X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/0604bdd5f693da9f1c78f9d9fa2276f0c7b6c17b..eae370b75207e1dc15f5f3f7ee7a402c7721d9a8:/src/librarian/document.py?ds=inline diff --git a/src/librarian/document.py b/src/librarian/document.py index 1bd249d..a3ee093 100644 --- a/src/librarian/document.py +++ b/src/librarian/document.py @@ -1,16 +1,26 @@ +import gettext +import os +import re from lxml import etree -from .builders import get_builder_class +import six from .parser import parser -from . import dcparser +from . import dcparser, DCNS, DirDocProvider +from .functions import lang_code_3to2 class WLDocument: - def __init__(self, tree=None, filename=None): - if filename is not None: - tree = etree.parse(filename, parser=parser) + def __init__(self, filename=None, url=None, provider=None): + source = filename or six.moves.urllib.request.urlopen(url) + tree = etree.parse(source, parser=parser) self.tree = tree tree.getroot().document = self - self.base_meta = dcparser.BookInfo({}, {}, validate_required=False) + self.base_meta = dcparser.BookInfo({}, { + DCNS('language'): ["pol"], + }, validate_required=False) + + self.provider = provider if provider is not None else DirDocProvider('.') + + self.tree.getroot().validate() @property def meta(self): @@ -19,6 +29,63 @@ class WLDocument: return self.tree.getroot().meta return master.meta - def build(self, builder_id, **kwargs): - return get_builder_class(builder_id)().build(self, **kwargs) - + @property + def children(self): + for part_uri in self.meta.parts or []: + yield type(self)( + filename=self.provider.by_uri(part_uri), + provider=self.provider + ) + + + def build(self, builder, base_url=None, **kwargs): + return builder(base_url=base_url).build(self, **kwargs) + + def _compat_assign_ordered_ids(self): + """ + Compatibility: ids in document order, to be roughly compatible with legacy + footnote ids. Just for testing consistency, change to some sane identifiers + at convenience. + """ + EXPR = re.compile(r'/\s', re.MULTILINE | re.UNICODE) + def _compat_assign_ordered_ids_in_elem(elem, i): + elem.attrib['_compat_ordered_id'] = str(i) + i += 1 + if getattr(elem, 'HTML_CLASS', None) == 'stanza': + if elem.text: + i += len(EXPR.split(elem.text)) - 1 + for sub in elem: + i = _compat_assign_ordered_ids_in_elem(sub, i) + if sub.tail: + i += len(EXPR.split(sub.tail)) - 1 + else: + if elem.tag in ('uwaga', 'extra'): + return i + for sub in elem: + i = _compat_assign_ordered_ids_in_elem(sub, i) + return i + + _compat_assign_ordered_ids_in_elem(self.tree.getroot(), 4) + + def _compat_assign_section_ids(self): + """ + Ids in master-section order. These need to be compatible with the + #secN anchors used by WL search results page to link to fragments. + """ + def _compat_assigns_section_ids_in_elem(elem, prefix='sec'): + for i, child in enumerate(elem): + idfier = '{}{}'.format(prefix, i + 1) + child.attrib['_compat_section_id'] = idfier + _compat_assigns_section_ids_in_elem(child, idfier + '-') + _compat_assigns_section_ids_in_elem(self.tree.getroot().master) + + + def editors(self): + persons = set(self.meta.editors + + self.meta.technical_editors) + for child in self.children: + persons.update(child.editors()) + if None in persons: + persons.remove(None) + return persons +