1 # -*- coding: utf-8 -*-
3 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
4 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
6 from xml.parsers.expat import ExpatError
7 from datetime import date
10 from librarian import (ValidationError, NoDublinCore, ParseError, DCNS, RDFNS,
13 import lxml.etree as etree # ElementTree API using libxml2
14 from lxml.etree import XMLSyntaxError
21 """Single person with last name and a list of first names."""
22 def __init__(self, last_name, *first_names):
23 self.last_name = last_name
24 self.first_names = first_names
27 def from_text(cls, text):
28 parts = [ token.strip() for token in text.split(',') ]
33 raise ValueError("Invalid person name. There should be at most one comma: \"%s\"." % text)
36 if len(parts[1]) == 0:
37 # there is no non-whitespace data after the comma
38 raise ValueError("Found a comma, but no names given: \"%s\" -> %r." % (text, parts))
39 names = [ name for name in parts[1].split() if len(name) ] # all non-whitespace tokens
40 return cls(surname, *names)
43 return u" ".join(self.first_names + (self.last_name,))
45 def __eq__(self, right):
46 return self.last_name == right.last_name and self.first_names == right.first_names
48 def __cmp__(self, other):
49 return cmp((self.last_name, self.first_names), (other.last_name, other.first_names))
52 return hash((self.last_name, self.first_names))
54 def __unicode__(self):
55 if len(self.first_names) > 0:
56 return '%s, %s' % (self.last_name, ' '.join(self.first_names))
61 return 'Person(last_name=%r, first_names=*%r)' % (self.last_name, self.first_names)
66 t = time.strptime(text, '%Y-%m-%d')
68 t = time.strptime(text, '%Y')
69 return date(t[0], t[1], t[2])
71 raise ValueError("Unrecognized date format. Try YYYY-MM-DD or YYYY.")
73 as_person = Person.from_text
76 if isinstance(text, unicode):
79 return text.decode('utf-8')
82 def __init__(self, uri, attr_name, validator=as_unicode, strict=None, multiple=False, salias=None, **kwargs):
85 self.validator = lambda x: validator(x)
86 self.strict = lambda x: strict(x)
87 self.multiple = multiple
90 self.required = kwargs.get('required', True) and not kwargs.has_key('default')
91 self.default = kwargs.get('default', [] if multiple else [None])
93 def validate_value(self, val, strict=False):
94 if strict and self.strict is not None:
95 validator = self.strict
97 validator = self.validator
100 if validator is None:
102 return [ validator(v) if v is not None else v for v in val ]
104 raise ValidationError("Multiple values not allowed for field '%s'" % self.uri)
106 raise ValidationError("Field %s has no value to assign. Check your defaults." % self.uri)
108 if validator is None or val[0] is None:
110 return validator(val[0])
111 except ValueError, e:
112 raise ValidationError("Field '%s' - invald value: %s" % (self.uri, e.message))
114 def validate(self, fdict, strict=False):
115 if not fdict.has_key(self.uri):
116 if not self.required:
119 raise ValidationError("Required field %s not found" % self.uri)
123 return self.validate_value(f, strict=strict)
125 def __eq__(self, other):
126 if isinstance(other, Field) and other.name == self.name:
132 def __new__(meta, classname, bases, class_dict):
133 fields = list(class_dict['FIELDS'])
135 for base in bases[::-1]:
136 if hasattr(base, 'FIELDS'):
137 for field in base.FIELDS[::-1]:
141 fields.insert(0, field)
143 class_dict['FIELDS'] = tuple(fields)
144 return super(DCInfo, meta).__new__(meta, classname, bases, class_dict)
147 class WorkInfo(object):
148 __metaclass__ = DCInfo
151 Field( DCNS('creator'), 'authors', as_person, salias='author', multiple=True),
152 Field( DCNS('title'), 'title'),
153 Field( DCNS('type'), 'type', required=False, multiple=True),
155 Field( DCNS('contributor.editor'), 'editors', \
156 as_person, salias='editor', multiple=True, default=[]),
157 Field( DCNS('contributor.technical_editor'), 'technical_editors',
158 as_person, salias='technical_editor', multiple=True, default=[]),
160 Field( DCNS('date'), 'created_at', as_date),
161 Field( DCNS('date.pd'), 'released_to_public_domain_at', as_date, required=False),
162 Field( DCNS('publisher'), 'publisher'),
164 Field( DCNS('language'), 'language'),
165 Field( DCNS('description'), 'description', required=False),
167 Field( DCNS('source'), 'source_name', required=False),
168 Field( DCNS('source.URL'), 'source_url', required=False),
169 Field( DCNS('identifier.url'), 'url', WLURI, strict=WLURI.strict),
170 Field( DCNS('rights.license'), 'license', required=False),
171 Field( DCNS('rights'), 'license_description'),
175 def from_string(cls, xml, *args, **kwargs):
176 from StringIO import StringIO
177 return cls.from_file(StringIO(xml), *args, **kwargs)
180 def from_file(cls, xmlfile, *args, **kwargs):
183 iter = etree.iterparse(xmlfile, ['start', 'end'])
184 for (event, element) in iter:
185 if element.tag == RDFNS('RDF') and event == 'start':
190 raise NoDublinCore("DublinCore section not found. \
191 Check if there are rdf:RDF and rdf:Description tags.")
193 # continue 'till the end of RDF section
194 for (event, element) in iter:
195 if element.tag == RDFNS('RDF') and event == 'end':
198 # if there is no end, Expat should yell at us with an ExpatError
200 # extract data from the element and make the info
201 return cls.from_element(desc_tag, *args, **kwargs)
202 except XMLSyntaxError, e:
204 except ExpatError, e:
208 def from_element(cls, rdf_tag, *args, **kwargs):
209 # the tree is already parsed, so we don't need to worry about Expat errors
211 desc = rdf_tag.find(".//" + RDFNS('Description'))
214 raise NoDublinCore("No DublinCore section found.")
216 for e in desc.getchildren():
217 fv = field_dict.get(e.tag, [])
219 field_dict[e.tag] = fv
221 return cls(desc.attrib, field_dict, *args, **kwargs)
223 def __init__(self, rdf_attrs, dc_fields, strict=False):
224 """rdf_attrs should be a dictionary-like object with any attributes of the RDF:Description.
225 dc_fields - dictionary mapping DC fields (with namespace) to list of text values for the
228 self.about = rdf_attrs.get(RDFNS('about'))
231 for field in self.FIELDS:
232 value = field.validate(dc_fields, strict=strict)
233 setattr(self, 'prop_' + field.name, value)
234 self.fmap[field.name] = field
235 if field.salias: self.fmap[field.salias] = field
237 def __getattribute__(self, name):
239 field = object.__getattribute__(self, 'fmap')[name]
240 value = object.__getattribute__(self, 'prop_'+field.name)
241 if field.name == name:
243 else: # singular alias
244 if not field.multiple:
245 raise "OUCH!! for field %s" % name
248 except (KeyError, AttributeError):
249 return object.__getattribute__(self, name)
251 def __setattr__(self, name, newvalue):
253 field = object.__getattribute__(self, 'fmap')[name]
254 if field.name == name:
255 object.__setattr__(self, 'prop_'+field.name, newvalue)
256 else: # singular alias
257 if not field.multiple:
258 raise "OUCH! while setting field %s" % name
260 object.__setattr__(self, 'prop_'+field.name, [newvalue])
261 except (KeyError, AttributeError):
262 return object.__setattr__(self, name, newvalue)
264 def update(self, field_dict):
265 """Update using field_dict. Verify correctness, but don't check if all
266 required fields are present."""
267 for field in self.FIELDS:
268 if field_dict.has_key(field.name):
269 setattr(self, field.name, field_dict[field.name])
271 def to_etree(self, parent = None):
272 """XML representation of this object."""
273 #etree._namespace_map[str(self.RDF)] = 'rdf'
274 #etree._namespace_map[str(self.DC)] = 'dc'
277 root = etree.Element(RDFNS('RDF'))
279 root = parent.makeelement(RDFNS('RDF'))
281 description = etree.SubElement(root, RDFNS('Description'))
284 description.set(RDFNS('about'), self.about)
286 for field in self.FIELDS:
287 v = getattr(self, field.name, None)
290 if len(v) == 0: continue
292 e = etree.Element(field.uri)
295 description.append(e)
297 e = etree.Element(field.uri)
299 description.append(e)
305 rdf['about'] = { 'uri': RDFNS('about'), 'value': self.about }
308 for field in self.FIELDS:
309 v = getattr(self, field.name, None)
312 if len(v) == 0: continue
313 v = [ unicode(x) for x in v if x is not None ]
317 dc[field.name] = {'uri': field.uri, 'value': v}
322 result = {'about': self.about}
323 for field in self.FIELDS:
324 v = getattr(self, field.name, None)
328 if len(v) == 0: continue
329 v = [ unicode(x) for x in v if x is not None ]
332 result[field.name] = v
335 v = getattr(self, field.salias)
336 if v is not None: result[field.salias] = unicode(v)
341 class BookInfo(WorkInfo):
343 Field( DCNS('audience'), 'audiences', salias='audience', multiple=True,
346 Field( DCNS('subject.period'), 'epochs', salias='epoch', multiple=True),
347 Field( DCNS('subject.type'), 'kinds', salias='kind', multiple=True),
348 Field( DCNS('subject.genre'), 'genres', salias='genre', multiple=True),
350 Field( DCNS('contributor.translator'), 'translators', \
351 as_person, salias='translator', multiple=True, default=[]),
352 Field( DCNS('relation.hasPart'), 'parts',
353 WLURI, strict=WLURI.strict, multiple=True, required=False),
354 Field( DCNS('relation.isVariantOf'), 'variant_of',
355 WLURI, strict=WLURI.strict, required=False),
357 Field( DCNS('relation.cover_image.url'), 'cover_url', required=False),
358 Field( DCNS('relation.cover_image.attribution'), 'cover_by', required=False),
359 Field( DCNS('relation.cover_image.source'), 'cover_source', required=False),
363 def parse(file_name, cls=BookInfo):
364 return cls.from_file(file_name)