move stats to L2
[librarian.git] / src / librarian / elements / base.py
index fb5e3b1..863436c 100644 (file)
@@ -1,11 +1,29 @@
-# -*- coding: utf-8
-
+# This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
+# Copyright © Fundacja Wolne Lektury. See NOTICE for more information.
+#
+import copy
 import re
 from lxml import etree
 from librarian import dcparser, RDFNS
+from librarian.util import get_translation
+
+def last_words(text, n):
+    words = []
+    for w in reversed(text.split()):
+        words.append(w)
+        if len(w) > 2:
+            n -= 1
+            if not n: break
+    if n:
+        return n, text
+    else:
+        return n, ' '.join(reversed(words))
 
 
 class WLElement(etree.ElementBase):
+    SECTION_PRECEDENCE = None
+    ASIDE = False
+
     TXT_TOP_MARGIN = 0
     TXT_BOTTOM_MARGIN = 0
     TXT_PREFIX = ""
@@ -14,18 +32,24 @@ class WLElement(etree.ElementBase):
     HTML_TAG = None
     HTML_ATTR = {}
     HTML_CLASS = None
-    HTML_SECTION = False
-    
+
+    EPUB_TAG = None
+    EPUB_ATTR = {}
+    EPUB_CLASS = None
+    EPUB_START_CHUNK = False
+   
     CAN_HAVE_TEXT = True
     STRIP = False
 
     text_substitutions = [
-        (u'---', u'—'),
-        (u'--', u'–'),
-        (u'...', u'…'),
-        (u',,', u'„'),
-        (u'"', u'”'),
+        ('---', '—'),
+        ('--', '–'),
+        #('...', '…'),  # Temporary turnoff for epub
+        (',,', '„'),
+        ('"', '”'),
         ('\ufeff', ''),
+
+        ("'", "\u2019"),    # This was enabled for epub.
     ]
 
     @property
@@ -37,7 +61,7 @@ class WLElement(etree.ElementBase):
             else:
                 self._meta_object = None
         return self._meta_object
-    
+
     @property
     def meta(self):
         if self.meta_object is not None:
@@ -47,18 +71,63 @@ class WLElement(etree.ElementBase):
                 return self.getparent().meta
             else:
                 return self.document.base_meta
+
+    @property
+    def gettext(self):
+        return get_translation(self.meta.language).gettext
+
+    def in_context_of(self, setting):
+        parent = self.getparent()
+        if parent is None:
+            return False
+        try:
+            return getattr(parent, setting)
+        except AttributeError:
+            return parent.in_context_of(setting)
+
+    def signal(self, signal):
+        parent = self.getparent()
+        if parent is not None:
+            parent.signal(signal)
     
-    def normalize_text(self, text):
+    def raw_printable_text(self, builder):
+        from librarian.html import raw_printable_text
+
+        # TODO: podtagi, wyroznienia, etc
+        t = ''
+        t += self.normalize_text(self.text, builder)
+        for c in self:
+            if not isinstance(c, WLElement):
+                continue
+            if c.tag not in ('pe', 'pa', 'pt', 'pr', 'motyw'):
+                t += c.raw_printable_text(builder)
+            t += self.normalize_text(c.tail, builder)
+        return t
+    
+    def normalize_text(self, text, builder):
         text = text or ''
         for e, s in self.text_substitutions:
             text = text.replace(e, s)
-        text = re.sub(r'\s+', ' ', text)
+            # FIXME: TEmporary turnoff
+#        text = re.sub(r'\s+', ' ', text)
+### TODO: Added now for epub
+
+        if getattr(builder, 'hyphenator', None) is not None:
+            newt = ''
+            wlist = re.compile(r'\w+|[^\w]', re.UNICODE).findall(text)
+            for w in wlist:
+                newt += builder.hyphenator.inserted(w, '\u00AD')
+            text = newt
+
+        if builder.orphans:
+            text = re.sub(r'(?<=\s\w)\s+', '\u00A0', text)
+
         return text
 
     def _build_inner(self, builder, build_method):
         child_count = len(self)
         if self.CAN_HAVE_TEXT and self.text:
-            text = self.normalize_text(self.text)
+            text = self.normalize_text(self.text, builder)
             if self.STRIP:
                 text = text.lstrip()
                 if not child_count:
@@ -68,7 +137,7 @@ class WLElement(etree.ElementBase):
             if isinstance(child, WLElement):
                 getattr(child, build_method)(builder)
             if self.CAN_HAVE_TEXT and child.tail:
-                text = self.normalize_text(child.tail)
+                text = self.normalize_text(child.tail, builder)
                 if self.STRIP and i == child_count - 1:
                     text = text.rstrip()
                 builder.push_text(text)
@@ -99,41 +168,156 @@ class WLElement(etree.ElementBase):
         # always copy the id attribute (?)
         if self.attrib.get('id'):
             attr['id'] = self.attrib['id']
+        elif getattr(self, 'SHOULD_HAVE_ID', False) and '_compat_section_id' in self.attrib:
+            attr['id'] = self.attrib['_compat_section_id']
         return attr
-        
-    def html_build(self, builder):
-        if self.HTML_SECTION:
-            builder.start_element(
-                'a', {"name": "f18", "class": "target"}
-            )
-            builder.push_text(" ")
-            builder.end_element()
-
-            builder.start_element(
-                "a", {"href": "#f18", "class": "anchor"}
-            )
-            builder.push_text("18")
-            builder.end_element()
-        
 
+    def html_build(self, builder):
         if self.HTML_TAG:
             builder.start_element(
                 self.HTML_TAG,
                 self.get_html_attr(builder),
             )
 
-        if self.HTML_SECTION:
+        self._html_build_inner(builder)
+        if self.HTML_TAG:
+            builder.end_element()
+
+    def _epub_build_inner(self, builder):
+        self._build_inner(builder, 'epub_build')
+
+    def get_epub_attr(self, builder):
+        attr = self.EPUB_ATTR.copy()
+        if self.EPUB_CLASS:
+            attr['class'] = self.EPUB_CLASS
+        return attr
+
+    def epub_build(self, builder):
+        from librarian.elements.masters import Master
+
+        # TEMPORARY
+        self.CAN_HAVE_TEXT = True
+        self.STRIP = False
+       
+        start_chunk = self.EPUB_START_CHUNK and isinstance(self.getparent(), Master)
+
+        if start_chunk:
+            builder.start_chunk()
+
+        fragment = None
+        if self.SECTION_PRECEDENCE and not self.in_context_of('NO_TOC'):
+            if not start_chunk:
+                fragment = 'sub%d' % builder.assign_section_number()
+                self.attrib['id'] = fragment
+
+            builder.add_toc_entry(
+                fragment,
+                self.raw_printable_text(builder),
+                self.SECTION_PRECEDENCE
+            )
+            
+        if self.EPUB_TAG:
+            attr = self.get_epub_attr(builder)
+            if fragment:
+                attr['id'] = fragment
             builder.start_element(
-                "a", {"name": "sec34"}
+                self.EPUB_TAG,
+                attr
             )
-            builder.end_element()
 
-        self._html_build_inner(builder)
-        if self.HTML_TAG:
+        self._epub_build_inner(builder)
+        if self.EPUB_TAG:
             builder.end_element()
 
+    def validate(self):
+        from librarian.elements.masters import Master
+        from librarian.elements.blocks import DlugiCytat, PoezjaCyt
+        from librarian.elements.footnotes import Footnote
+
+        if self.SECTION_PRECEDENCE:
+            assert isinstance(self.getparent(), (Master, DlugiCytat, PoezjaCyt, Footnote)), \
+                    'Header {} inside a <{}> instead of a master.'.format(
+                            etree.tostring(self, encoding='unicode'), self.getparent().tag)
+
+        for c in self:
+            if isinstance(c, WLElement):
+                c.validate()
+
+
     def sanitize(self):
         # TODO: Remove insanity here.
         for e in self:
             if isinstance(e, WLElement):
                 e.sanitize()
+
+    def snip(self, words, before=None, sub=False):
+        if sub and self.ASIDE:
+            return words, []
+
+        snippet = []
+        if before is not None:
+            i = self.index(before)
+        else:
+            i = len(self)
+
+        while i > 0:
+            i -= 1
+            if self[i].tail:
+                if words:
+                    words, text = last_words(self[i].tail, words)
+                    snippet = [('text', text)] + snippet
+
+            if words:
+                words, subsnip = self[i].snip(words, sub=True)
+                snippet = subsnip + snippet
+
+        if words and self.text:
+            words, text = last_words(self.text, words)
+            snippet = [('text', text)] + snippet
+                    
+        snippet = [('start', self.tag, self.attrib)] + snippet + [('end',)]
+
+        if not sub and words and not self.ASIDE:
+            # do we dare go up?
+            parent = self.getparent()
+            if parent is not None and parent.CAN_HAVE_TEXT:
+                words, parsnip = parent.snip(words, before=self)
+                return words, parsnip[:-1] + snippet + parsnip[-1:]
+
+        return words, snippet
+
+    def get_snippet(self, words=15):
+        from librarian.parser import parser
+
+        words, snippet = self.getparent().snip(words=words, before=self)
+        
+        cursor = snipelem = parser.makeelement('snippet')
+        snipelem._meta_object = self.meta
+        for s in snippet:
+            if s[0] == 'start':
+                elem = parser.makeelement(s[1], **s[2])
+                cursor.append(elem)
+                cursor = elem
+            elif s[0] == 'end':
+                cursor = cursor.getparent()
+            else:
+                if len(cursor):
+                    cursor[-1].tail = (cursor[-1].tail or '') + s[1]
+                else:
+                    cursor.text = (cursor.text or '') + s[1]
+
+        return snipelem
+
+    def get_link(self):
+        sec = getattr(self, 'SHOULD_HAVE_ID', False) and self.attrib.get('_compat_section_id')
+        if sec:
+            return sec
+        parent_index = self.getparent().index(self)
+        if parent_index:
+            return self.getparent()[parent_index - 1].get_link()
+        else:
+            return self.getparent().get_link()
+
+
+class Snippet(WLElement):
+    pass