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 __future__ import unicode_literals
8 from xml.parsers.expat import ExpatError
9 from datetime import date
10 from functools import total_ordering
14 from librarian.util import roman_to_int
16 from librarian import (ValidationError, NoDublinCore, ParseError, DCNS, RDFNS,
17 XMLNS, WLURI, WLNS, PLMETNS)
19 import lxml.etree as etree # ElementTree API using libxml2
20 from lxml.etree import XMLSyntaxError
23 class TextPlus(six.text_type):
34 @six.python_2_unicode_compatible
37 """Single person with last name and a list of first names."""
38 def __init__(self, last_name, *first_names):
39 self.last_name = last_name
40 self.first_names = first_names
43 def from_text(cls, text):
44 parts = [token.strip() for token in text.split(',')]
49 raise ValueError("Invalid person name. There should be at most one comma: \"%s\"." % text.encode('utf-8'))
52 if len(parts[1]) == 0:
53 # there is no non-whitespace data after the comma
54 raise ValueError("Found a comma, but no names given: \"%s\" -> %r." % (text, parts))
55 names = parts[1].split()
56 return cls(surname, *names)
59 return u" ".join(self.first_names + (self.last_name,))
61 def __eq__(self, right):
62 return self.last_name == right.last_name and self.first_names == right.first_names
64 def __lt__(self, other):
65 return (self.last_name, self.first_names) < (other.last_name, other.first_names)
68 return hash((self.last_name, self.first_names))
71 if len(self.first_names) > 0:
72 return '%s, %s' % (self.last_name, ' '.join(self.first_names))
77 return 'Person(last_name=%r, first_names=*%r)' % (self.last_name, self.first_names)
81 """Dates for digitization of pictures. It seems we need the following:
84 half centuries/decades: '2 poł. XVIII w.', 'XVII w., l. 20'
86 circa 'ok. 1813-1814', 'ok.1876-ok.1886
88 for now we will translate this to some single date losing information of course.
91 # check out the "N. poł X w." syntax
92 if isinstance(text, six.binary_type):
93 text = text.decode("utf-8")
95 century_format = u"(?:([12]) *poł[.]? +)?([MCDXVI]+) *w[.,]*(?: *l[.]? *([0-9]+))?"
96 vague_format = u"(?:po *|ok. *)?([0-9]{4})(-[0-9]{2}-[0-9]{2})?"
98 m = re.match(century_format, text)
99 m2 = re.match(vague_format, text)
103 century = roman_to_int(m.group(2))
105 if decade is not None:
106 raise ValueError("Bad date format. Cannot specify both half and decade of century")
108 t = ((century*100 + (half-1)*50), 1, 1)
110 decade = int(decade or 0)
111 t = ((century*100 + decade), 1, 1)
114 mon_day = m2.group(2)
116 t = time.strptime(year + mon_day, "%Y-%m-%d")
118 t = time.strptime(year, '%Y')
122 return DatePlus(t[0], t[1], t[2])
123 except ValueError as e:
124 raise ValueError("Unrecognized date format. Try YYYY-MM-DD or YYYY.")
128 return Person.from_text(text)
131 def as_unicode(text):
132 if isinstance(text, six.text_type):
135 return TextPlus(text.decode('utf-8'))
138 def as_wluri_strict(text):
139 return WLURI.strict(text)
143 def __init__(self, uri, attr_name, validator=as_unicode, strict=None, multiple=False, salias=None, **kwargs):
145 self.name = attr_name
146 self.validator = validator
148 self.multiple = multiple
151 self.required = kwargs.get('required', True) and 'default' not in kwargs
152 self.default = kwargs.get('default', [] if multiple else [None])
154 def validate_value(self, val, strict=False):
155 if strict and self.strict is not None:
156 validator = self.strict
158 validator = self.validator
161 if validator is None:
168 if hasattr(v, 'lang'):
169 setattr(nv, 'lang', v.lang)
170 new_values.append(nv)
173 raise ValidationError("Multiple values not allowed for field '%s'" % self.uri)
175 raise ValidationError("Field %s has no value to assign. Check your defaults." % self.uri)
177 if validator is None or val[0] is None:
179 nv = validator(val[0])
180 if hasattr(val[0], 'lang'):
181 setattr(nv, 'lang', val[0].lang)
183 except ValueError as e:
184 raise ValidationError("Field '%s' - invald value: %s" % (self.uri, e.message))
186 def validate(self, fdict, fallbacks=None, strict=False):
187 if fallbacks is None:
189 if self.uri not in fdict:
190 if not self.required:
191 # Accept single value for single fields and saliases.
192 if self.name in fallbacks:
194 f = fallbacks[self.name]
196 f = [fallbacks[self.name]]
197 elif self.salias and self.salias in fallbacks:
198 f = [fallbacks[self.salias]]
202 raise ValidationError("Required field %s not found" % self.uri)
206 return self.validate_value(f, strict=strict)
208 def __eq__(self, other):
209 if isinstance(other, Field) and other.name == self.name:
215 def __new__(mcs, classname, bases, class_dict):
216 fields = list(class_dict['FIELDS'])
218 for base in bases[::-1]:
219 if hasattr(base, 'FIELDS'):
220 for field in base.FIELDS[::-1]:
224 fields.insert(0, field)
226 class_dict['FIELDS'] = tuple(fields)
227 return super(DCInfo, mcs).__new__(mcs, classname, bases, class_dict)
230 class WorkInfo(six.with_metaclass(DCInfo, object)):
232 Field(DCNS('creator'), 'authors', as_person, salias='author', multiple=True),
233 Field(DCNS('title'), 'title'),
234 Field(DCNS('type'), 'type', required=False, multiple=True),
236 Field(DCNS('contributor.editor'), 'editors',
237 as_person, salias='editor', multiple=True, required=False),
238 Field(DCNS('contributor.technical_editor'), 'technical_editors',
239 as_person, salias='technical_editor', multiple=True, required=False),
240 Field(DCNS('contributor.funding'), 'funders', salias='funder', multiple=True, required=False),
241 Field(DCNS('contributor.thanks'), 'thanks', required=False),
243 Field(DCNS('date'), 'created_at'),
244 Field(DCNS('date.pd'), 'released_to_public_domain_at', as_date, required=False),
245 Field(DCNS('publisher'), 'publisher', multiple=True),
247 Field(DCNS('language'), 'language'),
248 Field(DCNS('description'), 'description', required=False),
250 Field(DCNS('source'), 'source_name', required=False),
251 Field(DCNS('source.URL'), 'source_urls', salias='source_url', multiple=True, required=False),
252 Field(DCNS('identifier.url'), 'url', WLURI, strict=as_wluri_strict),
253 Field(DCNS('rights.license'), 'license', required=False),
254 Field(DCNS('rights'), 'license_description'),
256 Field(PLMETNS('digitisationSponsor'), 'sponsors', multiple=True, required=False),
257 Field(WLNS('digitisationSponsorNote'), 'sponsor_note', required=False),
258 Field(WLNS('developmentStage'), 'stage', required=False),
262 def from_bytes(cls, xml, *args, **kwargs):
263 return cls.from_file(six.BytesIO(xml), *args, **kwargs)
266 def from_file(cls, xmlfile, *args, **kwargs):
269 iter = etree.iterparse(xmlfile, ['start', 'end'])
270 for (event, element) in iter:
271 if element.tag == RDFNS('RDF') and event == 'start':
276 raise NoDublinCore("DublinCore section not found. \
277 Check if there are rdf:RDF and rdf:Description tags.")
279 # continue 'till the end of RDF section
280 for (event, element) in iter:
281 if element.tag == RDFNS('RDF') and event == 'end':
284 # if there is no end, Expat should yell at us with an ExpatError
286 # extract data from the element and make the info
287 return cls.from_element(desc_tag, *args, **kwargs)
288 except XMLSyntaxError as e:
290 except ExpatError as e:
294 def from_element(cls, rdf_tag, *args, **kwargs):
295 # the tree is already parsed, so we don't need to worry about Expat errors
297 desc = rdf_tag.find(".//" + RDFNS('Description'))
300 raise NoDublinCore("No DublinCore section found.")
304 while p is not None and lang is None:
305 lang = p.attrib.get(XMLNS('lang'))
308 for e in desc.getchildren():
309 fv = field_dict.get(e.tag, [])
310 if e.text is not None:
312 if not isinstance(text, six.text_type):
313 text = text.decode('utf-8')
315 val.lang = e.attrib.get(XMLNS('lang'), lang)
317 meta_id = e.attrib.get('id')
318 if meta_id and meta_id.endswith('-id'):
319 field_dict[meta_id] = [val.replace('ISBN-', 'ISBN ')]
323 field_dict[e.tag] = fv
325 return cls(desc.attrib, field_dict, *args, **kwargs)
327 def __init__(self, rdf_attrs, dc_fields, fallbacks=None, strict=False):
328 """rdf_attrs should be a dictionary-like object with any attributes of the RDF:Description.
329 dc_fields - dictionary mapping DC fields (with namespace) to list of text values for the
332 self.about = rdf_attrs.get(RDFNS('about'))
335 for field in self.FIELDS:
336 value = field.validate(dc_fields, fallbacks=fallbacks, strict=strict)
337 setattr(self, 'prop_' + field.name, value)
338 self.fmap[field.name] = field
340 self.fmap[field.salias] = field
342 def __getattribute__(self, name):
344 field = object.__getattribute__(self, 'fmap')[name]
345 value = object.__getattribute__(self, 'prop_'+field.name)
346 if field.name == name:
348 else: # singular alias
349 if not field.multiple:
350 raise "OUCH!! for field %s" % name
352 return value[0] if value else None
353 except (KeyError, AttributeError):
354 return object.__getattribute__(self, name)
356 def __setattr__(self, name, newvalue):
358 field = object.__getattribute__(self, 'fmap')[name]
359 if field.name == name:
360 object.__setattr__(self, 'prop_'+field.name, newvalue)
361 else: # singular alias
362 if not field.multiple:
363 raise "OUCH! while setting field %s" % name
365 object.__setattr__(self, 'prop_'+field.name, [newvalue])
366 except (KeyError, AttributeError):
367 return object.__setattr__(self, name, newvalue)
369 def update(self, field_dict):
370 """Update using field_dict. Verify correctness, but don't check if all
371 required fields are present."""
372 for field in self.FIELDS:
373 if field.name in field_dict:
374 setattr(self, field.name, field_dict[field.name])
376 def to_etree(self, parent=None):
377 """XML representation of this object."""
378 # etree._namespace_map[str(self.RDF)] = 'rdf'
379 # etree._namespace_map[str(self.DC)] = 'dc'
382 root = etree.Element(RDFNS('RDF'))
384 root = parent.makeelement(RDFNS('RDF'))
386 description = etree.SubElement(root, RDFNS('Description'))
389 description.set(RDFNS('about'), self.about)
391 for field in self.FIELDS:
392 v = getattr(self, field.name, None)
398 e = etree.Element(field.uri)
400 e.text = six.text_type(x)
401 description.append(e)
403 e = etree.Element(field.uri)
404 e.text = six.text_type(v)
405 description.append(e)
410 rdf = {'about': {'uri': RDFNS('about'), 'value': self.about}}
413 for field in self.FIELDS:
414 v = getattr(self, field.name, None)
419 v = [six.text_type(x) for x in v if x is not None]
423 dc[field.name] = {'uri': field.uri, 'value': v}
428 result = {'about': self.about}
429 for field in self.FIELDS:
430 v = getattr(self, field.name, None)
436 v = [six.text_type(x) for x in v if x is not None]
439 result[field.name] = v
442 v = getattr(self, field.salias)
444 result[field.salias] = six.text_type(v)
449 class BookInfo(WorkInfo):
451 Field(DCNS('audience'), 'audiences', salias='audience', multiple=True, required=False),
453 Field(DCNS('subject.period'), 'epochs', salias='epoch', multiple=True, required=False),
454 Field(DCNS('subject.type'), 'kinds', salias='kind', multiple=True, required=False),
455 Field(DCNS('subject.genre'), 'genres', salias='genre', multiple=True, required=False),
457 Field(DCNS('subject.location'), 'location', required=False),
459 Field(DCNS('contributor.translator'), 'translators',
460 as_person, salias='translator', multiple=True, required=False),
461 Field(DCNS('relation.hasPart'), 'parts', WLURI, strict=as_wluri_strict, multiple=True, required=False),
462 Field(DCNS('relation.isVariantOf'), 'variant_of', WLURI, strict=as_wluri_strict, required=False),
464 Field(DCNS('relation.coverImage.url'), 'cover_url', required=False),
465 Field(DCNS('relation.coverImage.attribution'), 'cover_by', required=False),
466 Field(DCNS('relation.coverImage.source'), 'cover_source', required=False),
468 Field(WLNS('coverBarColor'), 'cover_bar_color', required=False),
469 Field(WLNS('coverBoxPosition'), 'cover_box_position', required=False),
470 Field(WLNS('coverClass'), 'cover_class', default=['default']),
471 Field(WLNS('coverLogoUrl'), 'cover_logo_urls', multiple=True, required=False),
473 Field('pdf-id', 'isbn_pdf', required=False),
474 Field('epub-id', 'isbn_epub', required=False),
475 Field('mobi-id', 'isbn_mobi', required=False),
476 Field('txt-id', 'isbn_txt', required=False),
477 Field('html-id', 'isbn_html', required=False),
481 def parse(file_name, cls=BookInfo):
482 return cls.from_file(file_name)