def train(self, req, log_id, spam=True):
environ = {}
for name, value in req.environ.items():
if not name.startswith('HTTP_'):
environ[name] = value
entry = LogEntry.fetch(self.env, log_id)
if entry:
self.log.debug('Marking as %s: %r submitted by "%s"',
spam and 'spam' or 'ham',
shorten_line(entry.content),
entry.author)
fakeenv = environ.copy()
for header in entry.headers.splitlines():
name, value = header.split(':', 1)
if name == 'Cookie': # breaks SimpleCookie somehow
continue
cgi_name = 'HTTP_%s' % name.strip().replace('-', '_').upper()
fakeenv[cgi_name] = value.strip()
fakeenv['REQUEST_METHOD'] = 'POST'
fakeenv['PATH_INFO'] = entry.path
fakeenv['wsgi.input'] = StringIO('')
fakeenv['REMOTE_ADDR'] = entry.ipnr
if entry.authenticated:
fakeenv['REMOTE_USER'] = entry.author
for strategy in self.strategies:
strategy.train(Request(fakeenv, None),
entry.author or 'anonymous',
entry.content, spam=spam)
entry.update(rejected=spam)
def _render_monitoring_entry(self, req, cat, page, entry_id):
req.perm.require('SPAM_MONITOR')
entry = LogEntry.fetch(self.env, entry_id)
if not entry:
raise HTTPNotFound(_("Log entry not found"))
previous = entry.get_previous()
if previous:
add_link(req, 'prev', req.href.admin(cat, page, previous.id),
_("Log Entry %(id)s", id=previous.id))
add_link(req, 'up', req.href.admin(cat, page), _("Log Entry List"))
next = entry.get_next()
if next:
add_link(req, 'next', req.href.admin(cat, page, next.id),
_("Log Entry %(id)s", id=next.id))
return {'entry': entry, 'train_only': self.train_only}
def delete(self, req, ids, stats):
if not isinstance(ids, list):
ids = [ids]
if stats:
for log_id in ids:
entry = LogEntry.fetch(self.env, log_id)
if entry:
type = "spam" if entry.rejected else "ham"
status = "delete"
for strategy in self.strategies:
count = entry.findreasonvalue(get_strategy_name(strategy))
if count:
spamstatus = count < 0
self._record_action(status, type, ("ok" if spamstatus == entry.rejected else "error"), strategy, 0)
else:
self._record_action(status, type, '', strategy, 0)
self._record_action(status, type, '', '', 0)
LogEntry.delete(self.env, ids)
def _render_monitoring_entry(self, req, cat, page, entry_id):
req.perm.assert_permission('SPAM_MONITOR')
entry = LogEntry.fetch(self.env, entry_id)
if not entry:
raise HTTPNotFound('Log entry not found')
previous = entry.get_previous()
if previous:
add_link(req, 'prev', req.href.admin(cat, page, previous.id),
'Log Entry %d' % previous.id)
add_link(req, 'up', req.href.admin(cat, page), 'Log Entry List')
next = entry.get_next()
if next:
add_link(req, 'next', req.href.admin(cat, page, next.id),
'Log Entry %d' % next.id)
return {'entry': entry}
def train(self, req, ids, spam=True, delete=False):
environ = {}
for name, value in req.environ.items():
if not name.startswith('HTTP_'):
environ[name] = value
if not isinstance(ids, list):
ids = [ids]
for log_id in ids:
start = time.time()
entry = LogEntry.fetch(self.env, log_id)
if entry:
extint = "trainint"
self.log.debug('Marking as %s: %r submitted by "%s"',
spam and 'spam' or 'ham',
shorten_line(entry.content),
entry.author)
fakeenv = environ.copy()
for header in entry.headers.splitlines():
name, value = header.split(':', 1)
if name == 'Cookie': # breaks SimpleCookie somehow
continue
cgi_name = 'HTTP_%s' % name.strip().replace('-', '_').upper()
fakeenv[cgi_name] = value.strip()
fakeenv['REQUEST_METHOD'] = 'POST'
fakeenv['PATH_INFO'] = entry.path
fakeenv['wsgi.input'] = StringIO('')
fakeenv['REMOTE_ADDR'] = entry.ipnr
if entry.authenticated:
fakeenv['REMOTE_USER'] = entry.author
type = "spam" if spam else "ham"
for strategy in self.strategies:
status = "trainskip"
if (self.use_external and self.train_external) or not strategy.is_external():
tim = time.time()
extint = "trainext"
res = strategy.train(Request(fakeenv, None),
entry.author or 'anonymous',
entry.content, entry.ipnr, spam=spam)
tim = time.time()-tim
if tim > 3:
self.log.warn('Training %s took %d seconds to '
'complete.', strategy, tim)
if res == -1:
status = "trainerror"
elif res == -2:
status = "traincond"
elif res > 0:
status = "train"
count = entry.findreasonvalue(get_strategy_name(strategy))
if count:
spamstatus = count < 0
self._record_action(status, type, ("ok" if spamstatus == spam else "error"), strategy, tim)
else:
self._record_action(status, type, '', strategy, tim)
self._record_action(extint, type, ("ok" if entry.rejected == spam else "error"), '', time.time()-start)
entry.update(rejected=spam)
if delete:
self.delete(req, log_id, True)
