ValueError("Picture %s has unknown mime type: %s" % (slug, mime_type))
  75             err.slug = slug
  76             err.mime_type = mime_type
  77             raise err
  78         ext = self.EXT[i]
  79         # add some common extensions tiff->tif, jpeg->jpg
  80         return path.join(self.dir, slug + '.' + ext)
  81
  82
  83 class WLPicture(object):
  84     def __init__(self, edoc, parse_dublincore=True, image_store=None):
  85         self.edoc = edoc
  86         self.image_store = image_store
  87
  88         root_elem = edoc.getroot()
  89
  90         dc_path = './/' + RDFNS('RDF')
  91
  92         if root_elem.tag != 'picture':
  93             raise ValidationError("Invalid root element. Found '%s', should be 'picture'" % root_elem.tag)
  94
  95         if parse_dublincore:
  96             self.rdf_elem = root_elem.find(dc_path)
  97
  98             if self.rdf_elem is None:
  99                 raise NoDublinCore('Document has no DublinCore - which is required.')
 100
 101             self.picture_info = PictureInfo.from_element(self.rdf_elem)
 102         else:
 103             self.picture_info = None
 104
 105     @classmethod
 106     def from_string(cls, xml, *args, **kwargs):
 107         return cls.from_file(StringIO(xml), *args, **kwargs)
 108
 109     @classmethod
 110     def from_file(cls, xmlfile, parse_dublincore=True, image_store=None):
 111
 112         # first, prepare for parsing
 113         if isinstance(xmlfile, basestring):
 114             file = open(xmlfile, 'rb')
 115             try:
 116                 data = file.read()
 117             finally:
 118                 file.close()
 119         else:
 120             data = xmlfile.read()
 121
 122         if not isinstance(data, unicode):
 123             data = data.decode('utf-8')
 124
 125         data = data.replace(u'\ufeff', '')
 126
 127         # assume images are in the same directory
 128         if image_store is None and xmlfile.name is not None:
 129             image_store = ImageStore(path.dirname(xmlfile.name))
 130
 131         try:
 132             parser = etree.XMLParser(remove_blank_text=False)
 133             tree = etree.parse(StringIO(data.encode('utf-8')), parser)
 134
 135             return cls(tree, parse_dublincore=parse_dublincore, image_store=image_store)
 136         except (ExpatError, XMLSyntaxError, XSLTApplyError), e:
 137             raise ParseError(e)
 138
 139     @property
 140     def mime_type(self):
 141         if self.picture_info is None:
 142             raise ValueError('DC is not loaded, hence we don\'t know the image type')
 143         return self.picture_info.mime_type
 144
 145     @property
 146     def slug(self):
 147         return self.picture_info.url.slug
 148
 149     @property
 150     def image_path(self):
 151         if self.image_store is None:
 152             raise ValueError("No image store associated with whis WLPicture.")
 153         return self.image_store.path(self.slug, self.mime_type)
 154
 155     def image_file(self, *args, **kwargs):
 156         return open(self.image_path, *args, **kwargs)
 157
 158     def partiter(self):
 159         """
 160         Iterates the parts of this picture and returns them and their metadata
 161         """
 162         for part in self.edoc.iter("div"):
 163             pd = {}
 164             pd['type'] = part.get('type')
 165             if pd['type'] == 'area':
 166                 pd['coords'] = ((int(part.get('x1')), int(part.get('y1'))),
 167                                 (int(part.get('x2')), int(part.get('y2'))))
 168
 169             pd['themes'] = []
 170             pd['object'] = None
 171             parent = part
 172             while True:
 173                 parent = parent.getparent()
 174                 if parent is None:
 175                     break
 176                 if parent.tag == 'sem':
 177                     if parent.get('type') == 'theme':
 178                         pd['themes'] += map(unicode.strip, unicode(parent.get('theme')).split(','))
 179                     elif parent.get('type') == 'object' and not pd['object']:
 180                         pd['object'] = parent.get('name')
 181             yield pd