X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/97ed31c1ea6c21e0ac6cad0bc25a3cf63ecdd1ad..6791e72f5bb378973798e1d410f7bd42a346cc9e:/librarian/picture.py?ds=inline diff --git a/librarian/picture.py b/librarian/picture.py index b770030..5d644d7 100644 --- a/librarian/picture.py +++ b/librarian/picture.py @@ -5,24 +5,22 @@ 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 - +from functools import * +from operator import * class WLPictureURI(WLURI): _re_wl_uri = re.compile('http://wolnelektury.pl/katalog/obraz/' - '(?P[-a-z0-9]+)(/(?P[a-z]{3}))?/?$') - - def __init__(self, *args, **kw): - super(WLPictureURI, self).__init__(*args, **kw) + '(?P[-a-z0-9]+)/?$') @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): @@ -33,20 +31,17 @@ 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('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), + Field(DCNS('identifier.url'), 'url', WLPictureURI, + strict=as_wlpictureuri_strict), ) - def validate(self): - """ - WorkInfo has a language validation code only, which we do not need. - """ - pass - class ImageStore(object): EXT = ['gif', 'jpeg', 'png', 'swf', 'psd', 'bmp' @@ -131,7 +126,9 @@ class WLPicture(object): parser = etree.XMLParser(remove_blank_text=False) tree = etree.parse(StringIO(data.encode('utf-8')), parser) - return cls(tree, parse_dublincore=parse_dublincore, image_store=image_store) + me = cls(tree, parse_dublincore=parse_dublincore, image_store=image_store) + me.load_frame_info() + return me except (ExpatError, XMLSyntaxError, XSLTApplyError), e: raise ParseError(e) @@ -149,32 +146,62 @@ 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) + 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 has_all_props(area, ['x1', 'x2', 'y1', 'y2']) == False: + 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"): + # 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 = {} 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') + + coords = self.get_sem_coords(part) + if coords is None: continue + pd['coords'] = coords + + def want_unicode(x): + if not isinstance(x, unicode): 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