def optimize_index(request):
if not users.is_current_user_admin():
return http.HttpResponse("no", status=403)
if request.method == "POST":
term = request.POST.get("term")
elif request.method == "GET":
term = request.GET.get("term")
else:
return http.HttpResponse("no", status=403)
if term:
search.optimize_index(term)
return http.HttpResponse("ok")
def optimize_index(request):
if not users.is_current_user_admin():
return http.HttpResponse("no", status=403)
if request.method == "POST":
term = request.POST.get("term")
elif request.method == "GET":
term = request.GET.get("term")
else:
return http.HttpResponse("no", status=403)
if term:
search.optimize_index(term)
return http.HttpResponse("ok")
def test_index_optimization(self):
# Create a bunch of test keys.
test_keys = [
db.Key.from_path("kind_dummy", "key%02d" % i) for i in range(12)
]
# Now create four test SearchMatches, all for the same term
# and over two different fields. Split the four sets of keys
# between them.
for i in range(4):
sm = models.SearchMatches(generation=search._GENERATION,
entity_kind="kind_dummy",
field="field%d" % (i % 2),
term="foo")
sm.matches.extend(test_keys[i::4])
sm.save()
# Optimize on our term
num_deleted = search.optimize_index("foo")
# We go from 4 objects to 2, so we should return 4-2=2.
self.assertEqual(2, num_deleted)
# There should now just one SearchMatches per field.
sm = None
query = models.SearchMatches.all().filter("term =", "foo")
all_sms = sorted(query.fetch(999), key=lambda sm: sm.field)
self.assertEqual(2, len(all_sms))
for sm in all_sms:
self.assertEqual(search._GENERATION, sm.generation)
self.assertEqual("kind_dummy", sm.entity_kind)
self.assertEqual("foo", sm.term)
self.assertEqual("field0", all_sms[0].field)
self.assertEqual(set(test_keys[0::2]), set(all_sms[0].matches))
self.assertEqual("field1", all_sms[1].field)
self.assertEqual(set(test_keys[1::2]), set(all_sms[1].matches))
# Create a SearchMatches for a stop word. Optimization should
# cause that object to be deleted.
sm = models.SearchMatches(generation=search._GENERATION,
entity_kind="kind_dummy",
field="field",
term="the")
sm.matches.extend(test_keys)
sm.save()
num_deleted = search.optimize_index("the")
self.assertEqual(1, num_deleted)
query = models.SearchMatches.all().filter("term =", "the")
self.assertEqual(0, query.count())
def test_index_optimization(self):
# Create a bunch of test keys.
test_keys = [db.Key.from_path("kind_dummy", "key%02d" % i)
for i in range(12)]
# Now create four test SearchMatches, all for the same term
# and over two different fields. Split the four sets of keys
# between them.
for i in range(4):
sm = models.SearchMatches(generation=search._GENERATION,
entity_kind="kind_dummy",
field="field%d" % (i%2),
term="foo")
sm.matches.extend(test_keys[i::4])
sm.save()
# Optimize on our term
num_deleted = search.optimize_index("foo")
# We go from 4 objects to 2, so we should return 4-2=2.
self.assertEqual(2, num_deleted)
# There should now just one SearchMatches per field.
sm = None
query = models.SearchMatches.all().filter("term =", "foo")
all_sms = sorted(query.fetch(999), key=lambda sm: sm.field)
self.assertEqual(2, len(all_sms))
for sm in all_sms:
self.assertEqual(search._GENERATION, sm.generation)
self.assertEqual("kind_dummy", sm.entity_kind)
self.assertEqual("foo", sm.term)
self.assertEqual("field0", all_sms[0].field)
self.assertEqual(set(test_keys[0::2]), set(all_sms[0].matches))
self.assertEqual("field1", all_sms[1].field)
self.assertEqual(set(test_keys[1::2]), set(all_sms[1].matches))
# Create a SearchMatches for a stop word. Optimization should
# cause that object to be deleted.
sm = models.SearchMatches(generation=search._GENERATION,
entity_kind="kind_dummy",
field="field",
term="the")
sm.matches.extend(test_keys)
sm.save()
num_deleted = search.optimize_index("the")
self.assertEqual(1, num_deleted)
query = models.SearchMatches.all().filter("term =", "the")
self.assertEqual(0, query.count())
chirpradio.connect()
terms_to_opt = set()
for line in codecs.open("index.data", "r", "utf-8"):
F = [f.strip() for f in line.split(",")]
if F[2] == "1":
continue
terms_to_opt.add(F[0])
deleted = 0
skipping = True
for term in terms_to_opt:
if skipping and term:
skipping = False
if skipping:
continue
attempt = 1
while True:
try:
n = search.optimize_index(term)
break
except (datastore_errors.Timeout, urllib2.URLError):
attempt += 1
print "Timeout on attempt %d for %s!" % (attempt,
term.encode("utf-8"))
time.sleep(2)
deleted += n
print term.encode("utf-8"), n, deleted
