fixes for edge cases
[librarian.git] / src / librarian / document.py
index 1c8f223..6cdf191 100644 (file)
@@ -1,16 +1,16 @@
 import gettext
 import os
 import re
 import gettext
 import os
 import re
-from urllib.request import urlopen
 from lxml import etree
 from lxml import etree
+import six
 from .parser import parser
 from .parser import parser
-from . import dcparser, DCNS
+from . import dcparser, DCNS, DirDocProvider
 from .functions import lang_code_3to2
 
 
 class WLDocument:
 from .functions import lang_code_3to2
 
 
 class WLDocument:
-    def __init__(self, filename=None, url=None):
-        source = filename or urlopen(url)
+    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
         tree = etree.parse(source, parser=parser)
         self.tree = tree
         tree.getroot().document = self
@@ -18,6 +18,10 @@ class WLDocument:
             DCNS('language'): ["pol"],
         }, validate_required=False)
 
             DCNS('language'): ["pol"],
         }, validate_required=False)
 
+        self.provider = provider if provider is not None else DirDocProvider('.')
+
+        self.tree.getroot().validate()
+
     @property
     def meta(self):
         # Allow metadata of the master element as document meta.
     @property
     def meta(self):
         # Allow metadata of the master element as document meta.
@@ -25,9 +29,44 @@ class WLDocument:
         return self.tree.getroot().meta
         return master.meta
 
         return self.tree.getroot().meta
         return master.meta
 
-    def build(self, builder, **kwargs):
-        return builder().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 assign_ids(self, existing=None):
+        # Find all existing IDs.
+        existing = existing or set()
+        que = [self.tree.getroot()]
+        while que:
+            item = que.pop(0)
+            try:
+                item.normalize_insides()
+            except AttributeError:
+                pass
+            existing.add(item.attrib.get('id'))
+            que.extend(item)
 
 
+        i = 1
+        que = [self.tree.getroot()]
+        while que:
+            item = que.pop(0)
+            que.extend(item)
+            if item.attrib.get('id'):
+                continue
+            if not getattr(item, 'SHOULD_HAVE_ID', False):
+                continue
+            while f'e{i}' in existing:
+                i += 1
+            item.attrib['id'] = f'e{i}'
+            i += 1
+    
     def _compat_assign_ordered_ids(self):
         """
         Compatibility: ids in document order, to be roughly compatible with legacy
     def _compat_assign_ordered_ids(self):
         """
         Compatibility: ids in document order, to be roughly compatible with legacy
@@ -62,7 +101,22 @@ class WLDocument:
         def _compat_assigns_section_ids_in_elem(elem, prefix='sec'):
             for i, child in enumerate(elem):
                 idfier = '{}{}'.format(prefix, i + 1)
         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
+                try:
+                    child.attrib['_compat_section_id'] = idfier
+                except:
+                    pass
                 _compat_assigns_section_ids_in_elem(child, idfier + '-')
         _compat_assigns_section_ids_in_elem(self.tree.getroot().master)
 
                 _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
+
+    def references(self):
+        return self.tree.findall('.//ref')