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]+)/?$')
  15
  16     @classmethod
  17     def from_slug(cls, slug):
  18         uri = 'http://wolnelektury.pl/katalog/obraz/%s/' % slug
  19         return cls(uri)
  20
  21 def as_wlpictureuri_strict(text):
  22     return WLPictureURI.strict(text)
  23
  24
  25 class PictureInfo(WorkInfo):
  26     """
  27     Dublin core metadata for a picture
  28     """
  29     FIELDS = (
  30         Field(DCNS('language'), 'language', required=False),
  31         Field(DCNS('subject.period'), 'epochs', salias='epoch', multiple=True),
  32         Field(DCNS('subject.type'), 'kinds', salias='kind', multiple=True),
  33         Field(DCNS('subject.genre'), 'genres', salias='genre', multiple=True, required=False),
  34
  35         Field(DCNS('format.dimensions'), 'dimensions', required=False),
  36         Field(DCNS('format.checksum.sha1'), 'sha1', required=True),
  37         Field(DCNS('description.medium'), 'medium', required=False),
  38         Field(DCNS('description.dimensions'), 'original_dimensions', required=False),
  39         Field(DCNS('format'), 'mime_type', required=False),
  40         Field(DCNS('identifier.url'), 'url', WLPictureURI,
  41             strict=as_wlpictureuri_strict),
  42         )
  43
  44
  45 class ImageStore(object):
  46     EXT = ['gif', 'jpeg', 'png', 'swf', 'psd', 'bmp'
  47             'tiff', 'tiff', 'jpc', 'jp2', 'jpf', 'jb2', 'swc',
  48             'aiff', 'wbmp', 'xbm']
  49     MIME = ['image/gif', 'image/jpeg', 'image/png',
  50             'application/x-shockwave-flash', 'image/psd', 'image/bmp',
  51             'image/tiff', 'image/tiff', 'application/octet-stream',
  52             'image/jp2', 'application/octet-stream', 'application/octet-stream',
  53             'application/x-shockwave-flash', 'image/iff', 'image/vnd.wap.wbmp', 'image/xbm']
  54
  55     def __init__(self, dir_):
  56         self.dir = dir_
  57         return super(ImageStore, self).__init__()
  58
  59     def path(self, slug, mime_type):
  60         """
  61         Finds file by slug and mime type in our iamge store.
  62         Returns a file objects (perhaps should return a filename?)
  63         """
  64         try:
  65             i = self.MIME.index(mime_type)
  66         except ValueError:
  67             err = ValueError("Picture %s has unknown mime type: %s" % (slug, mime_type))
  68             err.slug = slug
  69             err.mime_type = mime_type
  70             raise err
  71         ext = self.EXT[i]
  72         # add some common extensions tiff->tif, jpeg->jpg
  73         return path.join(self.dir, slug + '.' + ext)
  74
  75
  76 class WLPicture(object):
  77     def __init__(self, edoc, parse_dublincore=True, image_store=None):
  78         self.edoc = edoc
  79         self.image_store = image_store
  80
  81         root_elem = edoc.getroot()
  82
  83         dc_path = './/' + RDFNS('RDF')
  84
  85         if root_elem.tag != 'picture':
  86             raise ValidationError("Invalid root element. Found '%s', should be 'picture'" % root_elem.tag)
  87
  88         if parse_dublincore:
  89             self.rdf_elem = root_elem.find(dc_path)
  90
  91             if self.rdf_elem is None:
  92                 raise NoDublinCore('Document has no DublinCore - which is required.')
  93
  94             self.picture_info = PictureInfo.from_element(self.rdf_elem)
  95         else:
  96             self.picture_info = None
  97
  98     @classmethod
  99     def from_string(cls, xml, *args, **kwargs):
 100         return cls.from_file(StringIO(xml), *args, **kwargs)
 101
 102     @classmethod
 103     def from_file(cls, xmlfile, parse_dublincore=True, image_store=None):
 104
 105         # first, prepare for parsing
 106         if isinstance(xmlfile, basestring):
 107             file = open(xmlfile, 'rb')
 108             try:
 109                 data = file.read()
 110             finally:
 111                 file.close()
 112         else:
 113             data = xmlfile.read()
 114
 115         if not isinstance(data, unicode):
 116             data = data.decode('utf-8')
 117
 118         data = data.replace(u'\ufeff', '')
 119
 120         # assume images are in the same directory
 121         if image_store is None and xmlfile.name is not None:
 122             image_store = ImageStore(path.dirname(xmlfile.name))
 123
 124         try:
 125             parser = etree.XMLParser(remove_blank_text=False)
 126             tree = etree.parse(StringIO(data.encode('utf-8')), parser)
 127
 128             return cls(tree, parse_dublincore=parse_dublincore, image_store=image_store)
 129         except (ExpatError, XMLSyntaxError, XSLTApplyError), e:
 130             raise ParseError(e)
 131
 132     @property
 133     def mime_type(self):
 134         if self.picture_info is None:
 135             raise ValueError('DC is not loaded, hence we don\'t know the image type')
 136         return self.picture_info.mime_type
 137
 138     @property
 139     def slug(self):
 140         return self.picture_info.url.slug
 141
 142     @property
 143     def image_path(self):
 144         if self.image_store is None:
 145             raise ValueError("No image store associated with whis WLPicture.")
 146         return self.image_store.path(self.slug, self.mime_type)
 147
 148     def image_file(self, *args, **kwargs):
 149         return open(self.image_path, *args, **kwargs)
 150
 151     def partiter(self):
 152         """
 153         Iterates the parts of this picture and returns them and their metadata
 154         """
 155         for part in self.edoc.iter("div"):
 156             pd = {}
 157             pd['type'] = part.get('type')
 158             if pd['type'] == 'area':
 159                 pd['coords'] = ((int(part.get('x1')), int(part.get('y1'))),
 160                                 (int(part.get('x2')), int(part.get('y2'))))
 161
 162             pd['themes'] = []
 163             pd['object'] = None
 164             parent = part
 165             while True:
 166                 parent = parent.getparent()
 167                 if parent is None:
 168                     break
 169                 if parent.tag == 'sem':
 170                     if parent.get('type') == 'theme':
 171                         pd['themes'] += map(unicode.strip, unicode(parent.get('theme')).split(','))
 172                     elif parent.get('type') == 'object' and pd['object'] is None:
 173                         pd['object'] = parent.get('object')
 174             yield pd