lxml update
[librarian.git] / src / librarian / document.py
index 1bd249d..fbaf6ca 100644 (file)
@@ -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_slug(part_uri.slug),
+                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
+