X-Git-Url: https://git.sommitrealweird.co.uk/rss2maildir.git/blobdiff_plain/286e096cd216854c5a9bc8cdc580e5d3ec447f19..9337881574f8f9370fecabfeb9a52e4341568cd9:/rss2maildir.py

diff --git a/rss2maildir.py b/rss2maildir.py
index 739c1f3..9473dd0 100755
--- a/rss2maildir.py
+++ b/rss2maildir.py
@@ -48,42 +48,160 @@ from HTMLParser import HTMLParser
 
 class HTML2Text(HTMLParser):
     entities = {
-        "amp": "&",
-        "lt": "<",
-        "gt": ">",
-        "pound": "£",
-        "copy": "©",
-        "apos": "'",
-        "quot": "\"",
-        "nbsp": " ",
+        u'amp': u'&',
+        u'lt': u'<',
+        u'gt': u'>',
+        u'pound': u'£',
+        u'copy': u'©',
+        u'apos': u'\'',
+        u'quot': u'"',
+        u'nbsp': u' ',
+        u'ldquo': u'“',
+        u'rdquo': u'”',
+        u'lsquo': u'‘',
+        u'rsquo': u'’',
+        u'laquo': u'«',
+        u'raquo': u'»',
+        u'lsaquo': u'‹',
+        u'rsaquo': u'›',
+        u'bull': u'•',
+        u'middot': u'·',
+        u'deg': u'°',
+        u'helip': u'…',
+        u'trade': u'™',
+        u'reg': u'®',
+        u'agrave': u'à',
+        u'Agrave': u'À',
+        u'egrave': u'è',
+        u'Egrave': u'È',
+        u'igrave': u'ì',
+        u'Igrave': u'Ì',
+        u'ograve': u'ò',
+        u'Ograve': u'Ò',
+        u'ugrave': u'ù',
+        u'Ugrave': u'Ù',
+        u'aacute': u'á',
+        u'Aacute': u'Á',
+        u'eacute': u'é',
+        u'Eacute': u'É',
+        u'iacute': u'í',
+        u'Iacute': u'Í',
+        u'oacute': u'ó',
+        u'Oacute': u'Ó',
+        u'uacute': u'ú',
+        u'Uacute': u'Ú',
+        u'yactue': u'ý',
+        u'Yacute': u'Ý',
+        u'acirc': u'â',
+        u'Acirc': u'Â',
+        u'ecirc': u'ê',
+        u'Ecirc': u'Ê',
+        u'icirc': u'î',
+        u'Icirc': u'Î',
+        u'ocirc': u'ô',
+        u'Ocirc': u'Ô',
+        u'ucirc': u'û',
+        u'Ucirc': u'Û',
+        u'atilde': u'ã',
+        u'Atilde': u'Ã',
+        u'ntilde': u'ñ',
+        u'Ntilde': u'Ñ',
+        u'otilde': u'õ',
+        u'Otilde': u'Õ',
+        u'auml': u'ä',
+        u'Auml': u'Ä',
+        u'euml': u'ë',
+        u'Euml': u'Ë',
+        u'iuml': u'ï',
+        u'Iuml': u'Ï',
+        u'ouml': u'ö',
+        u'Ouml': u'Ö',
+        u'uuml': u'ü',
+        u'Uuml': u'Ü',
+        u'yuml': u'ÿ',
+        u'Yuml': u'Ÿ',
+        u'iexcl': u'¡',
+        u'iquest': u'¿',
+        u'ccedil': u'ç',
+        u'Ccedil': u'Ç',
+        u'oelig': u'œ',
+        u'OElig': u'Œ',
+        u'szlig': u'ß',
+        u'oslash': u'ø',
+        u'Oslash': u'Ø',
+        u'aring': u'Ã¥',
+        u'Aring': u'Å',
+        u'aelig': u'æ',
+        u'AElig': u'Æ',
+        u'thorn': u'þ',
+        u'THORN': u'Þ',
+        u'eth': u'ð',
+        u'ETH': u'Ð',
+        u'mdash': u'—',
+        u'ndash': u'–',
+        u'sect': u'§',
+        u'para': u'¶',
+        u'uarr': u'↑',
+        u'darr': u'↓',
+        u'larr': u'←',
+        u'rarr': u'→',
+        u'dagger': u'†',
+        u'Dagger': u'‡',
+        u'permil': u'‰',
+        u'prod': u'∏',
+        u'infin': u'∞',
+        u'radic': u'√',
+        u'there4': u'∴',
+        u'int': u'∫',
+        u'asymp': u'≈',
+        u'ne': u'≠',
+        u'equiv': '≡',
+        u'le': u'≤',
+        u'ge': u'≥',
+        u'loz': u'⋄',
+        u'sum': u'∑',
+        u'part': u'∂',
+        u'prime': u'′',
+        u'Prime': u'″',
+        u'harr': u'↔',
+        u'micro': u'µ',
+        u'not': u'¬',
+        u'plusmn': u'±',
+        u'divide': u'÷',
+        u'cent': u'¢',
+        u'euro': u'€',
         }
 
     blockleveltags = [
-        "h1",
-        "h2",
-        "h3",
-        "h4",
-        "h5",
-        "h6",
-        "pre",
-        "p",
-        "ul",
-        "ol",
-        "dl",
-        "br",
+        u'h1',
+        u'h2',
+        u'h3',
+        u'h4',
+        u'h5',
+        u'h6',
+        u'pre',
+        u'p',
+        u'ul',
+        u'ol',
+        u'dl',
+        u'li',
+        u'dt',
+        u'dd',
+        u'div',
+        u'blockquote',
         ]
 
     liststarttags = [
-        "ul",
-        "ol",
-        "dl",
+        u'ul',
+        u'ol',
+        u'dl',
         ]
 
     cancontainflow = [
-        "div",
-        "li",
-        "dd",
-        "blockquote",
+        u'div',
+        u'li',
+        u'dd',
+        u'blockquote',
     ]
 
     def __init__(self,textwidth=70):
@@ -92,6 +210,10 @@ class HTML2Text(HTMLParser):
         self.textwidth = textwidth
         self.opentags = []
         self.indentlevel = 0
+        self.ignorenodata = False
+        self.listcount = []
+        self.urls = []
+        self.images = {}
         HTMLParser.__init__(self)
 
     def handle_starttag(self, tag, attrs):
@@ -100,42 +222,159 @@ class HTML2Text(HTMLParser):
             # 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 tag_name == u'br':
+            if len(self.opentags) > 1 and self.opentags[-1] == u'li':
                 self.handle_curdata()
-                self.opentags.append(tag_name)
-                self.opentags.pop()
+
+            if tag_name == u'ol':
+                self.handle_curdata()
+                self.listcount.append(1)
+                self.listlevel = len(self.listcount) - 1
+
+            if tag_name == u'dl':
+                self.indentlevel = self.indentlevel + 4
+
+            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()
-                self.opentags.pop()
+                if tag_name not in self.cancontainflow:
+                    self.opentags.pop()
             self.opentags.append(tag_name)
         else:
-            self.handle_curdata()
-            self.opentags.append(tag_name)
+            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].decode('utf-8'))
+                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:
+                # we don't know the tag, so lets avoid handling it!
+                return 
 
     def handle_startendtag(self, tag, attrs):
         if tag.lower() == u'br':
-            self.tags.append(u'br')
-            self.handle_curdata() # just handle the data, don't do anything else
-            self.tags.pop()
+            self.handle_br()
+        elif tag.lower() == u'img':
+            self.handle_image(attrs)
+            return
+
+    def handle_br(self):
+            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].decode('utf-8')
+            elif attr[0] == 'src':
+                url = attr[1].decode('utf-8')
+        if url:
+            if alt:
+                if self.images.has_key(alt):
+                    if self.images[alt]["url"] == url:
+                        self.curdata = self.curdata \
+                            + u'|%s|' %(alt,)
+                    else:
+                        while self.images.has_key(alt):
+                            alt = alt + "_"
+                        self.images[alt] = {"url": url}
+                        self.curdata = self.curdata \
+                            + u'|%s|' %(alt,)
+                else:
+                    self.images[alt] = {"url": url}
+                    self.curdata = self.curdata \
+                        + u'|%s|' %(alt,)
+            else:
+                if self.images.has_key(url):
+                    self.curdata = self.curdata \
+                        + u'|%s|' %(url,)
+                else:
+                    self.images[url] = {}
+                    self.images[url]["url"] =url
+                    self.curdata = self.curdata \
+                        + u'|%s|' %(url,)
 
     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:
-                self.text = self.text + u'\n\n'
+                if tag_thats_done in [u'dt', u'dd', u'li'] \
+                    and len(self.text) > 1 \
+                    and self.text[-1] != u'\n':
+                        self.text = self.text + u'\n'
+                elif 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 = self.curdata.encode("utf-8").strip()
-            headingtext = u'\n'.join( \
-                textwrap.wrap(headingtext, self.textwidth))
+            headingtext = " ".join(self.curdata.split())
+            seperator = u'\n' + u' '*self.indentlevel
+            headingtext = seperator.join( \
+                textwrap.wrap( \
+                    headingtext, \
+                    self.textwidth - self.indentlevel \
+                    ) \
+                )
 
             if tag_thats_done == u'h2':
                 underlinechar = u'-'
@@ -143,54 +382,157 @@ class HTML2Text(HTMLParser):
                 underlinechar = u'~'
 
             if u'\n' in headingtext:
-                underline = underlinechar * self.textwidth
+                underline = u' ' * self.indentlevel \
+                    + underlinechar * (self.textwidth - self.indentlevel)
             else:
-                underline = underlinechar * len(headingtext)
+                underline = u' ' * self.indentlevel \
+                    + underlinechar * len(headingtext)
             self.text = self.text \
-                + headingtext.encode("utf-8") + u'\n' \
+                + headingtext + u'\n' \
                 + underline
-        elif tag_thats_done == "p":
-            paragraph = self.curdata.encode("utf-8").strip()
+        elif tag_thats_done in [u'p', u'div']:
+            paragraph = unicode( \
+                " ".join(self.curdata.strip().encode("utf-8").split()), \
+                "utf-8")
+            seperator = u'\n' + u' ' * self.indentlevel
             self.text = self.text \
-                + u'\n'.join(textwrap.wrap(paragraph, self.textwidth))
+                + u' ' * self.indentlevel \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        paragraph, self.textwidth - self.indentlevel))
         elif tag_thats_done == "pre":
-            self.text = self.text + self.curdata
-        elif tag_thats_done == "blockquote":
-            quote = self.curdata.encode("utf-8").strip()
+            self.text = self.text + unicode( \
+                self.curdata.encode("utf-8"), "utf-8")
+        elif tag_thats_done == u'blockquote':
+            quote = unicode( \
+                " ".join(self.curdata.encode("utf-8").strip().split()), \
+                "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 \
-                + u'> ' \
-                + u'> '.join(textwrap.wrap(quote, self.textwidth - 2))
+                + u'    ' \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        quote, \
+                        self.textwidth - self.indentlevel - 2 \
+                    )
+                )
+            self.curdata = u''
         elif tag_thats_done == "li":
-            item = self.curdata.encode("utf-8").strip()
+            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 \
+                + u' ' * self.indentlevel \
+                + listmarker \
+                + seperator.join( \
+                    textwrap.wrap( \
+                        item, \
+                        self.textwidth - self.indentlevel - listindent \
+                    ) \
+                )
+            self.curdata = u''
+        elif tag_thats_done == u'dt':
+            definition = unicode(" ".join( \
+                    self.curdata.encode("utf-8").strip().split()), \
+                "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 - 4) + definition + "::"
+            indentstring = u'\n' + u' ' * (self.indentlevel - 3)
             self.text = self.text \
-                + u' * ' \
-                + u'\n   '.join( \
-                    textwrap.wrap(item, self.textwidth - 3))
+                + indentstring.join(
+                    textwrap.wrap(definition, \
+                        self.textwidth - self.indentlevel - 4))
             self.curdata = u''
+        elif tag_thats_done == u'dd':
+            definition = unicode(" ".join( \
+                    self.curdata.encode("utf-8").strip().split()),
+                "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
+                self.text = self.text \
+                    + indentstring \
+                    + indentstring.join( \
+                        textwrap.wrap( \
+                            definition, \
+                            self.textwidth - self.indentlevel \
+                            ) \
+                        )
+                self.curdata = u''
+        elif tag_thats_done == u'a':
+            self.curdata = self.curdata + u'`__'
+            pass
         elif tag_thats_done in self.liststarttags:
             pass
-        else:
-            # we've got no idea what this tag does, so we'll
-            # make an assumption that we're not going to know later
-            if len(self.curdata) > 0:
-                self.text = self.text \
-                    + u' ... ' \
-                    + u'\n ... '.join( \
-                        textwrap.wrap(self.curdata, self.textwidth - 5))
-            self.curdata = u''
 
         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:
-            # closing tag we know nothing about.
-            # err. weird.
-            tagindex = 0
+            return
+        tag = tag.lower()
+
+        if tag in [u'br', u'img']:
+            return
+
+        if tag == u'dl':
+            self.indentlevel = self.indentlevel - 4
+
+        if tag in self.liststarttags:
+            if tag in [u'ol', u'dl', u'ul', u'dd']:
+                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:]:
@@ -203,27 +545,50 @@ class HTML2Text(HTMLParser):
             # 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]
+            self.opentags = self.opentags[:tagindex + 1]
+        else:
+            self.handle_curdata()
+            if self.opentags[-1] == tag:
+                self.opentags.pop()
 
     def handle_data(self, data):
-        self.curdata = self.curdata + unicode(data, "utf-8")
+        if len(self.opentags) == 0:
+            self.opentags.append(u'p')
+        self.curdata = self.curdata + data.decode("utf-8")
+
+    def handle_charref(self, name):
+        entity = unichr(int(name))
+        self.curdata = self.curdata + unicode(entity.encode('utf-8'), \
+            "utf-8")
 
     def handle_entityref(self, name):
         entity = name
-        if HTML2Text.entities.has_key(name.lower()):
-            entity = HTML2Text.entities[name.lower()]
-        elif name[0] == "#":
-            entity = unichr(int(name[1:]))
+        if HTML2Text.entities.has_key(name):
+            entity = HTML2Text.entities[name]
         else:
             entity = "&" + name + ";"
 
-        self.curdata = self.curdata + unicode(entity, "utf-8")
+        self.curdata = self.curdata + unicode(entity.encode('utf-8'), \
+            "utf-8")
 
     def gettext(self):
         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 = []
+        if len(self.images.keys()) > 0:
+            self.text = self.text + u'\n.. ' \
+                + u'\n.. '.join( \
+                    ["|%s| image:: %s" %(a, self.images[a]["url"]) \
+                for a in self.images.keys()]) + u'\n'
+            self.images = {}
         return self.text
 
 def open_url(method, url):
@@ -359,7 +724,9 @@ def parse_and_deliver(maildir, url, statedir):
         except:
             pass
         msg.add_header("Date", createddate)
-        msg.add_header("Subject", item["title"])
+        subj_gen = HTML2Text()
+        subj_gen.feed(item["title"].encode("utf-8"))
+        msg.add_header("Subject", subj_gen.gettext())
         msg.set_default_type("text/plain")
 
         htmlcontent = content.encode("utf-8")
@@ -429,7 +796,8 @@ def parse_and_deliver(maildir, url, statedir):
     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)
@@ -508,11 +876,13 @@ if __name__ == "__main__":
     elif scp.has_option("general", "state_dir"):
         new_state_dir = scp.get("general", "state_dir")
         try:
-            mode = os.stat(state_dir)[stat.ST_MODE]
+            mode = os.stat(new_state_dir)[stat.ST_MODE]
             if not stat.S_ISDIR(mode):
                 sys.stderr.write( \
                     "State directory (%s) is not a directory\n" %(state_dir))
                 sys.exit(1)
+            else:
+                state_dir = new_state_dir
         except:
             # try to create it
             try: