X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/97ed31c1ea6c21e0ac6cad0bc25a3cf63ecdd1ad..665c88719c108626edee864affe07b085d28b7b4:/librarian/picture.py?ds=sidebyside diff --git a/librarian/picture.py b/librarian/picture.py index b770030..10d2ae7 100644 --- a/librarian/picture.py +++ b/librarian/picture.py @@ -1,28 +1,29 @@ +# -*- coding: utf-8 -*- +from __future__ import unicode_literals -from dcparser import (as_person, as_date, Field, WorkInfo, DCNS) +from operator import and_ + +from .dcparser import Field, WorkInfo, DCNS from librarian import (RDFNS, ValidationError, NoDublinCore, ParseError, WLURI) from xml.parsers.expat import ExpatError from os import path -from StringIO import StringIO from lxml import etree -from lxml.etree import (XMLSyntaxError, XSLTApplyError) +from lxml.etree import (XMLSyntaxError, XSLTApplyError, Element) import re +import six class WLPictureURI(WLURI): - _re_wl_uri = re.compile('http://wolnelektury.pl/katalog/obraz/' - '(?P[-a-z0-9]+)(/(?P[a-z]{3}))?/?$') - - def __init__(self, *args, **kw): - super(WLPictureURI, self).__init__(*args, **kw) + _re_wl_uri = re.compile('http://wolnelektury.pl/katalog/obraz/(?P[-a-z0-9]+)/?$') @classmethod - def from_slug_and_lang(cls, slug, lang): + def from_slug(cls, slug): uri = 'http://wolnelektury.pl/katalog/obraz/%s/' % slug return cls(uri) - def filename_stem(self): - return self.slug + +def as_wlpictureuri_strict(text): + return WLPictureURI.strict(text) class PictureInfo(WorkInfo): @@ -33,25 +34,22 @@ class PictureInfo(WorkInfo): Field(DCNS('language'), 'language', required=False), Field(DCNS('subject.period'), 'epochs', salias='epoch', multiple=True), Field(DCNS('subject.type'), 'kinds', salias='kind', multiple=True), + Field(DCNS('subject.genre'), 'genres', salias='genre', multiple=True, required=False), + Field(DCNS('subject.style'), 'styles', salias='style', multiple=True, required=False), Field(DCNS('format.dimensions'), 'dimensions', required=False), + Field(DCNS('format.checksum.sha1'), 'sha1', required=True), Field(DCNS('description.medium'), 'medium', required=False), Field(DCNS('description.dimensions'), 'original_dimensions', required=False), Field(DCNS('format'), 'mime_type', required=False), - Field(DCNS('identifier.url'), 'url', WLPictureURI), - ) - - def validate(self): - """ - WorkInfo has a language validation code only, which we do not need. - """ - pass + Field(DCNS('identifier.url'), 'url', WLPictureURI, strict=as_wlpictureuri_strict) + ) class ImageStore(object): EXT = ['gif', 'jpeg', 'png', 'swf', 'psd', 'bmp' - 'tiff', 'tiff', 'jpc', 'jp2', 'jpf', 'jb2', 'swc', - 'aiff', 'wbmp', 'xbm'] + 'tiff', 'tiff', 'jpc', 'jp2', 'jpf', 'jb2', 'swc', + 'aiff', 'wbmp', 'xbm'] MIME = ['image/gif', 'image/jpeg', 'image/png', 'application/x-shockwave-flash', 'image/psd', 'image/bmp', 'image/tiff', 'image/tiff', 'application/octet-stream', @@ -59,8 +57,8 @@ class ImageStore(object): 'application/x-shockwave-flash', 'image/iff', 'image/vnd.wap.wbmp', 'image/xbm'] def __init__(self, dir_): + super(ImageStore, self).__init__() self.dir = dir_ - return super(ImageStore, self).__init__() def path(self, slug, mime_type): """ @@ -100,16 +98,17 @@ class WLPicture(object): self.picture_info = PictureInfo.from_element(self.rdf_elem) else: self.picture_info = None + self.frame = None @classmethod - def from_string(cls, xml, *args, **kwargs): - 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, parse_dublincore=True, image_store=None): # first, prepare for parsing - if isinstance(xmlfile, basestring): + if isinstance(xmlfile, six.text_type): file = open(xmlfile, 'rb') try: data = file.read() @@ -118,21 +117,23 @@ class WLPicture(object): else: data = xmlfile.read() - if not isinstance(data, unicode): + if not isinstance(data, six.text_type): data = data.decode('utf-8') data = data.replace(u'\ufeff', '') # assume images are in the same directory - if image_store is None and xmlfile.name is not None: + if image_store is None and getattr(xmlfile, 'name', None): image_store = ImageStore(path.dirname(xmlfile.name)) try: parser = etree.XMLParser(remove_blank_text=False) - tree = etree.parse(StringIO(data.encode('utf-8')), parser) + tree = etree.parse(six.BytesIO(data.encode('utf-8')), parser) - return cls(tree, parse_dublincore=parse_dublincore, image_store=image_store) - except (ExpatError, XMLSyntaxError, XSLTApplyError), e: + me = cls(tree, parse_dublincore=parse_dublincore, image_store=image_store) + me.load_frame_info() + return me + except (ExpatError, XMLSyntaxError, XSLTApplyError) as e: raise ParseError(e) @property @@ -149,32 +150,64 @@ class WLPicture(object): def image_path(self): if self.image_store is None: raise ValueError("No image store associated with whis WLPicture.") + return self.image_store.path(self.slug, self.mime_type) def image_file(self, *args, **kwargs): - return open(self.image_path, *args, **kwargs) + return open(self.image_path, 'rb', *args, **kwargs) + + def get_sem_coords(self, sem): + area = sem.find("div[@type='rect']") + if area is None: + area = sem.find("div[@type='whole']") + return [[0, 0], [-1, -1]] + + def has_all_props(node, props): + return reduce(and_, map(lambda prop: prop in node.attrib, props)) + + if not has_all_props(area, ['x1', 'x2', 'y1', 'y2']): + return None + + def n(prop): return int(area.get(prop)) + return [[n('x1'), n('y1')], [n('x2'), n('y2')]] def partiter(self): """ Iterates the parts of this picture and returns them and their metadata """ - for part in self.edoc.iter("div"): - pd = {} - pd['type'] = part.get('type') - if pd['type'] == 'area': - pd['coords'] = ((int(part.get('x1')), int(part.get('y1'))), - (int(part.get('x2')), int(part.get('y2')))) - - pd['themes'] = [] - pd['object'] = None - parent = part - while True: - parent = parent.getparent() - if parent is None: - break - if parent.tag == 'sem': - if parent.get('type') == 'theme': - pd['themes'] += map(unicode.strip, unicode(parent.get('theme')).split(',')) - elif parent.get('type') == 'object' and not pd['object']: - pd['object'] = parent.get('name') + # omg no support for //sem[(@type='theme') or (@type='object')] ? + for part in list(self.edoc.iterfind("//sem[@type='theme']")) +\ + list(self.edoc.iterfind("//sem[@type='object']")): + pd = {'type': part.get('type')} + + coords = self.get_sem_coords(part) + if coords is None: + continue + pd['coords'] = coords + + def want_unicode(x): + if not isinstance(x, six.text_type): + return x.decode('utf-8') + else: + return x + pd['object'] = part.attrib['type'] == 'object' and want_unicode(part.attrib.get('object', u'')) or None + pd['themes'] = part.attrib['type'] == 'theme' and [part.attrib.get('theme', u'')] or [] yield pd + + def load_frame_info(self): + k = self.edoc.find("//sem[@object='kadr']") + + if k is not None: + clip = self.get_sem_coords(k) + self.frame = clip + frm = Element("sem", {"type": "frame"}) + frm.append(k.iter("div").next()) + self.edoc.getroot().append(frm) + k.getparent().remove(k) + else: + frm = self.edoc.find("//sem[@type='frame']") + if frm: + self.frame = self.get_sem_coords(frm) + else: + self.frame = None + return self