class CorePy(object):
def __init__(self, path, predictorType):
super(CorePy, self).__init__()
self.image = ImageFactory()
self.path = path
if predictorType == "kppv":
self.predictor = Kppv()
# elif predictorType == "mlp":
# self.predictor = Mlp()
else:
self.predictor = None
self.max_distance = 0
def setImage(self, path_to_image):
self.image.initialize(path_to_image)
def predict_current(self):
predicted_classes, result = np.zeros((len(self.image.feature_list), 2)), 0
for x in range(0,len(self.image.feature_list)):
predicted_classes[x], distance = self.predictor.predict(self.image.feature_list[x])
result += predicted_classes[x]
if distance >= 0:
self.max_distance = max(self.max_distance, distance)
self.image.class_list = predicted_classes
pass
def train_predictor(self):
self.predictor.train(self.image.feature_list, self.image.class_list)
class CorePy(object):
def __init__(self, path, predictorType):
super(CorePy, self).__init__()
self.image = ImageFactory()
self.path = path
if predictorType == "kppv":
self.predictor = Kppv()
# elif predictorType == "mlp":
# self.predictor = Mlp()
else:
self.predictor = None
self.max_distance = 0
def setImage(self, path_to_image):
self.image.initialize(path_to_image)
def predict_current(self):
predicted_classes, result = np.zeros(
(len(self.image.feature_list), 2)), 0
for x in range(0, len(self.image.feature_list)):
predicted_classes[x], distance = self.predictor.predict(
self.image.feature_list[x])
result += predicted_classes[x]
if distance >= 0:
self.max_distance = max(self.max_distance, distance)
self.image.class_list = predicted_classes
pass
def train_predictor(self):
self.predictor.train(self.image.feature_list, self.image.class_list)
def __init__(self, path, predictorType):
super(CorePy, self).__init__()
self.image = ImageFactory()
self.path = path
if predictorType == "kppv":
self.predictor = Kppv()
# elif predictorType == "mlp":
# self.predictor = Mlp()
else:
self.predictor = None
self.max_distance = 0
def __init__(self, url):
self.log = logging.getLogger('%s.%s' %
(__name__, self.__class__.__name__))
self._managedObjects = {}
self.session = None
# create a connection and connect to qpidd
# TODO: (redmine 277) - Make this use actual amqp:// urls... currently, only host works
self.connection = cqpid.Connection(url, "{reconnect:True}")
self.connection.open()
# Create, configure, and open a QMFv2 agent session using the connection.
self.session = AgentSession(self.connection)
self.session.setVendor("redhat.com")
self.session.setProduct("imagefactory")
self.session.open()
# Initialize the parent class with the session.
AgentHandler.__init__(self, self.session)
# Register our schemata with the agent session.
self.session.registerSchema(ImageFactory.qmf_schema)
self.session.registerSchema(BuildAdaptor.qmf_schema)
self.session.registerSchema(BuildAdaptor.qmf_event_schema_status)
self.session.registerSchema(BuildAdaptor.qmf_event_schema_percentage)
self.session.registerSchema(BuildAdaptor.qmf_event_schema_build_failed)
# Now add the image factory object
self.image_factory = ImageFactory()
self.image_factory.agent = self
self.image_factory_addr = self.session.addData(
self.image_factory.qmf_object, "image_factory")
self.log.info("image_factory has qmf/qpid address: %s",
self.image_factory_addr)
def resize_image_in_folder(path):
if os.path.isfile(path):
if is_deal(path):
return;
imageFactory = ImageFactory("");
try:
imageFactory.getThumbByCut(232, 93, path, "_232x93");
imageFactory.getThumbByCut(80, 80, path, "_80x80");
except:
pass;
elif os.path.isdir(path):
files = os.listdir(path);
for imageFile in files:
resize_image_in_folder(path + "/" + imageFile);
else:
print("the given path is not a folder or a file :%s"%path);
def getContentBoxesWorker(socketID, filename):
# instantiate a publisher (send to nodejs server)
context = zmq.Context()
socket = context.socket(zmq.PUB)
socket.bind("tcp://127.0.0.1:5556")
# msg["socketID"] is clientID, msg["filename"] is path
Im = ImageFactory()
path = "../public/" + filename
Im.initialize(path)
# initialize class_list
Im.class_list = np.zeros((len(Im.feature_list), 2))
for x in range(0, len(Im.feature_list)):
Im.class_list[x][0] = 1.0
# create JSON string before ...
json_string = '{"socketID":"' + socketID + '", "contentBoxes" :' + tojson([Im.content_list, Im.class_list]) + "}"
socket.send_string(json_string)
pass
def run():
prepare_directories()
download_file()
image_factory = ImageFactory(resize=256, crop=224)
print("Formatting train set")
format_set("train", TRAIN_FILE, image_factory)
print("Formatting val set")
format_set("val", VAL_FILE, image_factory)
def __init__(self, path, predictorType):
super(CorePy, self).__init__()
self.image = ImageFactory()
self.path = path
if predictorType == "kppv":
self.predictor = Kppv()
# elif predictorType == "mlp":
# self.predictor = Mlp()
else:
self.predictor = None
self.max_distance = 0
def getContentBoxesWorker(socketID, filename):
# instantiate a publisher (send to nodejs server)
context = zmq.Context()
socket = context.socket(zmq.PUB)
socket.bind("tcp://127.0.0.1:5556")
# msg["socketID"] is clientID, msg["filename"] is path
Im = ImageFactory()
path = "../public/" + filename
Im.initialize(path)
# initialize class_list
Im.class_list = np.zeros((len(Im.feature_list), 2))
for x in range(0, len(Im.feature_list)):
Im.class_list[x][0] = 1.0
# create JSON string before ...
json_string = '{"socketID":"' + socketID + '", "contentBoxes" :' + tojson(
[Im.content_list, Im.class_list]) + '}'
socket.send_string(json_string)
pass
