Updated book2txt.py.

[wolnelektury.git] / lib / librarian / html.py

diff --git a/lib/librarian/html.py b/lib/librarian/html.py
index c75dd6b..8d256ee 100644 (file)
--- a/lib/librarian/html.py
+++ b/lib/librarian/html.py
@@ -53,8 +53,13 @@ def transform(input_filename, output_filename):
     doc = etree.parse(doc_file, parser)
 
     result = doc.xslt(style)
-    add_anchors(result.getroot())
-    result.write(output_filename, xml_declaration=True, pretty_print=True, encoding='utf-8')
+    if result.find('//h1') is not None:
+        add_anchors(result.getroot())
+        add_table_of_contents(result.getroot())
+        result.write(output_filename, xml_declaration=False, pretty_print=True, encoding='utf-8')
+        return True
+    else:
+        return False
 
 
 class Fragment(object):
@@ -108,7 +113,7 @@ def extract_fragments(input_filename):
 
     for event, element in etree.iterparse(input_filename, events=('start', 'end')):
         # Process begin and end elements
-        if element.tag == 'span' and element.get('class', '') in ('theme-begin', 'theme-end'):
+        if element.get('class', '') in ('theme-begin', 'theme-end'):
             if not event == 'end': continue # Process elements only once, on end event
 
             # Open new fragment
@@ -116,9 +121,9 @@ def extract_fragments(input_filename):
                 fragment = Fragment(id=element.get('fid'), themes=element.text)
 
                 # Append parents
-                if element.getparent().tag != 'body':
+                if element.getparent().get('id', None) != 'book-text':
                     parents = [element.getparent()]
-                    while parents[-1].getparent().tag != 'body':
+                    while parents[-1].getparent().get('id', None) != 'book-text':
                         parents.append(parents[-1].getparent())
 
                     parents.reverse()
@@ -157,33 +162,81 @@ def extract_fragments(input_filename):
     return closed_fragments, open_fragments
 
 
-def add_anchor(element, number):
-    anchor = etree.Element('a', href='#f%d' % number)
-    anchor.set('class', 'anchor')
-    anchor.text = str(number)
-    if element.text:
-        anchor.tail = element.text
-        element.text = u''
-    element.insert(0, anchor)
+def add_anchor(element, prefix, with_link=True, with_target=True, link_text=None):
+    if with_link:
+        if link_text is None:
+            link_text = prefix
+        anchor = etree.Element('a', href='#%s' % prefix)
+        anchor.set('class', 'anchor')
+        anchor.text = unicode(link_text)
+        if element.text:
+            anchor.tail = element.text
+            element.text = u''
+        element.insert(0, anchor)
     
-    anchor_target = etree.Element('a', name='f%d' % number)
-    element.insert(0, anchor_target)
+    if with_target:
+        anchor_target = etree.Element('a', name='%s' % prefix)
+        anchor_target.set('class', 'target')
+        anchor_target.text = u' '
+        if element.text:
+            anchor_target.tail = element.text
+            element.text = u''
+        element.insert(0, anchor_target)
+
+
+def any_ancestor(element, test):
+    for ancestor in element.iterancestors():
+        if test(ancestor):
+            return True
+    return False
 
 
 def add_anchors(root):
     counter = 1
     for element in root.iterdescendants():
-        if element.getparent().tag in 'div' and 'note' in element.getparent().get('class', ''):
-            continue
-        if element.getparent().tag in 'blockquote':
+        if any_ancestor(element, lambda e: e.get('class') in ('note', 'motto', 'motto_podpis', 'dedication')
+        or e.tag == 'blockquote'):
             continue
         
         if element.tag == 'p' and 'verse' in element.get('class', ''):
             if counter == 1 or counter % 5 == 0:
-                add_anchor(element, counter)
+                add_anchor(element, "f%d" % counter, link_text=counter)
             counter += 1
         elif 'paragraph' in element.get('class', ''):
-            add_anchor(element, counter)
+            add_anchor(element, "f%d" % counter, link_text=counter)
             counter += 1
 
 
+def add_table_of_contents(root):
+    sections = []
+    counter = 1
+    for element in root.iterdescendants():
+        if element.tag in ('h2', 'h3'):
+            if any_ancestor(element, lambda e: e.get('id') in ('footnotes',)):
+                continue
+            
+            if element.tag == 'h3' and len(sections) and sections[-1][1] == 'h2':
+                sections[-1][3].append((counter, element.tag, ''.join(element.xpath('text()')), []))
+            else:
+                sections.append((counter, element.tag, ''.join(element.xpath('text()')), []))
+            add_anchor(element, "s%d" % counter, with_link=False)
+            counter += 1
+    
+    toc = etree.Element('div')
+    toc.set('id', 'toc')
+    toc_header = etree.SubElement(toc, 'h2')
+    toc_header.text = u'Spis treści'
+    toc_list = etree.SubElement(toc, 'ol')
+
+    for n, section, text, subsections in sections:
+        section_element = etree.SubElement(toc_list, 'li')
+        add_anchor(section_element, "s%d" % n, with_target=False, link_text=text)
+        
+        if len(subsections):
+            subsection_list = etree.SubElement(section_element, 'ol')
+            for n, subsection, text, _ in subsections:
+                subsection_element = etree.SubElement(subsection_list, 'li')
+                add_anchor(subsection_element, "s%d" % n, with_target=False, link_text=text)
+    
+    root.insert(0, toc)
+