X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/3df78775bc9e2e4d861bc0eaa4deadd354dda518..a2b90373ee7aa46769fb6faad75bac1ebef3758f:/librarian/dcparser.py diff --git a/librarian/dcparser.py b/librarian/dcparser.py index 12bb24f..92afc01 100644 --- a/librarian/dcparser.py +++ b/librarian/dcparser.py @@ -3,22 +3,27 @@ # This file is part of Librarian, licensed under GNU Affero GPLv3 or later. # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information. # +from __future__ import unicode_literals + from xml.parsers.expat import ExpatError from datetime import date +from functools import total_ordering import time import re +import six from librarian.util import roman_to_int from librarian import (ValidationError, NoDublinCore, ParseError, DCNS, RDFNS, XMLNS, WLURI, WLNS, PLMETNS) -import lxml.etree as etree # ElementTree API using libxml2 +import lxml.etree as etree # ElementTree API using libxml2 from lxml.etree import XMLSyntaxError -class TextPlus(unicode): +class TextPlus(six.text_type): pass + class DatePlus(date): pass @@ -26,6 +31,8 @@ class DatePlus(date): # ============== # = Converters = # ============== +@six.python_2_unicode_compatible +@total_ordering class Person(object): """Single person with last name and a list of first names.""" def __init__(self, last_name, *first_names): @@ -34,18 +41,18 @@ class Person(object): @classmethod def from_text(cls, text): - parts = [ token.strip() for token in text.split(',') ] + parts = [token.strip() for token in text.split(',')] if len(parts) == 1: surname = parts[0] names = [] elif len(parts) != 2: - raise ValueError("Invalid person name. There should be at most one comma: \"%s\"." % text) + raise ValueError("Invalid person name. There should be at most one comma: \"%s\"." % text.encode('utf-8')) else: surname = parts[0] if len(parts[1]) == 0: # there is no non-whitespace data after the comma raise ValueError("Found a comma, but no names given: \"%s\" -> %r." % (text, parts)) - names = [ name for name in parts[1].split() if len(name) ] # all non-whitespace tokens + names = parts[1].split() return cls(surname, *names) def readable(self): @@ -54,13 +61,13 @@ class Person(object): def __eq__(self, right): return self.last_name == right.last_name and self.first_names == right.first_names - def __cmp__(self, other): - return cmp((self.last_name, self.first_names), (other.last_name, other.first_names)) + def __lt__(self, other): + return (self.last_name, self.first_names) < (other.last_name, other.first_names) def __hash__(self): return hash((self.last_name, self.first_names)) - def __unicode__(self): + def __str__(self): if len(self.first_names) > 0: return '%s, %s' % (self.last_name, ' '.join(self.first_names)) else: @@ -69,6 +76,7 @@ class Person(object): def __repr__(self): return 'Person(last_name=%r, first_names=*%r)' % (self.last_name, self.first_names) + def as_date(text): """Dates for digitization of pictures. It seems we need the following: ranges: '1350-1450', @@ -81,7 +89,8 @@ for now we will translate this to some single date losing information of course. """ try: # check out the "N. poł X w." syntax - if isinstance(text, str): text = text.decode("utf-8") + if isinstance(text, six.binary_type): + text = text.decode("utf-8") century_format = u"(?:([12]) *poł[.]? +)?([MCDXVI]+) *w[.,]*(?: *l[.]? *([0-9]+))?" vague_format = u"(?:po *|ok. *)?([0-9]{4})(-[0-9]{2}-[0-9]{2})?" @@ -91,7 +100,7 @@ for now we will translate this to some single date losing information of course. if m: half = m.group(1) decade = m.group(3) - century = roman_to_int(str(m.group(2))) + century = roman_to_int(m.group(2)) if half is not None: if decade is not None: raise ValueError("Bad date format. Cannot specify both half and decade of century") @@ -111,21 +120,25 @@ for now we will translate this to some single date losing information of course. raise ValueError return DatePlus(t[0], t[1], t[2]) - except ValueError, e: + except ValueError as e: raise ValueError("Unrecognized date format. Try YYYY-MM-DD or YYYY.") + def as_person(text): return Person.from_text(text) + def as_unicode(text): - if isinstance(text, unicode): + if isinstance(text, six.text_type): return text else: return TextPlus(text.decode('utf-8')) + def as_wluri_strict(text): return WLURI.strict(text) + class Field(object): def __init__(self, uri, attr_name, validator=as_unicode, strict=None, multiple=False, salias=None, **kwargs): self.uri = uri @@ -135,7 +148,7 @@ class Field(object): self.multiple = multiple self.salias = salias - self.required = kwargs.get('required', True) and not kwargs.has_key('default') + self.required = kwargs.get('required', True) and 'default' not in kwargs self.default = kwargs.get('default', [] if multiple else [None]) def validate_value(self, val, strict=False): @@ -167,13 +180,13 @@ class Field(object): if hasattr(val[0], 'lang'): setattr(nv, 'lang', val[0].lang) return nv - except ValueError, e: + except ValueError as e: raise ValidationError("Field '%s' - invald value: %s" % (self.uri, e.message)) def validate(self, fdict, fallbacks=None, strict=False): if fallbacks is None: fallbacks = {} - if not fdict.has_key(self.uri): + if self.uri not in fdict: if not self.required: # Accept single value for single fields and saliases. if self.name in fallbacks: @@ -199,7 +212,7 @@ class Field(object): class DCInfo(type): - def __new__(meta, classname, bases, class_dict): + def __new__(mcs, classname, bases, class_dict): fields = list(class_dict['FIELDS']) for base in bases[::-1]: @@ -211,46 +224,43 @@ class DCInfo(type): fields.insert(0, field) class_dict['FIELDS'] = tuple(fields) - return super(DCInfo, meta).__new__(meta, classname, bases, class_dict) + return super(DCInfo, mcs).__new__(mcs, classname, bases, class_dict) -class WorkInfo(object): - __metaclass__ = DCInfo - +class WorkInfo(six.with_metaclass(DCInfo, object)): FIELDS = ( - Field( DCNS('creator'), 'authors', as_person, salias='author', multiple=True), - Field( DCNS('title'), 'title'), - Field( DCNS('type'), 'type', required=False, multiple=True), - - Field( DCNS('contributor.editor'), 'editors', \ - as_person, salias='editor', multiple=True, default=[]), - Field( DCNS('contributor.technical_editor'), 'technical_editors', - as_person, salias='technical_editor', multiple=True, default=[]), - Field( DCNS('contributor.funding'), 'funders', - salias='funder', multiple=True, default=[]), - Field( DCNS('contributor.thanks'), 'thanks', required=False), - - Field( DCNS('date'), 'created_at', as_date), - Field( DCNS('date.pd'), 'released_to_public_domain_at', as_date, required=False), - Field( DCNS('publisher'), 'publisher'), - - Field( DCNS('language'), 'language'), - Field( DCNS('description'), 'description', required=False), - - Field( DCNS('source'), 'source_name', required=False), - Field( DCNS('source.URL'), 'source_url', required=False), - Field( DCNS('identifier.url'), 'url', WLURI, strict=as_wluri_strict), - Field( DCNS('rights.license'), 'license', required=False), - Field( DCNS('rights'), 'license_description'), - - Field( PLMETNS('digitisationSponsor'), 'sponsors', multiple=True, default=[]), - Field( WLNS('digitisationSponsorNote'), 'sponsor_note', required=False), + Field(DCNS('creator'), 'authors', as_person, salias='author', multiple=True), + Field(DCNS('title'), 'title'), + Field(DCNS('type'), 'type', required=False, multiple=True), + + Field(DCNS('contributor.editor'), 'editors', + as_person, salias='editor', multiple=True, required=False), + Field(DCNS('contributor.technical_editor'), 'technical_editors', + as_person, salias='technical_editor', multiple=True, required=False), + Field(DCNS('contributor.funding'), 'funders', salias='funder', multiple=True, required=False), + Field(DCNS('contributor.thanks'), 'thanks', required=False), + + Field(DCNS('date'), 'created_at'), + Field(DCNS('date.pd'), 'released_to_public_domain_at', as_date, required=False), + Field(DCNS('publisher'), 'publisher', multiple=True), + + Field(DCNS('language'), 'language'), + Field(DCNS('description'), 'description', required=False), + + Field(DCNS('source'), 'source_name', required=False), + Field(DCNS('source.URL'), 'source_urls', salias='source_url', multiple=True, required=False), + Field(DCNS('identifier.url'), 'url', WLURI, strict=as_wluri_strict), + Field(DCNS('rights.license'), 'license', required=False), + Field(DCNS('rights'), 'license_description'), + + Field(PLMETNS('digitisationSponsor'), 'sponsors', multiple=True, required=False), + Field(WLNS('digitisationSponsorNote'), 'sponsor_note', required=False), + Field(WLNS('developmentStage'), 'stage', required=False), ) @classmethod - def from_string(cls, xml, *args, **kwargs): - from StringIO import StringIO - return cls.from_file(StringIO(xml), *args, **kwargs) + def from_bytes(cls, xml, *args, **kwargs): + return cls.from_file(six.BytesIO(xml), *args, **kwargs) @classmethod def from_file(cls, xmlfile, *args, **kwargs): @@ -275,9 +285,9 @@ class WorkInfo(object): # extract data from the element and make the info return cls.from_element(desc_tag, *args, **kwargs) - except XMLSyntaxError, e: + except XMLSyntaxError as e: raise ParseError(e) - except ExpatError, e: + except ExpatError as e: raise ParseError(e) @classmethod @@ -287,7 +297,7 @@ class WorkInfo(object): desc = rdf_tag.find(".//" + RDFNS('Description')) if desc is None: - raise NoDublinCore("No DublinCore section found.") + raise NoDublinCore("There must be a '%s' element inside the RDF." % RDFNS('Description')) lang = None p = desc @@ -299,10 +309,14 @@ class WorkInfo(object): fv = field_dict.get(e.tag, []) if e.text is not None: text = e.text - if not isinstance(text, unicode): + if not isinstance(text, six.text_type): text = text.decode('utf-8') val = TextPlus(text) val.lang = e.attrib.get(XMLNS('lang'), lang) + if e.tag == 'meta': + meta_id = e.attrib.get('id') + if meta_id and meta_id.endswith('-id'): + field_dict[meta_id] = [val.replace('ISBN-', 'ISBN ')] else: val = e.text fv.append(val) @@ -319,11 +333,11 @@ class WorkInfo(object): self.fmap = {} for field in self.FIELDS: - value = field.validate(dc_fields, fallbacks=fallbacks, - strict=strict) + value = field.validate(dc_fields, fallbacks=fallbacks, strict=strict) setattr(self, 'prop_' + field.name, value) self.fmap[field.name] = field - if field.salias: self.fmap[field.salias] = field + if field.salias: + self.fmap[field.salias] = field def __getattribute__(self, name): try: @@ -331,7 +345,7 @@ class WorkInfo(object): value = object.__getattribute__(self, 'prop_'+field.name) if field.name == name: return value - else: # singular alias + else: # singular alias if not field.multiple: raise "OUCH!! for field %s" % name @@ -344,7 +358,7 @@ class WorkInfo(object): field = object.__getattribute__(self, 'fmap')[name] if field.name == name: object.__setattr__(self, 'prop_'+field.name, newvalue) - else: # singular alias + else: # singular alias if not field.multiple: raise "OUCH! while setting field %s" % name @@ -356,13 +370,13 @@ class WorkInfo(object): """Update using field_dict. Verify correctness, but don't check if all required fields are present.""" for field in self.FIELDS: - if field_dict.has_key(field.name): + if field.name in field_dict: setattr(self, field.name, field_dict[field.name]) - def to_etree(self, parent = None): + def to_etree(self, parent=None): """XML representation of this object.""" - #etree._namespace_map[str(self.RDF)] = 'rdf' - #etree._namespace_map[str(self.DC)] = 'dc' + # etree._namespace_map[str(self.RDF)] = 'rdf' + # etree._namespace_map[str(self.DC)] = 'dc' if parent is None: root = etree.Element(RDFNS('RDF')) @@ -378,32 +392,33 @@ class WorkInfo(object): v = getattr(self, field.name, None) if v is not None: if field.multiple: - if len(v) == 0: continue + if len(v) == 0: + continue for x in v: e = etree.Element(field.uri) if x is not None: - e.text = unicode(x) + e.text = six.text_type(x) description.append(e) else: e = etree.Element(field.uri) - e.text = unicode(v) + e.text = six.text_type(v) description.append(e) return root def serialize(self): - rdf = {} - rdf['about'] = { 'uri': RDFNS('about'), 'value': self.about } + rdf = {'about': {'uri': RDFNS('about'), 'value': self.about}} dc = {} for field in self.FIELDS: v = getattr(self, field.name, None) if v is not None: if field.multiple: - if len(v) == 0: continue - v = [ unicode(x) for x in v if x is not None ] + if len(v) == 0: + continue + v = [six.text_type(x) for x in v if x is not None] else: - v = unicode(v) + v = six.text_type(v) dc[field.name] = {'uri': field.uri, 'value': v} rdf['fields'] = dc @@ -416,44 +431,50 @@ class WorkInfo(object): if v is not None: if field.multiple: - if len(v) == 0: continue - v = [ unicode(x) for x in v if x is not None ] + if len(v) == 0: + continue + v = [six.text_type(x) for x in v if x is not None] else: - v = unicode(v) + v = six.text_type(v) result[field.name] = v if field.salias: v = getattr(self, field.salias) - if v is not None: result[field.salias] = unicode(v) + if v is not None: + result[field.salias] = six.text_type(v) return result class BookInfo(WorkInfo): FIELDS = ( - Field( DCNS('audience'), 'audiences', salias='audience', multiple=True, - required=False), - - Field( DCNS('subject.period'), 'epochs', salias='epoch', multiple=True, - required=False), - Field( DCNS('subject.type'), 'kinds', salias='kind', multiple=True, - required=False), - Field( DCNS('subject.genre'), 'genres', salias='genre', multiple=True, - required=False), - - Field( DCNS('contributor.translator'), 'translators', \ - as_person, salias='translator', multiple=True, default=[]), - Field( DCNS('relation.hasPart'), 'parts', - WLURI, strict=as_wluri_strict, multiple=True, required=False), - Field( DCNS('relation.isVariantOf'), 'variant_of', - WLURI, strict=as_wluri_strict, required=False), - - Field( DCNS('relation.coverImage.url'), 'cover_url', required=False), - Field( DCNS('relation.coverImage.attribution'), 'cover_by', required=False), - Field( DCNS('relation.coverImage.source'), 'cover_source', required=False), + Field(DCNS('audience'), 'audiences', salias='audience', multiple=True, required=False), + + Field(DCNS('subject.period'), 'epochs', salias='epoch', multiple=True, required=False), + Field(DCNS('subject.type'), 'kinds', salias='kind', multiple=True, required=False), + Field(DCNS('subject.genre'), 'genres', salias='genre', multiple=True, required=False), + + Field(DCNS('subject.location'), 'location', required=False), + + Field(DCNS('contributor.translator'), 'translators', + as_person, salias='translator', multiple=True, required=False), + Field(DCNS('relation.hasPart'), 'parts', WLURI, strict=as_wluri_strict, multiple=True, required=False), + Field(DCNS('relation.isVariantOf'), 'variant_of', WLURI, strict=as_wluri_strict, required=False), + + Field(DCNS('relation.coverImage.url'), 'cover_url', required=False), + Field(DCNS('relation.coverImage.attribution'), 'cover_by', required=False), + Field(DCNS('relation.coverImage.source'), 'cover_source', required=False), # WLCover-specific. - Field( WLNS('coverBarColor'), 'cover_bar_color', required=False), - Field( WLNS('coverBoxPosition'), 'cover_box_position', required=False), + Field(WLNS('coverBarColor'), 'cover_bar_color', required=False), + Field(WLNS('coverBoxPosition'), 'cover_box_position', required=False), + Field(WLNS('coverClass'), 'cover_class', default=['default']), + Field(WLNS('coverLogoUrl'), 'cover_logo_urls', multiple=True, required=False), + + Field('pdf-id', 'isbn_pdf', required=False), + Field('epub-id', 'isbn_epub', required=False), + Field('mobi-id', 'isbn_mobi', required=False), + Field('txt-id', 'isbn_txt', required=False), + Field('html-id', 'isbn_html', required=False), )