def message_keywords(self, session, msg_mid, msg_id, msg, msg_date,
mailbox=None):
keywords = []
textpart = None
for part in msg.walk():
charset = part.get_charset() or 'iso-8859-1'
if part.get_content_type() == 'text/plain':
textpart = self.try_decode(part.get_payload(None, True), charset)
elif part.get_content_type() == 'text/html':
payload = self.try_decode(part.get_payload(None, True), charset)
if len(payload) > 3:
try:
textpart = lxml.html.fromstring(payload).text_content()
except:
session.ui.warning('=%s/%s has bogus HTML.' % (msg_mid, msg_id))
textpart = payload
else:
textpart = payload
elif 'pgp' in part.get_content_type():
keywords.append('pgp:has')
att = part.get_filename()
if att:
att = self.try_decode(att, charset)
keywords.append('attachment:has')
keywords.extend([t+':att' for t in re.findall(WORD_REGEXP, att.lower())])
textpart = (textpart or '') + ' ' + att
if textpart:
# FIXME: Does this lowercase non-ASCII characters correctly?
# FIXME: What about encrypted content?
keywords.extend(re.findall(WORD_REGEXP, textpart.lower()))
# FIXME: Do this better.
if '-----BEGIN PGP' in textpart and '-----END PGP' in textpart:
keywords.append('pgp:has')
mdate = datetime.date.fromtimestamp(msg_date)
keywords.append('%s:year' % mdate.year)
keywords.append('%s:month' % mdate.month)
keywords.append('%s:day' % mdate.day)
keywords.append('%s-%s-%s:date' % (mdate.year, mdate.month, mdate.day))
keywords.append('%s:id' % msg_id)
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'subject').lower()))
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'from').lower()))
if mailbox: keywords.append('%s:mailbox' % mailbox.lower())
keywords.append('%s:hprint' % HeaderPrint(msg))
for key in msg.keys():
key_lower = key.lower()
if key_lower not in BORING_HEADERS:
words = set(re.findall(WORD_REGEXP, self.hdr(msg, key).lower()))
words -= STOPLIST
keywords.extend(['%s:%s' % (t, key_lower) for t in words])
if 'list' in key_lower:
keywords.extend(['%s:list' % t for t in words])
return (set(keywords) - STOPLIST)
def message_keywords(self, session, msg_mid, msg_id, msg, msg_date,
mailbox=None):
keywords = []
textpart = None
payload = [None]
for part in msg.walk():
payload[0] = None
ctype = part.get_content_type()
charset = part.get_charset() or 'iso-8859-1'
def _loader(p):
if payload[0] is None:
payload[0] = self.try_decode(p.get_payload(None, True), charset)
return payload[0]
if ctype == 'text/plain':
textpart = _loader(part)
elif ctype == 'text/html':
_loader(part)
if len(payload[0]) > 3:
try:
textpart = lxml.html.fromstring(payload[0]).text_content()
except:
session.ui.warning('=%s/%s has bogus HTML.' % (msg_mid, msg_id))
textpart = payload[0]
else:
textpart = payload[0]
elif 'pgp' in part.get_content_type():
keywords.append('pgp:has')
att = part.get_filename()
if att:
att = self.try_decode(att, charset)
keywords.append('attachment:has')
keywords.extend([t+':att' for t in re.findall(WORD_REGEXP, att.lower())])
textpart = (textpart or '') + ' ' + att
if textpart:
# FIXME: Does this lowercase non-ASCII characters correctly?
# FIXME: What about encrypted content?
keywords.extend(re.findall(WORD_REGEXP, textpart.lower()))
# FIXME: Do this better.
if '-----BEGIN PGP' in textpart and '-----END PGP' in textpart:
keywords.append('pgp:has')
for extract in plugins.get_text_kw_extractors():
keywords.extend(extract(self, msg, ctype, lambda: textpart))
for extract in plugins.get_data_kw_extractors():
keywords.extend(extract(self, msg, ctype, att, part,
lambda: _loader(part)))
keywords.append('%s:id' % msg_id)
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'subject').lower()))
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'from').lower()))
if mailbox: keywords.append('%s:mailbox' % mailbox.lower())
keywords.append('%s:hprint' % HeaderPrint(msg))
for key in msg.keys():
key_lower = key.lower()
if key_lower not in BORING_HEADERS:
emails = ExtractEmails(self.hdr(msg, key).lower())
words = set(re.findall(WORD_REGEXP, self.hdr(msg, key).lower()))
words -= STOPLIST
keywords.extend(['%s:%s' % (t, key_lower) for t in words])
keywords.extend(['%s:%s' % (e, key_lower) for e in emails])
keywords.extend(['%s:email' % e for e in emails])
if 'list' in key_lower:
keywords.extend(['%s:list' % t for t in words])
for extract in plugins.get_meta_kw_extractors():
keywords.extend(extract(self, msg_mid, msg, msg_date))
return (set(keywords) - STOPLIST)
def message_keywords(self,
session,
msg_mid,
msg_id,
msg,
msg_date,
mailbox=None):
keywords = []
payload = [None]
for part in msg.walk():
textpart = payload[0] = None
ctype = part.get_content_type()
charset = part.get_charset() or 'iso-8859-1'
def _loader(p):
if payload[0] is None:
payload[0] = self.try_decode(p.get_payload(None, True),
charset)
return payload[0]
if ctype == 'text/plain':
textpart = _loader(part)
elif ctype == 'text/html':
_loader(part)
if len(payload[0]) > 3:
try:
textpart = lxml.html.fromstring(
payload[0]).text_content()
except:
session.ui.warning('=%s/%s has bogus HTML.' %
(msg_mid, msg_id))
textpart = payload[0]
else:
textpart = payload[0]
elif 'pgp' in part.get_content_type():
keywords.append('pgp:has')
att = part.get_filename()
if att:
att = self.try_decode(att, charset)
keywords.append('attachment:has')
keywords.extend(
[t + ':att' for t in re.findall(WORD_REGEXP, att.lower())])
textpart = (textpart or '') + ' ' + att
if textpart:
# FIXME: Does this lowercase non-ASCII characters correctly?
# FIXME: What about encrypted content?
keywords.extend(re.findall(WORD_REGEXP, textpart.lower()))
# FIXME: Do this better.
if '-----BEGIN PGP' in textpart and '-----END PGP' in textpart:
keywords.append('pgp:has')
for extract in plugins.get_text_kw_extractors():
keywords.extend(extract(self, msg, ctype,
lambda: textpart))
for extract in plugins.get_data_kw_extractors():
keywords.extend(
extract(self, msg, ctype, att, part,
lambda: _loader(part)))
keywords.append('%s:id' % msg_id)
keywords.extend(
re.findall(WORD_REGEXP,
self.hdr(msg, 'subject').lower()))
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'from').lower()))
if mailbox: keywords.append('%s:mailbox' % mailbox.lower())
keywords.append('%s:hprint' % HeaderPrint(msg))
for key in msg.keys():
key_lower = key.lower()
if key_lower not in BORING_HEADERS:
emails = ExtractEmails(self.hdr(msg, key).lower())
words = set(re.findall(WORD_REGEXP,
self.hdr(msg, key).lower()))
words -= STOPLIST
keywords.extend(['%s:%s' % (t, key_lower) for t in words])
keywords.extend(['%s:%s' % (e, key_lower) for e in emails])
keywords.extend(['%s:email' % e for e in emails])
if 'list' in key_lower:
keywords.extend(['%s:list' % t for t in words])
for extract in plugins.get_meta_kw_extractors():
keywords.extend(extract(self, msg_mid, msg, msg_date))
return (set(keywords) - STOPLIST)
def index_message(self, session, msg_mid, msg_id, msg, msg_date,
mailbox=None, compact=True, filter_hooks=[]):
keywords = []
for part in msg.walk():
charset = part.get_charset() or 'iso-8859-1'
if part.get_content_type() == 'text/plain':
textpart = self.try_decode(part.get_payload(None, True), charset)
elif part.get_content_type() == 'text/html':
payload = self.try_decode(part.get_payload(None, True), charset)
if len(payload) > 3:
try:
textpart = lxml.html.fromstring(payload).text_content()
except:
session.ui.warning('=%s/%s has bogus HTML.' % (msg_mid, msg_id))
textpart = payload
else:
textpart = payload
else:
textpart = None
att = part.get_filename()
if att:
att = self.try_decode(att, charset)
keywords.append('attachment:has')
keywords.extend([t+':att' for t in re.findall(WORD_REGEXP, att.lower())])
textpart = (textpart or '') + ' ' + att
if textpart:
# FIXME: Does this lowercase non-ASCII characters correctly?
keywords.extend(re.findall(WORD_REGEXP, textpart.lower()))
mdate = datetime.date.fromtimestamp(msg_date)
keywords.append('%s:year' % mdate.year)
keywords.append('%s:month' % mdate.month)
keywords.append('%s:day' % mdate.day)
keywords.append('%s-%s-%s:date' % (mdate.year, mdate.month, mdate.day))
keywords.append('%s:id' % msg_id)
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'subject').lower()))
keywords.extend(re.findall(WORD_REGEXP, self.hdr(msg, 'from').lower()))
if mailbox: keywords.append('%s:mailbox' % mailbox.lower())
for key in msg.keys():
key_lower = key.lower()
if key_lower not in BORING_HEADERS:
words = set(re.findall(WORD_REGEXP, self.hdr(msg, key).lower()))
words -= STOPLIST
keywords.extend(['%s:%s' % (t, key_lower) for t in words])
if 'list' in key_lower:
keywords.extend(['%s:list' % t for t in words])
keywords = set(keywords)
keywords -= STOPLIST
for hook in filter_hooks:
keywords = hook(session, msg_mid, msg, keywords)
for word in keywords:
try:
PostingList.Append(session, word, msg_mid, compact=compact)
except UnicodeDecodeError:
# FIXME: we just ignore garbage
pass
return keywords
