first take on The Great Import

[redakcja.git] / apps / catalogue / xml_tools.py

diff --git a/apps/catalogue/xml_tools.py b/apps/catalogue/xml_tools.py
index 928e57b..522806b 100755 (executable)
--- a/apps/catalogue/xml_tools.py
+++ b/apps/catalogue/xml_tools.py
@@ -1,3 +1,5 @@
+# -*- coding: utf-8 -*-
+from copy import deepcopy
 from functools import wraps
 import re
 
@@ -139,10 +141,11 @@ def compile_text(parts):
     for next_text in parts:
         if not next_text:
             continue
-        # trim the end, because there's more non-empty text
-        # don't trim beginning, if `text' is the first non-empty part
-        texts.append(_trim(text, trim_begin=trim_begin))
-        trim_begin = True
+        if text:
+            # trim the end, because there's more non-empty text
+            # don't trim beginning, if `text' is the first non-empty part
+            texts.append(_trim(text, trim_begin=trim_begin))
+            trim_begin = True
         text = next_text
     # don't trim the end, because there's no more text coming after `text'
     # only trim beginning if it's not still the first non-empty
@@ -156,7 +159,7 @@ def change_master(text, master):
     """
     e = etree.fromstring(text)
     e[-1].tag = master
-    return etree.tostring(e, encoding="utf-8")
+    return unicode(etree.tostring(e, encoding="utf-8"), 'utf-8')
 
 
 def basic_structure(text, master):
@@ -167,7 +170,7 @@ def basic_structure(text, master):
 </utwor>''' % (TRIM_BEGIN, TRIM_END))
     e[0].tag = master
     e[0][0].tail = "\n"*3 + text + "\n"*3
-    return etree.tostring(e, encoding="utf-8")
+    return unicode(etree.tostring(e, encoding="utf-8"), 'utf-8')
 
 
 def add_trim_begin(text):
@@ -182,7 +185,7 @@ def add_trim_begin(text):
     master.insert(0, trim_tag)
     trim_tag.tail = '\n\n\n' + (master.text or '')
     master.text = '\n'
-    return etree.tostring(e, encoding="utf-8")
+    return unicode(etree.tostring(e, encoding="utf-8"), 'utf-8')
 
 
 def add_trim_end(text):
@@ -201,4 +204,63 @@ def add_trim_end(text):
         prev.tail = (prev.tail or '') + '\n\n\n'
     else:
         master.text = (master.text or '') + '\n\n\n'
-    return etree.tostring(e, encoding="utf-8")
+    return unicode(etree.tostring(e, encoding="utf-8"), 'utf-8')
+
+
+def split_xml(text):
+    """Splits text into chapters.
+
+    All this stuff really must go somewhere else.
+
+    """
+    src = etree.fromstring(text)
+    chunks = []
+
+    splitter = u'naglowek_rozdzial'
+    parts = src.findall('.//naglowek_rozdzial')
+    while parts:
+        # copy the document
+        copied = deepcopy(src)
+
+        element = parts[-1]
+
+        # find the chapter's title
+        name_elem = deepcopy(element)
+        for tag in 'extra', 'motyw', 'pa', 'pe', 'pr', 'pt', 'uwaga':
+            for a in name_elem.findall('.//' + tag):
+                a.text=''
+                del a[:]
+        name = etree.tostring(name_elem, method='text', encoding='utf-8')
+
+        # in the original, remove everything from the start of the last chapter
+        parent = element.getparent()
+        del parent[parent.index(element):]
+        element, parent = parent, parent.getparent()
+        while parent is not None:
+            del parent[parent.index(element) + 1:]
+            element, parent = parent, parent.getparent()
+
+        # in the copy, remove everything before the last chapter
+        element = copied.findall('.//naglowek_rozdzial')[-1]
+        parent = element.getparent()
+        while parent is not None:
+            parent.text = None
+            while parent[0] is not element:
+                del parent[0]
+            element, parent = parent, parent.getparent()
+        chunks[:0] = [[name,
+            unicode(etree.tostring(copied, encoding='utf-8'), 'utf-8')
+            ]]
+
+        parts = src.findall('.//naglowek_rozdzial')
+
+    chunks[:0] = [[u'początek',
+        unicode(etree.tostring(src, encoding='utf-8'), 'utf-8')
+        ]]
+
+    for ch in chunks[1:]:
+        ch[1] = add_trim_begin(ch[1])
+    for ch in chunks[:-1]:
+        ch[1] = add_trim_end(ch[1])
+
+    return chunks