def finishtest(self):
if options["TestDriver", "show_histograms"]:
printhist("all in this training set:",
self.trained_ham_hist, self.trained_spam_hist)
self.global_ham_hist += self.trained_ham_hist
self.global_spam_hist += self.trained_spam_hist
self.trained_ham_hist = Hist()
self.trained_spam_hist = Hist()
self.ntimes_finishtest_called += 1
if options["TestDriver", "save_trained_pickles"]:
fname = "%s%d.pik" % (options["TestDriver", "pickle_basename"],
self.ntimes_finishtest_called)
print(" saving pickle to", fname)
pickle_write(fname, self.classifier, 1)
def _save_caches(self):
# XXX Note that these caches are never refreshed, which might not
# XXX be a good thing long-term (if a previously invalid URL
# XXX becomes valid, for example).
for name, data in [(self.bad_url_cache_name, self.bad_urls),
(self.http_error_cache_name, self.http_error_urls),]:
pickle_write(name, data)
def alldone(self):
if options["TestDriver", "show_histograms"]:
besthamcut, bestspamcut = printhist("all runs:",
self.global_ham_hist,
self.global_spam_hist)
else:
besthamcut = options["Categorization", "ham_cutoff"]
bestspamcut = options["Categorization", "spam_cutoff"]
self.global_ham_hist.compute_stats()
self.global_spam_hist.compute_stats()
nham = self.global_ham_hist.n
nspam = self.global_spam_hist.n
nfp = len(self.falsepos)
nfn = len(self.falseneg)
nun = len(self.unsure)
print "-> <stat> all runs false positives:", nfp
print "-> <stat> all runs false negatives:", nfn
print "-> <stat> all runs unsure:", nun
print "-> <stat> all runs false positive %:", (nfp * 1e2 / nham)
print "-> <stat> all runs false negative %:", (nfn * 1e2 / nspam)
print "-> <stat> all runs unsure %:", (nun * 1e2 / (nham + nspam))
print "-> <stat> all runs cost: $%.2f" % (
nfp * options["TestDriver", "best_cutoff_fp_weight"] +
nfn * options["TestDriver", "best_cutoff_fn_weight"] +
nun * options["TestDriver", "best_cutoff_unsure_weight"])
options["Categorization", "ham_cutoff"] = besthamcut
options["Categorization", "spam_cutoff"] = bestspamcut
print self.cc
if options["TestDriver", "save_histogram_pickles"]:
for f, h in (('ham', self.global_ham_hist),
('spam', self.global_spam_hist)):
fname = "%s_%shist.pik" % (options["TestDriver",
"pickle_basename"], f)
print " saving %s histogram pickle to %s" % (f, fname)
pickle_write(fname, h, 1)
def main(args):
try:
opts, args = getopt.getopt(args, "hd:t:",
["type=", "help", "database="])
except getopt.GetoptError as msg:
usage(msg)
return 1
mapfile = None
mboxtype = None
for opt, arg in opts:
if opt in ("-h", "--help"):
usage()
return 0
elif opt in ("-d", "--database"):
mapfile = arg
elif opt in ("-t", "--type"):
mboxtype = arg
if mapfile is None:
usage("'-d mapfile' is required")
return 1
if mboxtype is None:
usage("'-t ham|spam' is required")
return 1
if mboxtype not in ("ham", "spam"):
usage("mboxtype must be 'ham' or 'spam'")
return 1
try:
mapd = pickle_read(mapfile)
except IOError:
mapd = {}
for f in args:
mapmessages(f, mboxtype, mapd)
pickle_write(mapfile, mapd)
def _save_caches(self):
# XXX Note that these caches are never refreshed, which might not
# XXX be a good thing long-term (if a previously invalid URL
# XXX becomes valid, for example).
for name, data in [
(self.bad_url_cache_name, self.bad_urls),
(self.http_error_cache_name, self.http_error_urls),
]:
pickle_write(name, data)
def close(self):
if options["globals", "verbose"]:
print >> sys.stderr, "saving", len(self.cache),
print >> sys.stderr, "items to", self.cachefile,
if self.hits + self.misses:
print >> sys.stderr, "%.2f%% hit rate" % \
(100 * self.hits / (self.hits + self.misses)),
print >> sys.stderr
pickle_write(self.cachefile, self.cache)
def close(self):
if options["globals", "verbose"]:
print("saving", len(self.cache), end=' ', file=sys.stderr)
print("items to", self.cachefile, end=' ', file=sys.stderr)
if self.hits + self.misses:
print("%.2f%% hit rate" % \
(100 * self.hits / (self.hits + self.misses)), end=' ', file=sys.stderr)
print(file=sys.stderr)
pickle_write(self.cachefile, self.cache)
def close(self):
if options["globals", "verbose"]:
print >> sys.stderr, "saving", len(self.cache),
print >> sys.stderr, "items to", self.cachefile,
if self.hits + self.misses:
print >> sys.stderr, "%.2f%% hit rate" % \
(100 * self.hits / (self.hits + self.misses)),
print >> sys.stderr
pickle_write(self.cachefile, self.cache)
def close(self):
if self.printStatsAtEnd:
self.printStats()
if self.cachefile:
pickle_write(self.cachefile, self.caches)
def finishtest(self):
if options["TestDriver", "show_histograms"]:
printhist("all in this training set:", self.trained_ham_hist,
self.trained_spam_hist)
self.global_ham_hist += self.trained_ham_hist
self.global_spam_hist += self.trained_spam_hist
self.trained_ham_hist = Hist()
self.trained_spam_hist = Hist()
self.ntimes_finishtest_called += 1
if options["TestDriver", "save_trained_pickles"]:
fname = "%s%d.pik" % (options["TestDriver", "pickle_basename"],
self.ntimes_finishtest_called)
print " saving pickle to", fname
pickle_write(fname, self.classifier, 1)
def alldone(self):
if options["TestDriver", "show_histograms"]:
besthamcut, bestspamcut = printhist("all runs:",
self.global_ham_hist,
self.global_spam_hist)
else:
besthamcut = options["Categorization", "ham_cutoff"]
bestspamcut = options["Categorization", "spam_cutoff"]
self.global_ham_hist.compute_stats()
self.global_spam_hist.compute_stats()
nham = self.global_ham_hist.n
nspam = self.global_spam_hist.n
nfp = len(self.falsepos)
nfn = len(self.falseneg)
nun = len(self.unsure)
print "-> <stat> all runs false positives:", nfp
print "-> <stat> all runs false negatives:", nfn
print "-> <stat> all runs unsure:", nun
print "-> <stat> all runs false positive %:", (nfp * 1e2 / nham)
print "-> <stat> all runs false negative %:", (nfn * 1e2 / nspam)
print "-> <stat> all runs unsure %:", (nun * 1e2 / (nham + nspam))
print "-> <stat> all runs cost: $%.2f" % (
nfp * options["TestDriver", "best_cutoff_fp_weight"] +
nfn * options["TestDriver", "best_cutoff_fn_weight"] +
nun * options["TestDriver", "best_cutoff_unsure_weight"])
# Set back the options for the delayed calculations in self.cc
options["Categorization", "ham_cutoff"] = besthamcut
options["Categorization", "spam_cutoff"] = bestspamcut
print self.cc
if options["TestDriver", "save_histogram_pickles"]:
for f, h in (('ham', self.global_ham_hist),
('spam', self.global_spam_hist)):
fname = "%s_%shist.pik" % (options["TestDriver",
"pickle_basename"], f)
print " saving %s histogram pickle to %s" % (f, fname)
pickle_write(fname, h, 1)
return 0
elif opt in ("-d", "--database"):
mapfile = arg
elif opt in ("-t", "--type"):
mboxtype = arg
if mapfile is None:
usage("'-d mapfile' is required")
return 1
if mboxtype is None:
usage("'-t ham|spam' is required")
return 1
if mboxtype not in ("ham", "spam"):
usage("mboxtype must be 'ham' or 'spam'")
return 1
try:
mapd = pickle_read(mapfile)
except IOError:
mapd = {}
for f in args:
mapmessages(f, mboxtype, mapd)
pickle_write(mapfile, mapd)
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))
def close(self):
if self.printStatsAtEnd:
self.printStats()
if self.cachefile:
pickle_write(self.cachefile, self.caches)
def main(args):
try:
opts, args = getopt.getopt(args, "b:sh")
except getopt.error as msg:
usage(msg)
return 1
best = 0
skipspam = False
for opt, arg in opts:
if opt == "-h":
usage()
return 0
if opt == "-b":
best = int(arg)
elif opt == "-s":
skipspam = True
if len(args) != 3:
usage("require ham, spam and unsure message piles")
return 1
ham, spam, unsure = args
choices = ["best.pck"]
if "HOME" in os.environ:
home = os.environ["HOME"]
choices.append(os.path.join(home, "tmp", "best.pck"))
choices.append(os.path.join(home, "best.pck"))
choices.append(None)
for bestfile in choices:
if bestfile is None:
break
if os.path.exists(bestfile):
break
try:
file(bestfile, "w")
except IOError:
pass
else:
os.unlink(bestfile)
if bestfile is None:
usage("can't find a place to write best.pck file")
return 1
print("establish base training")
learn(ham, h, False)
learn(spam, h, True)
print("scoring")
if best:
last_scores = pickle_read(bestfile)
last_scores = list(last_scores.items())
last_scores.sort()
msgids = set()
for (k, v) in last_scores[-best:]:
msgids.update(set(v))
else:
msgids = None
scores = {}
try:
score(unsure, h, cls, scores, msgids, skipspam)
except KeyboardInterrupt:
pass
if not best:
pickle_write(bestfile, scores)
return 0
try:
print "Replicating..."
db.Replicate(rdbname)
print "Done"
except pywintypes.com_error:
print "Could not replicate"
if doClassify:
classifyInbox(vinbox, vtrainspam, bayes, ldbname, notesindex, log)
print "The Spambayes database currently has %s Spam and %s Ham" \
% (bayes.nspam, bayes.nham)
bayes.store()
pickle_write(idxname, notesindex)
if log:
log.LogAction("Finished running spambayes")
if __name__ == '__main__':
try:
opts, args = getopt.getopt(sys.argv[1:], 'htcPd:p:l:r:f:o:i:W:L:')
except getopt.error, msg:
print >> sys.stderr, str(msg) + '\n\n' + __doc__
sys.exit()
ldbname = None # local notes database name
rdbname = None # remote notes database location
sbfname = None # spambayes folder name
except IOError:
pass
else:
os.unlink(bestfile)
if bestfile is None:
usage("can't find a place to write best.pck file")
return 1
print "establish base training"
learn(ham, h, False)
learn(spam, h, True)
print "scoring"
if best:
last_scores = pickle_read(bestfile)
last_scores = last_scores.items()
last_scores.sort()
msgids = set()
for (k, v) in last_scores[-best:]:
msgids.update(set(v))
else:
msgids = None
scores = {}
try:
score(unsure, h, cls, scores, msgids, skipspam)
except KeyboardInterrupt:
pass
if not best:
pickle_write(bestfile, scores)
return 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))
try:
print "Replicating..."
db.Replicate(rdbname)
print "Done"
except pywintypes.com_error:
print "Could not replicate"
if doClassify:
classifyInbox(vinbox, vtrainspam, bayes, ldbname, notesindex, log)
print "The Spambayes database currently has %s Spam and %s Ham" \
% (bayes.nspam, bayes.nham)
bayes.store()
pickle_write(idxname, notesindex)
if log:
log.LogAction("Finished running spambayes")
if __name__ == '__main__':
try:
opts, args = getopt.getopt(sys.argv[1:], 'htcPd:p:l:r:f:o:i:W:L:')
except getopt.error, msg:
print >> sys.stderr, str(msg) + '\n\n' + __doc__
sys.exit()
ldbname = None # local notes database name
rdbname = None # remote notes database location
sbfname = None # spambayes folder name
def store(self):
"""Store self as a pickle"""
if options["globals", "verbose"]:
print("Persisting", self.db_name, "as a pickle", file=sys.stderr)
pickle_write(self.db_name, self, PICKLE_TYPE)
def store(self):
pickle_write(self.db_name, self.db, self.mode)
def run(bdbname, useDBM, ldbname, rdbname, foldname, doTrain, doClassify,
pwd, idxname, logname):
bayes = storage.open_storage(bdbname, useDBM)
try:
notesindex = pickle_read(idxname)
except IOError as e:
if e.errno != errno.ENOENT:
raise
notesindex = {}
print("%s file not found, this is a first time run" % (idxname,))
print("No classification will be performed")
need_replicate = False
sess = win32com.client.Dispatch("Lotus.NotesSession")
try:
if pwd:
sess.initialize(pwd)
else:
sess.initialize()
except pywintypes.com_error:
print("Session aborted")
sys.exit()
try:
db = sess.GetDatabase(rdbname, ldbname)
except pywintypes.com_error:
if rdbname:
print("Could not open database remotely, trying locally")
try:
db = sess.GetDatabase("", ldbname)
need_replicate = True
except pywintypes.com_error:
print("Could not open database")
sys.exit()
else:
raise
log = sess.CreateLog("SpambayesAgentLog")
try:
log.OpenNotesLog("", logname)
except pywintypes.com_error:
print("Could not open log")
log = None
if log:
log.LogAction("Running spambayes")
vinbox = db.getView('($Inbox)')
vspam = db.getView("%s\Spam" % (foldname,))
vham = db.getView("%s\Ham" % (foldname,))
vtrainspam = db.getView("%s\Train as Spam" % (foldname,))
vtrainham = db.getView("%s\Train as Ham" % (foldname,))
if doTrain:
processAndTrain(vtrainspam, vspam, bayes, True, notesindex, log)
processAndTrain(vtrainham, vham, bayes, False, notesindex, log)
if need_replicate:
try:
print("Replicating...")
db.Replicate(rdbname)
print("Done")
except pywintypes.com_error:
print("Could not replicate")
if doClassify:
classifyInbox(vinbox, vtrainspam, bayes, ldbname, notesindex, log)
print("The Spambayes database currently has %s Spam and %s Ham" \
% (bayes.nspam, bayes.nham))
bayes.store()
pickle_write(idxname, notesindex)
if log:
log.LogAction("Finished running spambayes")
elif opt in ("-d", "--database"):
mapfile = arg
elif opt in ("-t", "--type"):
mboxtype = arg
if mapfile is None:
usage("'-d mapfile' is required")
return 1
if mboxtype is None:
usage("'-t ham|spam' is required")
return 1
if mboxtype not in ("ham", "spam"):
usage("mboxtype must be 'ham' or 'spam'")
return 1
try:
mapd = pickle_read(mapfile)
except IOError:
mapd = {}
for f in args:
mapmessages(f, mboxtype, mapd)
pickle_write(mapfile, mapd)
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))
def store(self):
pickle_write(self.db_name, self.db, self.mode)
def _save_caches(self):
for name, data in [(self.bad_url_cache_name, self.bad_urls),
(self.http_error_cache_name, self.http_error_urls),]:
pickle_write(name, data)
