def log(self, *args):
from lib import logger, LOG_NOTICE
for arg in args:
if type(arg) in (str, unicode):
text = arg
else:
text = repr(arg)
logger(LOG_NOTICE, text)
def log(self, *args):
from lib import logger, LOG_NOTICE
for arg in args:
if type(arg) in (str, unicode):
text = arg
else:
text = repr(arg)
logger(LOG_NOTICE, text)
def test_loglevel_error():
trace = "tests#test_logger#test_loglevel_info"
info("Enter", trace)
init_logger("ERROR")
assert logger().getEffectiveLevel() == logging.ERROR
info("Exit", trace)
def test_loglevel_info():
trace = "tests#test_logger#test_loglevel_info"
info("Enter", trace)
init_logger("INFO")
assert logger().getEffectiveLevel() == logging.INFO
info("Exit", trace)
def POST(self, pluginName=None, *args, **kwargs):
"""Classify an image using given plugin name and return predictions.
@param pluginName: Optional name of plugin to lookup. Raise an error
if not supplied or not a valid name.
@return predictions: list of strings for predicted labels, ordered as
most likely first.
"""
if pluginName is None:
raise cherrypy.HTTPError(405,
"POST method requires a plugin name.")
try:
plugin = PLUGINS[pluginName]
except KeyError:
raise cherrypy.HTTPError(
400,
"Expected plugin name as one of {names}, but got: '{actual}'.".
format(names=CONFIGURED_PLUGINS, actual=pluginName))
if plugin is None:
raise cherrypy.HTTPError(
500,
"That plugin has not been setup on the server. Ensure it has"
" a valid graph file and that this is indicated in the"
" model conf file.")
startTime = time.time()
predictions = plugin.process(**kwargs)
maxResults = conf.getint('predictions', 'maxResults')
predictions = predictions[:maxResults]
duration = time.time() - startTime
msg = "Completed request. Name: {name}. Duration: {duration:4.3f}s."\
.format(
name=pluginName,
duration=duration
)
logger(msg, context="SERVICES.CLASSIFY.PREDICTION")
cherrypy.response.status = 201
return predictions
def process(self, imagePath=None, imageFile=None, x=None, y=None):
"""Do category prediction on the configured plugin instance using
input image and co-ordinate point.
Expects an image and mark co-ordinates, then does a prediction
for categories and returns as a list, ordered from highest to lowest
probability.
@param imagePath: Default None. path to local image on server, as
a string.
@param imageFile: Default None. Multi-party cherrpy request body.
This is expected to have a file attribute with image data.
@param x: Default None. The X co-ordinate of the mark on the image, as
a percentage value from 0 to 100. As an integer.
@param y: Default None. The Y co-ordinate of the mark on the image, as
a percentage value from 0 to 100. As an integer.
@return predictions: list of categories prediction values, as strings.
"""
start = time.time()
if imagePath:
image = imagePath
filename = os.path.basename(image)
elif imageFile:
image = imageFile.file
filename = imageFile.filename
else:
raise ValueError("Expected value for either`imagePath` or"
" `imageFile` parameters.")
preProcessedImg = self._preProcessImg(image, x, y)
predictions = self._doPrediction(preProcessedImg)
msg = "Completed prediction. Duration: {duration:4.3f}s."\
" Filename: {filename}. Results: {results}.".format(
duration=time.time() - start,
filename=filename,
results=json.dumps(predictions)
)
logger(msg=msg, context=self.getContext())
return predictions
def test_canAppendLoggers():
time = MagicMock()
time.dateTimeIso.return_value = '1'
logger = lib.logger.config(time=time,
include=['.*'],
exclude=[],
enabled=True)
assert logger(append='1')(append='2')(
append='3')('abc') == '[1][1:2:3] abc'
def test_canSpecifyCustomName():
time = MagicMock()
time.dateTimeIso.return_value = '1'
logger = lib.logger.config(time=time,
include=['.*'],
exclude=[],
enabled=True)
assert logger(append='1')(append='2')(append='3')(
'abc', name='custom') == '[1][1:2:3:custom] abc'
def _loadGraph(self, modelPath):
"""Prepare the plugin's graph definition so we can use its tensors.
The time taken for this method to compelte is sent to the error log.
This is a private method which be called internally on model
initialisation. It unpersists a TensorFlow graph from the proto-buf
file, so that we can use it in a TensorFlow session to do predictions.
According to the `tf.import_graph_def` docs, it expects a `GraphDef`
proto containing operations to be imported in the default graph, so
they be extracted as a `tf.Tensor` and `tf.Operation` objects.
@param modelPath: path to the model graph file, as a string.
@return graph: a tf.Graph instance which contains the graph definition
for the required model.
"""
startTime = time.time()
# Create a new graph then set it as the default.
graph = tf.Graph()
with graph.as_default():
with tf.gfile.FastGFile(modelPath, 'rb') as fIn:
# Create an empty graph-def then load the proto-buf file into
# the graph-def.
graphDef = tf.GraphDef()
graphDef.ParseFromString(fIn.read())
# Import the graph-def to the default TensorFlow graph.
# Name is a prefix for the tensor and should be left empty.
tf.import_graph_def(graphDef, name='')
duration = time.time() - startTime
msg = "Loaded model. Duration: {0:4.3f}s.".format(duration)
logger(msg, context=self.getContext())
return graph
def log(self, msg):
from lib import logger
logger(LOG_NOTICE, msg)
def _add_logger(self):
self.logger = logger('')
def __init__(self):
self.mml = MultiMediaLib()
self.mh = MediaHandler()
self.logger = logger()
def __init__(self):
self.logger = logger()
self.video_codec_list = ['avi', 'cam', 'mkv', 'mov', 'mpeg', 'mpg', 'mpe', 'svi', 'wmv',]
self.music_codec_list = ['jpeg', 'png', 'ppm', 'jpg', 'tga']
self.image_codec_list = []
self.doc_codec_list = []
import time, machine, update, wifisettings, nodesettings, lib.requests, lib.logger, lib.requests, lib.timew, time, os, machine
from umqtt.robust import MQTTClient
from machine import Pin, WDT
led = Pin(2, Pin.OUT)
led.off()
t = lib.timew.Time(time=time)
# Configure Logger
logger = lib.logger.config(enabled=nodesettings.settings['debug'],
include=nodesettings.settings['logInclude'],
exclude=nodesettings.settings['logExclude'],
time=t)
log = logger(append='boot')
log("The current time is %s" % t.human())
mqtt_pth = nodesettings.settings['controllerName']
client = MQTTClient(mqtt_pth,
wifisettings.settings['mqtt_ip'],
port=wifisettings.settings['mqtt_port'])
mq_c = False
for i in range(1):
try:
client.connect()
mq_c = True
except:
log('MQTT: cannot connect')
led.on()
time.sleep(1)
if not mq_c:
