3 files changed, 258 insertions, 189 deletions

diff --git a/wiki2html.py b/wiki2html.py
index 907e3b1..7fa97b7 100644
--- a/wiki2html.py
+++ b/wiki2html.py
@@ -26,12 +26,44 @@ class HtmlWikiMarkup (WikiMarkup):
            2. [[official position]]s : final 's' gets after closing </a> tag.
            Should be before.
     """
+
+    # FIXME: Awful kludge
+    image_kw = [ 'Image',
+                 'Grafika',
+                 'Bild',
+                 'Εικόνα',
+                 'Dosiero',
+                 'Slika',
+                 'Resim'
+                 ]
+    
+    ST_INIT = 0
+    ST_PARA = 1
+    ST_OPEN = 2
+    
+    state = []
+
+    def opara(self):
+        if self.state[-1] == self.ST_PARA:
+            self.state[-1] = self.ST_OPEN
+            return "<p>"
+        else:
+            return ""
         
+    def cpara(self):
+        state = self.state.pop();
+        self.state.append(self.ST_INIT)
+        if state == self.ST_OPEN:
+            return "</p>"
+        else:
+            return ""
+    
+            
     def target(self, t):
         (qual,sep,tgt) = t.partition(':')
         r = None
         if tgt != '':
-            if qual in ('Image', 'Grafika'):
+            if qual in self.image_kw:
                 t = self.image_base + urllib.quote(tgt) + '/250px-' + urllib.quote(tgt)
             elif qual == "Media":
                 t = self.media_base + '/' + tgt
@@ -96,11 +128,30 @@ class HtmlWikiMarkup (WikiMarkup):
                         self.fmtok(tok[1], env),
                         self.envel[env[1]])
 
-    def str_seq(self, tok, env):
+    def str_para(self, tok, env):
+        s = self.cpara()
+        self.state.append(self.ST_PARA)
+        return s
+
+    def fmtok(self, tok, env):
+        if type(tok) != TupleType:
+            return ""
+        if tok[0] in [ self.ENV, self.HDR ]:
+            s = self.cpara()
+        elif tok[0] == self.BAR:
+            s = self.str_para(tok, env)
+        elif tok[0] in [ self.NIL, self.SEQ ]:
             s = ""
-            for t in tok[1:]:
-                s += self.fmtok(t, env)
-            return s
+        else:
+            s = self.opara()
+        s1 = WikiMarkup.fmtok(self, tok, env)
+        if s1:
+            s += s1
+        return s
+    
+    def __str__(self):
+        self.state = [ self.ST_PARA ]
+        return WikiMarkup.__str__(self) + self.cpara()
 
 
 
@@ -109,7 +160,7 @@ class HtmlWiktionaryMarkup (HtmlWikiMarkup):
  A class for translating Wiktionary articles into HTML.
  This version does not do much, except that it tries to correctly
  format templates. But "tries" does not mean "does". The heuristics
- used here is clearly not enogh to cope with it.
+ used here is clearly not enough to cope with it.
 
  1. FIXME:    
  The right solution would be to have a database of templates with their
@@ -134,12 +185,14 @@ class HtmlWiktionaryMarkup (HtmlWikiMarkup):
     seq_pos = 0
     
     def str_seq(self, tok, env):
-            s = ""
+            str = ""
             self.seq_pos=0
             for t in tok[1:]:
-                s += self.fmtok(t, env)
-                self.seq_pos += 1
-            return s
+                s = self.fmtok(t, env)
+                if s:
+                    str += s
+                    self.seq_pos += 1
+            return str
     
     def str_tmpl(self, tok, env):
         arg = self.fmtok(tok[1], env)
diff --git a/wiki2text.py b/wiki2text.py
index e943f32..3669bd7 100644
--- a/wiki2text.py
+++ b/wiki2text.py
@@ -136,6 +136,9 @@ class TextWikiMarkup (WikiMarkup):
             return "" + self.indent(lev,
                                       "- " + self.fmtok(tok[1], env))
 
+    def str_para(self, tok, env):
+        return "\n"
+    
     def __str__(self):
         return self.fmtok(self.tree, None)
 
diff --git a/wikimarkup.py b/wikimarkup.py
index d9ae7cc..e2a1cab 100644
--- a/wikimarkup.py
+++ b/wikimarkup.py
@@ -35,10 +35,11 @@ class BaseWikiMarkup:
     """
 A base class for handling Wiki markups.
 It handles:
-   1. basic block markup (headers, numbered and unnumbered lists,
+   1. paragraphs;
+   2. basic block markup (headers, numbered and unnumbered lists,
       indentations);
-   2. basic inline markup (bold, italic);
-   3. basic reference markup (links, templates, external links).
+   3. basic inline markup (bold, italic);
+   4. basic reference markup (links, templates, external links).
  It does NOT handle:
    1. pseudo-html markup (<nowiki></nowiki>, and similar);
    2. leading spaces meaning ``preserve formatting'';
@@ -90,6 +91,8 @@ It handles:
     ITEM = 10
     # Sequence: seq
     SEQ  = 11
+    # Paragraph
+    PARA = 12
     
     # Environment types:
     # Unnumbered list
@@ -130,6 +133,10 @@ It handles:
                 self.putback(line)
                 break
 
+            if line == '\n':
+                yield(self.PARA,)
+                continue
+            
             m = eltbeg.match(line)
             if m:
                 if m.group(0)[0] in self.envtypes:
@@ -247,8 +254,6 @@ It handles:
             return toktype, self.expandtok(tok[1])
         elif toktype == self.HDR:
             return toktype, tok[1], self.expandtok(tok[2])
-        elif toktype == self.BAR:
-            return tok
         elif toktype == self.ENV:
             return toktype,tok[1],tok[2],self.expandtok(tok[3])
         elif toktype == self.SEQ:
@@ -264,6 +269,8 @@ It handles:
                         subtree.append(x)
                 return tuple(subtree) if len(subtree) > 2 else \
                        subtree[1] if len(subtree) == 2 else None
+        else:
+            return tok
 
     def parse(self):
         tree = [self.SEQ]
@@ -314,6 +321,8 @@ It handles:
             elif toktype == self.ITEM:
                 print "ITEM"
                 self.prtok(tok[1], indent+1)
+            elif toktype == self.PARA:
+                print "PARA"
     
     def output(self):
         self.prtok(self.tree, 0)
@@ -377,314 +386,316 @@ class WikiMarkup (BaseWikiMarkup):
 
     # ISO 639 
     langtab = {
-        "aa": "Afar", # Afar
-        "ab": "Аҧсуа", # Abkhazian
-        "ae": None, # Avestan
-        "af": "Afrikaans", # Afrikaans
-        "ak": "Akana", # Akan # or ak_CI
+        "aa": "Afar",            # Afar
+        "ab": "Аҧсуа",           # Abkhazian
+        "ae": None,              # Avestan
+        "af": "Afrikaans",       # Afrikaans
+        "ak": "Akana",           # Akan
         "als": "Alemannisch",
-        "am": "አማርኛ", # Amharic
-        "an": "Aragonés", # Aragonese
+        "am": "አማርኛ",            # Amharic
+        "an": "Aragonés",        # Aragonese
         "ang": "Englisc",
-        "ar": "العربية" , # Arabic
+        "ar": "العربية" ,        # Arabic
         "arc": "ܐܪܡܝܐ",
-        "as": "অসমীয়া", # Assamese
+        "as": "অসমীয়া",         # Assamese
         "ast": "Asturian",