* Change all entity refs in to unicode strings

[rss2maildir.git] / rss2maildir.py

diff --git a/rss2maildir.py b/rss2maildir.py
index 322dc5751d2f77efe51f301b5dba9a4b2799381b..910b8a913cf64ef9bc714cb2eefd72c28179a1ec 100755 (executable)
--- a/rss2maildir.py
+++ b/rss2maildir.py
@@ -48,189 +48,370 @@ from HTMLParser import HTMLParser
 
 class HTML2Text(HTMLParser):
     entities = {
 
 class HTML2Text(HTMLParser):
     entities = {

-        "amp": "&",
-        "lt": "<",
-        "gt": ">",
-        "pound": "£",
-        "copy": "©",
-        "apos": "'",
-        "quot": "\"",
-        "nbsp": " ",
+        u'amp': "&",
+        u'lt': "<",
+        u'gt': ">",
+        u'pound': "£",
+        u'copy': "©",
+        u'apos': "'",
+        u'quot': "\"",
+        u'nbsp': " ",

         }
 
         }
 

+    blockleveltags = [
+        u'h1',
+        u'h2',
+        u'h3',
+        u'h4',
+        u'h5',
+        u'h6',
+        u'pre',
+        u'p',
+        u'ul',
+        u'ol',
+        u'dl',
+        u'div',
+        #u'blockquote',
+        ]
+
+    liststarttags = [
+        u'ul',
+        u'ol',
+        u'dl',
+        ]
+
+    cancontainflow = [
+        u'div',
+        u'li',
+        u'dd',
+        u'blockquote',
+    ]
+

     def __init__(self,textwidth=70):
     def __init__(self,textwidth=70):

-        self.inheadingone = False
-        self.inheadingtwo = False
-        self.inotherheading = False
-        self.inparagraph = True
-        self.inblockquote = False
-        self.inlink = False

         self.text = u''
         self.text = u''

-        self.currentparagraph = u''
-        self.headingtext = u''
-        self.blockquote = u''
-        self.inpre = False
-        self.inul = False
-        self.initem = False
-        self.item = u''
+        self.curdata = u''

         self.textwidth = textwidth
         self.textwidth = textwidth

+        self.opentags = []
+        self.indentlevel = 0
+        self.ignorenodata = False
+        self.listcount = []
+        self.urls = []

         HTMLParser.__init__(self)
 
     def handle_starttag(self, tag, attrs):
         HTMLParser.__init__(self)
 
     def handle_starttag(self, tag, attrs):

-        if tag.lower() == "h1":
-            self.inheadingone = True
-            self.inparagraph = False
-        elif tag.lower() == "h2":
-            self.inheadingtwo = True
-            self.inparagraph = False
-        elif tag.lower() in ["h3", "h4", "h5", "h6"]:
-            self.inotherheading = True
-            self.inparagraph = False
-        elif tag.lower() == "a":
-            self.inlink = True
-        elif tag.lower() == "br":
-            self.handle_br()
-        elif tag.lower() == "blockquote":
-            self.inblockquote = True
-            self.text = self.text + u'\n'
-        elif tag.lower() == "p":
-            if self.text != "":
-                self.text = self.text + u'\n\n'
-            if self.inparagraph:
-                self.text = self.text \
-                    + u'\n'.join(textwrap.wrap(self.currentparagraph, self.textwidth))
-            self.currentparagraph = u''
-            self.inparagraph = True
-        elif tag.lower() == "pre":
-            self.text = self.text + "\n"
-            self.inpre = True
-            self.inparagraph = False
-            self.inblockquote = False
-        elif tag.lower() == "ul":
-            self.item = u''
-            self.inul = True
-            self.text = self.text + "\n"
-        elif tag.lower() == "li":
-            if not self.initem:
-                self.initem = True
-                self.item = u''
+        tag_name = tag.lower()
+        if tag_name in self.blockleveltags:
+            # handle starting a new block - unless we're in a block element
+            # that can contain other blocks, we'll assume that we want to close
+            # the container
+            if len(self.opentags) > 1 and self.opentags[-1] == u'li':
+                self.handle_curdata()
+
+            if tag_name == u'ol':
+                self.handle_curdata()
+                self.listcount.append(1)
+                self.listlevel = len(self.listcount) - 1
+
+            if tag_name in self.liststarttags:
+                smallist = self.opentags[-3:-1]
+                smallist.reverse()
+                for prev_listtag in smallist:
+                    if prev_listtag in [u'dl', u'ol']:
+                        self.indentlevel = self.indentlevel + 4
+                        break
+                    elif prev_listtag == u'ul':
+                        self.indentlevel = self.indentlevel + 3
+                        break
+
+            if len(self.opentags) > 0:
+                self.handle_curdata()
+                if tag_name not in self.cancontainflow:
+                    self.opentags.pop()
+            self.opentags.append(tag_name)
+        else:
+            if tag_name == "span":
+                return
+            listcount = 0
+            try:
+                listcount = self.listcount[-1]
+            except:
+                pass
+
+            if tag_name == u'dd' and len(self.opentags) > 1 \
+                and self.opentags[-1] == u'dt':
+                self.handle_curdata()
+                self.opentags.pop()
+            elif tag_name == u'dt' and len(self.opentags) > 1 \
+                and self.opentags[-1] == u'dd':
+                self.handle_curdata()
+                self.opentags.pop()
+            elif tag_name == u'a':
+                for attr in attrs:
+                    if attr[0].lower() == u'href':
+                        self.urls.append(attr[1])
+                self.curdata = self.curdata + u'`'
+                self.opentags.append(tag_name)
+                return
+            elif tag_name == u'img':
+                self.handle_image(attrs)
+                return
+            elif tag_name == u'br':
+                self.handle_br()
+                return

             else:
             else:

-                self.text = self.text \
-                    + u' * ' \
-                    + u'\n   '.join([a.strip() for a in \
-                        textwrap.wrap(self.item, self.textwidth - 3)]) \
-                    + u'\n'
-                self.item = u''
-                self.initem = True
+                # we don't know the tag, so lets avoid handling it!
+                return 

 
     def handle_startendtag(self, tag, attrs):
 
     def handle_startendtag(self, tag, attrs):

-        if tag.lower() == "br":
+        if tag.lower() == u'br':

             self.handle_br()
             self.handle_br()

+        elif tag.lower() == u'img':
+            self.handle_image(attrs)
+            return

 
     def handle_br(self):
 
     def handle_br(self):

-            if self.inparagraph:
-                self.text = self.text \
-                + u'\n'.join( \
-                    [a \
-                        for a in textwrap.wrap( \
-                            self.currentparagraph, self.textwidth) \
-                    ] \
-                ) \
-                + u'\n'
-                self.currentparagraph = u''
-            elif self.inblockquote:
-                self.text = self.text \
-                    + u'\n> ' \
-                    + u'\n> '.join( \
-                        [a \
-                            for a in textwrap.wrap( \
-                                self.blockquote.encode("utf-8") \
-                                , 68) \
-                        ] \
+            self.handle_curdata()
+            self.opentags.append(u'br')
+            self.handle_curdata()
+            self.opentags.pop()
+
+    def handle_image(self, attrs):
+        alt = u''
+        url = u''
+        for attr in attrs:
+            if attr[0] == 'alt':
+                alt = attr[1]
+            elif attr[0] == 'src':
+                url = attr[1]
+        if url:
+            self.curdata = self.curdata \
+                + u' [img:' \
+                + unicode( \
+                    url.encode('utf-8'), \
+                    'utf-8')
+            if alt:
+                self.curdata = self.curdata \
+                    + u'(' \
+                    + unicode( \
+                        alt.encode('utf-8'), \
+                        'utf-8') \
+                    + u')'
+            self.curdata = self.curdata \
+                + u']'
+
+    def handle_curdata(self):
+
+        if len(self.opentags) == 0:
+            return
+
+        tag_thats_done = self.opentags[-1]
+
+        if len(self.curdata) == 0:
+            return
+
+        if tag_thats_done == u'br':
+            if len(self.text) == 0 or self.text[-1] != '\n':
+                self.text = self.text + '\n'
+                self.ignorenodata = True
+            return
+
+        if len(self.curdata.strip()) == 0:
+            return
+
+        if tag_thats_done in self.blockleveltags:
+            newlinerequired = self.text != u''
+            if self.ignorenodata:
+                newlinerequired = False
+            self.ignorenodata = False
+            if newlinerequired \
+                and len(self.text) > 2 \
+                and self.text[-1] != u'\n' \
+                and self.text[-2] != u'\n':
+                    self.text = self.text + u'\n\n'
+
+        if tag_thats_done in ["h1", "h2", "h3", "h4", "h5", "h6"]:
+            underline = u''
+            underlinechar = u'='
+            headingtext = unicode( \
+                self.curdata.encode("utf-8").strip(), "utf-8")
+            seperator = u'\n' + u' '*self.indentlevel
+            headingtext = seperator.join( \
+                textwrap.wrap( \
+                    headingtext, \
+                    self.textwidth - self.indentlevel \

                     ) \
                     ) \

-                    + u'\n'
-                self.blockquote = u''
-            else:
-                self.text = self.text + "\n"
+                )

 
 

-    def handle_endtag(self, tag):
-        if tag.lower() == "h1":
-            self.inheadingone = False
-            self.text = self.text \
-                + u'\n\n' \
-                + self.headingtext.encode("utf-8") \
-                + u'\n' \
-                + u'=' * len(self.headingtext.encode("utf-8").strip())
-            self.headingtext = u''
-        elif tag.lower() == "h2":
-            self.inheadingtwo = False
+            if tag_thats_done == u'h2':
+                underlinechar = u'-'
+            elif tag_thats_done != u'h1':
+                underlinechar = u'~'
+
+            if u'\n' in headingtext:
+                underline = u' ' * self.indentlevel \
+                    + underlinechar * (self.textwidth - self.indentlevel)
+            else:
+                underline = u' ' * self.indentlevel \
+                    + underlinechar * len(headingtext)

             self.text = self.text \
             self.text = self.text \

-                + u'\n\n' \
-                + self.headingtext.encode("utf-8") \
-                + u'\n' \
-                + u'-' * len(self.headingtext.encode("utf-8").strip())
-            self.headingtext = u''
-        elif tag.lower() in ["h3", "h4", "h5", "h6"]:
-            self.inotherheading = False
+                + headingtext.encode("utf-8") + u'\n' \
+                + underline
+        elif tag_thats_done in [u'p', u'div']:
+            paragraph = unicode( \
+                self.curdata.strip().encode("utf-8"), "utf-8")
+            seperator = u'\n' + u' ' * self.indentlevel

             self.text = self.text \
             self.text = self.text \

-                + u'\n\n' \
-                + self.headingtext.encode("utf-8") \
-                + u'\n' \
-                + u'~' * len(self.headingtext.encode("utf-8").strip())
-            self.headingtext = u''
-        elif tag.lower() == "p":
+                + u' ' * self.indentlevel \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        paragraph, self.textwidth - self.indentlevel))
+        elif tag_thats_done == "pre":
+            self.text = self.text + unicode( \
+                self.curdata.encode("utf-8"), "utf-8")
+        elif tag_thats_done == u'blockquote':
+            quote = unicode( \
+                self.curdata.encode("utf-8").strip(), "utf-8")
+            seperator = u'\n' + u' ' * self.indentlevel + u'> '
+            if len(self.text) > 0 and self.text[-1] != u'\n':
+                self.text = self.text + u'\n'

             self.text = self.text \
             self.text = self.text \

-                + u'\n'.join(textwrap.wrap( \
-                    self.currentparagraph, self.textwidth) \
+                + u'> ' \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        quote, \
+                        self.textwidth - self.indentlevel - 2 \
+                    )

                 )
                 )

-            self.inparagraph = False
-            self.currentparagraph = u''
-        elif tag.lower() == "blockquote":
+            self.curdata = u''
+        elif tag_thats_done == "li":
+            item = unicode(self.curdata.encode("utf-8").strip(), "utf-8")
+            if len(self.text) > 0 and self.text[-1] != u'\n':
+                self.text = self.text + u'\n'
+            # work out if we're in an ol rather than a ul
+            latesttags = self.opentags[-4:]
+            latesttags.reverse()
+            isul = None
+            for thing in latesttags:
+                if thing == 'ul':
+                    isul = True
+                    break
+                elif thing == 'ol':
+                    isul = False
+                    break
+
+            listindent = 3
+            if not isul:
+                listindent = 4
+
+            listmarker = u' * '
+            if isul == False:
+                listmarker = u' %2d. ' %(self.listcount[-1])
+                self.listcount[-1] = self.listcount[-1] + 1
+
+            seperator = u'\n' \
+                + u' ' * self.indentlevel \
+                + u' ' * listindent

             self.text = self.text \
             self.text = self.text \

-                + u'\n> ' \
-                + u'\n> '.join( \
-                    [a.strip() \
-                        for a in textwrap.wrap( \
-                            self.blockquote, self.textwidth - 2)] \
+                + u' ' * self.indentlevel \
+                + listmarker \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        item, \
+                        self.textwidth - self.indentlevel - listindent \

                     ) \
                     ) \

-                + u'\n'
-            self.inblockquote = False
-            self.blockquote = u''
-        elif tag.lower() == "pre":
-            self.inpre = False
-        elif tag.lower() == "li":
-            self.initem = False
-            if self.item != u'':
+                )
+            self.curdata = u''
+        elif tag_thats_done == u'dt':
+            definition = unicode(self.curdata.encode("utf-8").strip(), "utf-8")
+            if len(self.text) > 0 and self.text[-1] != u'\n':
+                self.text = self.text + u'\n\n'
+            elif len(self.text) > 1 and self.text[-2] != u'\n':
+                self.text = self.text + u'\n'
+            definition = u' ' * self.indentlevel + definition + "::"
+            indentstring = u'\n' + u' ' * (self.indentlevel + 1)
+            self.text = self.text \
+                + indentstring.join(
+                    textwrap.wrap(definition, \
+                        self.textwidth - self.indentlevel - 1))
+            self.curdata = u''
+        elif tag_thats_done == u'dd':
+            definition = unicode(self.curdata.encode("utf-8").strip(), "utf-8")
+            if len(definition) > 0:
+                if len(self.text) > 0 and self.text[-1] != u'\n':
+                    self.text = self.text + u'\n'
+                indentstring = u'\n' + u' ' * (self.indentlevel + 4)

                 self.text = self.text \
                 self.text = self.text \

-                    + u' * ' \
-                    + u'\n   '.join( \
-                        [a.strip() for a in textwrap.wrap(self.item, self.textwidth - 3)]) \
-                    + u'\n'
-            self.item = u''
-        elif tag.lower() == "ul":
-            self.inul = False
+                    + u' ' * (self.indentlevel + 4) \
+                    + indentstring.join( \
+                        textwrap.wrap( \
+                            definition, \
+                            self.textwidth - self.indentlevel - 4 \
+                            ) \
+                        )
+                self.curdata = u''
+        elif tag_thats_done == u'a':
+            self.curdata = self.curdata + u'`__'
+            pass
+        elif tag_thats_done in self.liststarttags:
+            pass

 
 

-    def handle_data(self, data):
-        if self.inheadingone or self.inheadingtwo or self.inotherheading:
-            self.headingtext = self.headingtext \
-                + unicode(data, "utf-8").strip() \
-                + u' '
-        elif self.inblockquote:
-            self.blockquote = self.blockquote \
-                + unicode(data, "utf-8").strip() \
-                + u' '
-        elif self.initem:
-            self.item = self.item + unicode(data, "utf-8")
-        elif self.inparagraph:
-            self.currentparagraph = self.currentparagraph \
-                + unicode(data, "utf-8").strip() \
-                + u' '
-        elif self.inpre:
-            self.text = self.text + unicode(data, "utf-8")
+        if tag_thats_done in self.blockleveltags:
+            self.curdata = u''
+
+        self.ignorenodata = False
+
+    def handle_endtag(self, tag):
+        self.ignorenodata = False
+        if tag == "span":
+            return
+
+        try:
+            tagindex = self.opentags.index(tag)
+        except:
+            return
+        tag = tag.lower()
+
+        if tag in [u'br', u'img']:
+            return
+
+        if tag in self.liststarttags:
+            if tag in [u'ol', u'dl', u'ul']:
+                self.handle_curdata()
+                # find if there was a previous list level
+                smalllist = self.opentags[:-1]
+                smalllist.reverse()
+                for prev_listtag in smalllist:
+                    if prev_listtag in [u'ol', u'dl']:
+                        self.indentlevel = self.indentlevel - 4
+                        break
+                    elif prev_listtag == u'ul':
+                        self.indentlevel = self.indentlevel - 3
+                        break
+
+        if tag == u'ol':
+            self.listcount = self.listcount[:-1]
+
+        while tagindex < len(self.opentags) \
+            and tag in self.opentags[tagindex+1:]:
+            try:
+                tagindex = self.opentags.index(tag, tagindex+1)
+            except:
+                # well, we don't want to do that then
+                pass
+        if tagindex != len(self.opentags) - 1:
+            # Assuming the data was for the last opened tag first
+            self.handle_curdata()
+            # Now kill the list to be a slice before this tag was opened
+            self.opentags = self.opentags[:tagindex + 1]

         else:
         else:

-            isallwhitespace = data.strip() == ""
-            if not isallwhitespace:
-                self.text = self.text + unicode(data, "utf-8").strip() + u' ' 
+            self.handle_curdata()
+            if self.opentags[-1] == tag:
+                self.opentags.pop()
+
+    def handle_data(self, data):
+        if len(self.opentags) == 0:
+            self.opentags.append(u'p')
+        self.curdata = self.curdata + unicode(data, "utf-8")

 
     def handle_entityref(self, name):
         entity = name
 
     def handle_entityref(self, name):
         entity = name


@@ -241,21 +422,21 @@ class HTML2Text(HTMLParser):
         else:
             entity = "&" + name + ";"
 
         else:
             entity = "&" + name + ";"
 

-        if self.inparagraph:
-            self.currentparagraph = self.currentparagraph \
-                + unicode(entity, "utf-8")
-        elif self.inblockquote:
-            self.blockquote = self.blockquote + unicode(entity, "utf-8")
-        else:
-            self.text = self.text + unicode(entity, "utf-8")
+        self.curdata = self.curdata + unicode(entity, "utf-8")

 
     def gettext(self):
 
     def gettext(self):

-        data = self.text
-        if self.inparagraph:
-            data = data + "\n".join(textwrap.wrap(self.currentparagraph, self.textwidth))
-        if data[-1] != '\n':
-            data = data + '\n'
-        return data
+        self.handle_curdata()
+        if len(self.text) == 0 or self.text[-1] != u'\n':
+            self.text = self.text + u'\n'
+        self.opentags = []
+        if len(self.text) > 0:
+            while len(self.text) > 1 and self.text[-1] == u'\n':
+                self.text = self.text[:-1]
+            self.text = self.text + u'\n'
+        if len(self.urls) > 0:
+            self.text = self.text + u'\n__ ' + u'\n__ '.join(self.urls) + u'\n'
+            self.urls = []
+        return self.text

 
 def open_url(method, url):
     redirectcount = 0
 
 def open_url(method, url):
     redirectcount = 0


@@ -460,7 +641,8 @@ def parse_and_deliver(maildir, url, statedir):
     if headers:
         data = []
         for header in headers:
     if headers:
         data = []
         for header in headers:

-            if header[0] in ["content-md5", "etag", "last-modified", "content-length"]:
+            if header[0] in \
+                ["content-md5", "etag", "last-modified", "content-length"]:

                 data.append((header[0], header[1]))
         if len(data) > 0:
             data = urllib.urlencode(data)
                 data.append((header[0], header[1]))
         if len(data) > 0:
             data = urllib.urlencode(data)