some cleaning

[librarian.git] / librarian / parser.py

diff --git a/librarian/parser.py b/librarian/parser.py
index 55b4e4b..4cdaa79 100644 (file)
--- a/librarian/parser.py
+++ b/librarian/parser.py
@@ -1,6 +1,10 @@
 # -*- coding: utf-8 -*-
+#
+# This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
+#
 from librarian import ValidationError, NoDublinCore,  ParseError
-from librarian import RDFNS, DCNS
+from librarian import RDFNS
 from librarian import dcparser
 
 from xml.parsers.expat import ExpatError
@@ -17,9 +21,9 @@ class WLDocument(object):
         self.edoc = edoc
 
         root_elem = edoc.getroot()
-       
+
         dc_path = './/' + RDFNS('RDF')
-        
+
         if root_elem.tag != 'utwor':
             raise ValidationError("Invalid root element. Found '%s', should be 'utwor'" % root_elem.tag)
 
@@ -28,14 +32,14 @@ class WLDocument(object):
 
             if self.rdf_elem is None:
                 raise NoDublinCore('Document has no DublinCore - which is required.')
-            
+
             self.book_info = dcparser.BookInfo.from_element(self.rdf_elem)
         else:
             self.book_info = None
-    
+
     @classmethod
-    def from_string(cls, xml, swap_endlines=False, parse_dublincore=True):
-        return cls.from_file(StringIO(xml), swap_endlines, parse_dublincore=parse_dublincore)
+    def from_string(cls, xml, *args, **kwargs):
+        return cls.from_file(StringIO(xml), *args, **kwargs)
 
     @classmethod
     def from_file(cls, xmlfile, swap_endlines=False, parse_dublincore=True):
@@ -53,27 +57,49 @@ class WLDocument(object):
         if not isinstance(data, unicode):
             data = data.decode('utf-8')
 
-        if swap_endlines:
-            data = cls.LINE_SWAP_EXPR.sub(u'<br />\n', data)
-    
+        data = data.replace(u'\ufeff', '')
+
         try:
-            parser = etree.XMLParser(remove_blank_text=True)
-            return cls(etree.parse(StringIO(data), parser), parse_dublincore=parse_dublincore)
-        except (ExpatError, XMLSyntaxError, XSLTApplyError), e:
-            raise ParseError(e)                  
+            parser = etree.XMLParser(remove_blank_text=False)
+            tree = etree.parse(StringIO(data), parser)
 
-    def part_as_text(self, path):
-        # convert the path to XPath        
-        print "[L] Retrieving part:", path
+            if swap_endlines:
+                cls.swap_endlines(tree)
 
-        elems = self.edoc.xpath(self.path_to_xpath(path))
-        print "[L] xpath", elems
-        
-        if len(elems) == 0:
-            return None        
+            return cls(tree, parse_dublincore=parse_dublincore)
+        except (ExpatError, XMLSyntaxError, XSLTApplyError), e:
+            raise ParseError(e)
 
-        return etree.tostring(elems[0], encoding=unicode, pretty_print=True)
+    @classmethod
+    def swap_endlines(cls, tree):
+        # only swap inside stanzas
+        for elem in tree.iter('strofa'):
+            for child in list(elem):
+                if child.tail:
+                    chunks = cls.LINE_SWAP_EXPR.split(child.tail)
+                    ins_index = elem.index(child) + 1
+                    while len(chunks) > 1:
+                        ins = etree.Element('br')
+                        ins.tail = chunks.pop()
+                        elem.insert(ins_index, ins)
+                    child.tail = chunks.pop(0)
+            if elem.text:
+                chunks = cls.LINE_SWAP_EXPR.split(elem.text)
+                while len(chunks) > 1:
+                    ins = etree.Element('br')
+                    ins.tail = chunks.pop()
+                    elem.insert(0, ins)
+                elem.text = chunks.pop(0)
+
+    def chunk(self, path):
+        # convert the path to XPath
+        expr = self.path_to_xpath(path)
+        elems = self.edoc.xpath(expr)
 
+        if len(elems) == 0:
+            return None
+        else:
+            return elems[0]
 
     def path_to_xpath(self, path):
         parts = []
@@ -84,7 +110,7 @@ class WLDocument(object):
                 parts.append(part)
             else:
                 tag, n = match.groups()
-                parts.append("node()[position() = %d and name() = '%s']" % (int(n), tag) )
+                parts.append("*[%d][name() = '%s']" % (int(n)+1, tag) )
 
         if parts[0] == '.':
             parts[0] = ''
@@ -95,8 +121,9 @@ class WLDocument(object):
         return self.edoc.xslt(stylesheet, **options)
 
     def update_dc(self):
-        parent = self.rdf_elem.getparent()
-        parent.replace( self.rdf_elem, self.book_info.to_etree(parent) )
+        if self.book_info:
+            parent = self.rdf_elem.getparent()
+            parent.replace( self.rdf_elem, self.book_info.to_etree(parent) )
 
     def serialize(self):
         self.update_dc()
@@ -108,10 +135,20 @@ class WLDocument(object):
         for key, data in chunk_dict.iteritems():
             try:
                 xpath = self.path_to_xpath(key)
-                node = self.edoc.xpath(xpath)[0]                
-                repl = etree.fromstring(data)
+                node = self.edoc.xpath(xpath)[0]
+                repl = etree.fromstring(u"<%s>%s</%s>" %(node.tag, data, node.tag) )
                 node.getparent().replace(node, repl);
             except Exception, e:
                 unmerged.append( repr( (key, xpath, e) ) )
 
-        return unmerged
\ No newline at end of file
+        return unmerged
+
+    def clean_ed_note(self):
+        """ deletes forbidden tags from nota_red """
+
+        for node in self.edoc.xpath('|'.join('//nota_red//%s' % tag for tag in
+                    ('pa', 'pe', 'pr', 'pt', 'begin', 'end', 'motyw'))):
+            tail = node.tail
+            node.clear()
+            node.tag = 'span'
+            node.tail = tail