fix impovement for IT #3356 - workaround rare problem

[librarian.git] / librarian / epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index b17ca0e..8dba5ac 100644 (file)
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -16,15 +16,51 @@ import zipfile
 from tempfile import mkdtemp, NamedTemporaryFile
 from shutil import rmtree
 
 from tempfile import mkdtemp, NamedTemporaryFile
 from shutil import rmtree
 

-from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, OutputFile
+from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, DCNS, OutputFile

 from librarian.cover import DefaultEbookCover
 
 from librarian import functions, get_resource
 
 from librarian.cover import DefaultEbookCover
 
 from librarian import functions, get_resource
 

+from librarian.hyphenator import Hyphenator
+

 functions.reg_person_name()
 functions.reg_lang_code_3to2()
 
 functions.reg_person_name()
 functions.reg_lang_code_3to2()
 

-
+def set_hyph_language(source_tree):
+    def get_short_lng_code(text):
+        result = ''
+        text = ''.join(text)
+        with open(get_resource('res/ISO-639-2_8859-1.txt'), 'rb') as f:
+            for line in f:
+                list = line.strip().split('|')
+                if list[0] == text:
+                    result=list[2]
+        if result == '':
+            return text
+        else:
+            return result
+    bibl_lng = etree.XPath('//dc:language//text()', namespaces = {'dc':str(DCNS)})(source_tree)
+    short_lng = get_short_lng_code(bibl_lng[0])   
+    try:
+        return Hyphenator(get_resource('res/hyph-dictionaries/hyph_' + short_lng + '.dic'))
+    except:
+        pass
+    
+def hyphenate_and_fix_conjunctions(source_tree, hyph):
+    if hyph is not None:
+        texts = etree.XPath('/utwor/*[2]//text()')(source_tree)
+        for t in texts:
+            parent = t.getparent()
+            newt = ''
+            wlist = re.compile(r'\w+|[^\w]', re.UNICODE).findall(t)
+            for w in wlist:
+                newt += hyph.inserted(w, u'\u00AD')       
+            newt = re.sub(r'(?<=\s\w)\s+', u'\u00A0', newt)
+            if t.is_text:
+                parent.text = newt
+            elif t.is_tail:
+                parent.tail = newt
+        

 def inner_xml(node):
     """ returns node's text and children as a string
 
 def inner_xml(node):
     """ returns node's text and children as a string
 


@@ -78,7 +114,6 @@ def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None
     def replace_chars(text):
         if text is None:
             return None

-        #text = re.sub(r"(?<=\s\w)\s+", u"\u00a0", text) #fix for hanging single letter conjunctions – for future use.

         return text.replace(u"\ufeff", u"")\
                    .replace("---", u"\u2014")\
                    .replace("--", u"\u2013")\
         return text.replace(u"\ufeff", u"")\
                    .replace("---", u"\u2014")\
                    .replace("--", u"\u2013")\


@@ -249,7 +284,10 @@ class TOC(object):
 
             nav_label = nav_map.makeelement(NCXNS('navLabel'))
             text = nav_map.makeelement(NCXNS('text'))
 
             nav_label = nav_map.makeelement(NCXNS('navLabel'))
             text = nav_map.makeelement(NCXNS('text'))

-            text.text = re.sub(r'\n', ' ', child.name)
+            if child.name is not None:
+                text.text = re.sub(r'\n', ' ', child.name)
+            else:
+                text.text = child.name

             nav_label.append(text)
             nav_point.append(nav_label)
 
             nav_label.append(text)
             nav_point.append(nav_label)
 


@@ -310,7 +348,7 @@ def chop(main_text):
                 yield part_xml
                 last_node_part = True
                 main_xml_part[:] = [deepcopy(one_part)]
                 yield part_xml
                 last_node_part = True
                 main_xml_part[:] = [deepcopy(one_part)]

-            elif not last_node_part and name in ("naglowek_scena"):
+            elif not last_node_part and name == "naglowek_scena":

                 yield part_xml
                 main_xml_part[:] = [deepcopy(one_part)]
             else:
                 yield part_xml
                 main_xml_part[:] = [deepcopy(one_part)]
             else:


@@ -335,7 +373,7 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s
 
     toc = TOC()
     for element in chunk_xml[0]:
 
     toc = TOC()
     for element in chunk_xml[0]:

-        if element.tag in ("naglowek_czesc"):
+        if element.tag == "naglowek_czesc":

             toc.add(node_name(element), "part%d.html#book-text" % chunk_no)
         elif element.tag in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
             toc.add(node_name(element), "part%d.html" % chunk_no)
             toc.add(node_name(element), "part%d.html#book-text" % chunk_no)
         elif element.tag in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
             toc.add(node_name(element), "part%d.html" % chunk_no)


@@ -370,7 +408,11 @@ def transform(wldoc, verbose=False,
         """ processes one input file and proceeds to its children """
 
         replace_characters(wldoc.edoc.getroot())
         """ processes one input file and proceeds to its children """
 
         replace_characters(wldoc.edoc.getroot())

-
+        
+        hyphenator = set_hyph_language(wldoc.edoc.getroot())
+        hyphenate_and_fix_conjunctions(wldoc.edoc.getroot(), hyphenator)
+        
+        

         # every input file will have a TOC entry,
         # pointing to starting chunk
         toc = TOC(wldoc.book_info.title, "part%d.html" % chunk_counter)
         # every input file will have a TOC entry,
         # pointing to starting chunk
         toc = TOC(wldoc.book_info.title, "part%d.html" % chunk_counter)