def ExecuteRefinement_Task(pickled_model_coarse, pickled_parameters, min_size,
max_size):
sample = GenerateSample()
'''overwrite the old model serializer with the unpickled one'''
model_serializer_coarse = pickle.loads(pickled_model_coarse)
model_coarse = KratosMultiphysics.Model()
model_serializer_coarse.Load("ModelSerialization", model_coarse)
del (model_serializer_coarse)
'''overwrite the old parameters serializer with the unpickled one'''
serialized_parameters = pickle.loads(pickled_parameters)
parameters_refinement = KratosMultiphysics.Parameters()
serialized_parameters.Load("ParametersSerialization",
parameters_refinement)
del (serialized_parameters)
simulation_coarse = MonteCarloAnalysis(model_coarse, parameters_refinement,
sample)
simulation_coarse.Run()
QoI = EvaluateQuantityOfInterest(simulation_coarse)
'''refine'''
model_refined = refinement.compute_refinement_from_analysisstage_object(
simulation_coarse, min_size, max_size)
'''initialize'''
simulation = MonteCarloAnalysis(model_refined, parameters_refinement,
sample)
simulation.Initialize()
'''serialize model and pickle it'''
serialized_model = KratosMultiphysics.StreamSerializer()
serialized_model.Save("ModelSerialization", simulation.model)
pickled_model_refined = pickle.dumps(serialized_model, 2)
return QoI, pickled_model_refined
def restore_object(cls, obj):
"""
recreate objects bases on classes currently avaliable
"""
if obj == None:
return obj
for a in cls.allowed:
if isinstance(obj, a):
return obj
if isinstance(obj, list):
return [cls.restore_object(x) for x in obj]
if isinstance(obj, dict):
if "_type" in obj:
t = obj["_type"]
# restored registered classes
if t in cls.classes and "_content" in obj:
model = cls.classes[t]
return model.serial_loads(obj)
# restored pickled classes
if "_pickle" in obj:
# object restoration
ins = pickle.loads(str(obj["_pickle"]))
return ins
else:
# recursive dictionary restore
return {x: cls.restore_object(y) for x, y in obj.iteritems()}
def normalize_rpc(self, rpc):
try:
_rpc = pickle.loads(rpc)
assert isinstance(_rpc, dict), 'RPC request format error.'
return rpc
except:
return rpc
def ExecuteExactMonteCarloAnalysis_Task(pickled_model, pickled_parameters):
'''overwrite the old model serializer with the unpickled one'''
model_serializer = pickle.loads(pickled_model)
current_model = KratosMultiphysics.Model()
model_serializer.Load("ModelSerialization", current_model)
del (model_serializer)
'''overwrite the old parameters serializer with the unpickled one'''
serialized_parameters = pickle.loads(pickled_parameters)
current_parameters = KratosMultiphysics.Parameters()
serialized_parameters.Load("ParametersSerialization", current_parameters)
del (serialized_parameters)
sample = 1.0
simulation = MonteCarloAnalysis(current_model, current_parameters, sample)
simulation.Run()
ExactExpectedValueQoI = 0.25 * EvaluateQuantityOfInterest(simulation)
return ExactExpectedValueQoI
def restore_object(cls, obj):
"""
recreate objects bases on classes currently avaliable
"""
if obj == None:
return obj
for a in cls.allowed:
if isinstance(obj, a):
return obj
if isinstance(obj, list):
return [cls.restore_object(x) for x in obj]
if isinstance(obj, dict):
if "_type" in obj:
t = obj["_type"]
# restored registered classes
if t in cls.classes and "_content" in obj:
model = cls.classes[t]
return model.serial_loads(obj)
# restored pickled classes
if "_pickle" in obj:
# object restoration
ins = pickle.loads(str(obj["_pickle"]).encode("utf-8"))
return ins
else:
# recursive dictionary restore
return {x: cls.restore_object(y) for x, y in obj.items()}
def ExecuteInstance_Task(pickled_model,pickled_parameters,heat_flux_list,instance):
# overwrite the old model serializer with the unpickled one
model_serializer = pickle.loads(pickled_model)
current_model = KratosMultiphysics.Model()
model_serializer.Load("ModelSerialization",current_model)
del(model_serializer)
# overwrite the old parameters serializer with the unpickled one
serialized_parameters = pickle.loads(pickled_parameters)
current_parameters = KratosMultiphysics.Parameters()
serialized_parameters.Load("ParametersSerialization",current_parameters)
del(serialized_parameters)
# get sample
sample = GetValueFromListList(heat_flux_list,instance) # take
simulation = SimulationScenario(current_model,current_parameters,sample)
simulation.Run()
QoI = simulation.EvaluateQuantityOfInterest()
return QoI
def ExecuteInstanceAux_Task(pickled_model,pickled_project_parameters,current_analysis_stage,current_level):
# overwrite the old model serializer with the unpickled one
model_serializer = pickle.loads(pickled_model)
current_model = KratosMultiphysics.Model()
model_serializer.Load("ModelSerialization",current_model)
del(model_serializer)
# overwrite the old parameters serializer with the unpickled one
serialized_project_parameters = pickle.loads(pickled_project_parameters)
current_project_parameters = KratosMultiphysics.Parameters()
serialized_project_parameters.Load("ParametersSerialization",current_project_parameters)
del(serialized_project_parameters)
# initialize the MonteCarloResults class
mc_results_class = MonteCarloResults(current_level)
sample = generator.GenerateSample()
simulation = current_analysis_stage(current_model,current_project_parameters,sample)
simulation.Run()
QoI = simulation.EvaluateQuantityOfInterest()
mc_results_class.QoI[current_level].append(QoI) # saving results in the corresponding list, for MC only list of level 0
return mc_results_class
def ExecuteInstanceAux_Task(pickled_model, pickled_project_parameters, sample,
current_analysis_stage, current_level):
# def ExecuteInstanceAux_Task(serialized_model,serialized_project_parameters,current_analysis_stage,current_level):
time_0 = time.time()
# overwrite the old model serializer with the unpickled one
serialized_model = pickle.loads(pickled_model)
current_model = KratosMultiphysics.Model()
serialized_model.Load("ModelSerialization", current_model)
del (serialized_model)
# overwrite the old parameters serializer with the unpickled one
serialized_project_parameters = pickle.loads(pickled_project_parameters)
current_project_parameters = KratosMultiphysics.Parameters()
serialized_project_parameters.Load("ParametersSerialization",
current_project_parameters)
del (serialized_project_parameters)
time_1 = time.time()
# initialize the MonteCarloResults class
mc_results_class = MonteCarloResults(current_level)
simulation = current_analysis_stage(current_model,
current_project_parameters, sample)
simulation.Run()
QoI = simulation.EvaluateQuantityOfInterest()
time_2 = time.time()
mc_results_class.QoI[current_level].append(
QoI
) # saving results in the corresponding list, for MC only list of level 0
# post process execution times
# print("\n","#"*50," EXECUTE INSTANCE TASK TIMES ","#"*50,"\n")
deserialization_time = time_1 - time_0
Kratos_run_time = time_2 - time_1
total_task_time = time_2 - time_0
# print("[LEVEL] current level:",current_level)
# print("[TIMER] total task time:", total_task_time)
# print("[TIMER] Kratos run time:",Kratos_run_time)
# print("[TIMER] Serializer time:",deserialization_time)
# print("RATIOs: time of interest / total task time")
# print("[RATIO] Relative serializer time:",deserialization_time/total_task_time)
# print("[RATIO] Relative Kratos run time:",Kratos_run_time/total_task_time)
# print("\n","#"*50," END EXECUTE INSTANCE TASK TIMES ","#"*50,"\n")
# end post process execution times
return mc_results_class
def receive(self):
while True:
try:
data, addr = self.socket.recvfrom(1024)
r = pickle.loads(data)
for i in r:
if not i in self.addrUser:
print(addr, "client!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!")
self.addrUser[i] = True
return r
except socket.error:
break
return {}
def __getitem__(self, url):
"""Load data from disk for this URL
"""
path = self.url_to_path(url)
if os.path.exists(path):
with open(path, 'rb') as fp:
data = fp.read()
if self.compress:
data = zlib.decompress(data)
result, timestamp = pickle.loads(data)
print(result, 'exist')
#检验文件的过期时间（用自定义的has_expired()函数判断）
if self.has_expired(timestamp):
raise KeyError(url + ' has expired')
return result
else:
# url不存在的情况
raise KeyError(url + ' does not exist')
def __init__(self, collection_name, prefix='db_', ext='pkl'):
"""Build pickle file name and load data if exists.
:param collection_name: Collection name
:param prefix: File prefix; defaults to 'db_'
:param ext: File extension; defaults to 'pkl'
"""
# Build filename
self.filename = prefix + collection_name + '.' + ext
# Initialize empty store
self.store = {}
# Load file if exists
if os.path.exists(self.filename):
with open(self.filename, 'rb') as fp:
data = fp.read()
self.store = pickle.loads(data)
def __init__(self, collection_name, prefix='db_', ext='pkl'):
"""Build pickle file name and load data if exists.
:param collection_name: Collection name
:param prefix: File prefix; defaults to 'db_'
:param ext: File extension; defaults to 'pkl'
"""
# Build filename
self.filename = prefix + collection_name + '.' + ext
# Initialize empty store
self.store = {}
# Load file if exists
if os.path.exists(self.filename):
with open(self.filename, 'rb') as fp:
data = fp.read()
self.store = pickle.loads(data)
def test_model_serialization_with_pickling(self):
current_model = KratosMultiphysics.Model()
model_part = current_model.CreateModelPart("Main")
model_part.AddNodalSolutionStepVariable(KratosMultiphysics.TEMPERATURE)
model_part.CreateSubModelPart("Inlets")
model_part.CreateSubModelPart("Temp")
model_part.CreateNewNode(1, 0.0, 0.0, 0.0)
other = current_model.CreateModelPart("Other")
other.AddNodalSolutionStepVariable(KratosMultiphysics.PRESSURE)
other.CreateNewNode(1, 0.0, 0.0, 0.0)
serializer = KratosMultiphysics.StreamSerializer()
serializer.Save("ModelSerialization", current_model)
del (current_model)
# ######## here we pickle the serializer
try:
import cpickle as pickle # Use cPickle on Python 2.7
except ImportError:
import pickle
#pickle dataserialized_data
pickled_data = pickle.dumps(
serializer, 2
) #second argument is the protocol and is NECESSARY (according to pybind11 docs)
#overwrite the old serializer with the unpickled one
serializer = pickle.loads(pickled_data)
loaded_model = KratosMultiphysics.Model()
serializer.Load("ModelSerialization", loaded_model)
self.assertTrue(loaded_model["Main"].HasNodalSolutionStepVariable(
KratosMultiphysics.TEMPERATURE))
self.assertTrue(loaded_model["Other"].HasNodalSolutionStepVariable(
KratosMultiphysics.PRESSURE))
self.assertTrue(loaded_model.HasModelPart("Main.Inlets"))
self.assertTrue(loaded_model.HasModelPart("Main.Temp"))
self.assertTrue(1 in loaded_model["Main"].Nodes)
self.assertTrue(1 in loaded_model["Other"].Nodes)
def unpickle(self, data):
try:
return pickle.loads(str(data))
except:
return None
def _ToObj(objVacio,datos):#lo necesito para poderlo usar de forma generica con los objetos complejos return pickle.loads(datos)
def unpickle(self, data):
try:
return pickle.loads(str(data))
except:
return None
def compute(database, model, sid, uid, workdir, hostname, username, password):
"""
Computes the Time Series model. It fetches the necessary files from a
remote server that were computed by the slycat-agent-compute-timeseries.py
script.
:param database:
:param model:
:param sid: session ID
:param uid: user ID
:param workdir:
:param hostname:
:param username:
:param password:
"""
workdir += "/slycat/pickle" # route to the slycat directory
try:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
model["model_compute_time"] = datetime.datetime.utcnow().isoformat()
slycat.web.server.update_model(database, model)
sid, inputs = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/arrayset_inputs.pickle" % (workdir, uid))
inputs = pickle.loads(inputs)
slycat.web.server.put_model_arrayset(database, model, inputs["aid"])
attributes = inputs["attributes"]
slycat.web.server.put_model_array(database, model, inputs["aid"], 0, attributes, inputs["dimensions"])
sid, data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/inputs_attributes_data.pickle" % (workdir, uid))
attributes_data = pickle.loads(data)
# TODO this can become multi processored
for attribute in range(len(attributes)):
slycat.web.server.put_model_arrayset_data(database, model, inputs["aid"], "0/%s/..." % attribute,
[attributes_data[attribute]])
clusters = json.loads(
slycat.web.server.get_remote_file(sid, "%s/slycat_timeseries_%s/file_clusters.json" % (workdir, uid)))
clusters_file = json.JSONDecoder().decode(clusters["file"])
timeseries_count = json.JSONDecoder().decode(clusters["timeseries_count"])
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_clusters.out" % (workdir, uid),
clusters["aid"], clusters["parser"])
# TODO this can become multi processored
for file_name in clusters_file:
sid, file_cluster_data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/file_cluster_%s.json" % (
workdir, uid, file_name))
file_cluster_attr = json.loads(file_cluster_data)
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_cluster_%s.out" % (workdir, uid, file_name),
file_cluster_attr["aid"], file_cluster_attr["parser"])
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset(database, model, "preview-%s" % file_name)
sid, waveform_dimensions_data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_dimensions.pickle" % (
workdir, uid, file_name))
waveform_dimensions_array = pickle.loads(waveform_dimensions_data)
sid, waveform_attributes_data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_attributes.pickle" % (
workdir, uid, file_name))
waveform_attributes_array = pickle.loads(waveform_attributes_data)
sid, waveform_times_data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_times.pickle" % (
workdir, uid, file_name))
waveform_times_array = pickle.loads(waveform_times_data)
sid, waveform_values_data = get_remote_file(sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_values.pickle" % (
workdir, uid, file_name))
waveform_values_array = pickle.loads(waveform_values_data)
cherrypy.log.error("timeseries_count=%s" % timeseries_count)
# TODO this can become multi processored
for index in range(int(timeseries_count)):
try:
slycat.web.server.put_model_array(database, model, "preview-%s" % file_name, index,
waveform_attributes_array[index],
waveform_dimensions_array[index])
slycat.web.server.put_model_arrayset_data(database, model, "preview-%s" % file_name,
"%s/0/...;%s/1/..." % (index, index),
[waveform_times_array[index],
waveform_values_array[index]])
except:
cherrypy.log.error("failed on index: %s" % index)
pass
except:
fail_model(model["_id"], "Timeseries model compute exception: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception type: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" % sys.exc_info()[1])
cherrypy.log.error("Timeseries model compute exception traceback: %s" % sys.exc_info()[2])
#encoding:utf-8
try:
import cpickle as pickle
except:
import pickle
import pprint
data=[{'a':'A','b':2.0,'c':2.5}]
print 'BEFORE Date:',
pprint.pprint(data)
#pickle只包含ASCII字符
data_string=pickle.dumps(data)
data2=pickle.loads(data_string)
print 'AFTER pickle: ',
pprint.pprint(data2)
print 'SAME?:',(data is data2)
print 'EQL?:',(data==data2)
def compute(model_id, password):
"""
Computes the Time Series model. It fetches the necessary files from a
remote server that were computed by the slycat-agent-compute-timeseries.py
script.
:param model_id: uid for the model in the database
:param sid: session ID
:param uid: user ID
:param workdir:
:param hostname:
:param username:
:param password:
"""
cherrypy.log.error("in thread")
# workdir += "/slycat/pickle" # route to the slycat directory
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", True)
computing = True
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
model["model_compute_time"] = datetime.datetime.utcnow().isoformat()
slycat.web.server.update_model(database, model, state="waiting")
tries = 10
while computing:
tries = tries - 1
if tries <= 0:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", False)
computing = False
fail_model(model_id, "Exceeded max number of tries to pull data over to the server.")
cherrypy.log.error("[TIMESERIES] Exceeded max number of tries to pull data.")
raise Exception("Exceeded max number of tries to pull data over to the server.")
try:
uid = slycat.web.server.get_model_parameter(database, model, "pickle_uid")
workdir_raw = slycat.web.server.get_model_parameter(database, model, "working_directory")
workdir = workdir_raw + "pickle"
hostname = slycat.web.server.get_model_parameter(database, model, "hostname")
username = slycat.web.server.get_model_parameter(database, model, "username")
# get an active session
sid = get_sid(hostname, model)
cherrypy.log.error("sid:%s uid:%s work_dir:%s host:%s user:%s" % (
sid, uid, workdir, hostname, username))
inputs = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/arrayset_inputs.pickle" % (workdir, uid))
cherrypy.log.error("got inputs")
inputs = pickle.loads(inputs)
slycat.web.server.put_model_arrayset(database, model, inputs["aid"])
attributes = inputs["attributes"]
slycat.web.server.put_model_array(database, model, inputs["aid"], 0, attributes, inputs["dimensions"])
data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/inputs_attributes_data.pickle" % (workdir, uid))
attributes_data = pickle.loads(data)
# TODO this can become multi processored
for attribute in range(len(attributes)):
slycat.web.server.put_model_arrayset_data(database, model, inputs["aid"], "0/%s/..." % attribute,
[attributes_data[attribute]])
clusters = json.loads(
slycat.web.server.get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/file_clusters.json" % (workdir, uid)))
clusters_file = json.JSONDecoder().decode(clusters["file"])
timeseries_count = json.JSONDecoder().decode(clusters["timeseries_count"])
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_clusters.out" % (workdir, uid),
clusters["aid"], clusters["parser"], client=model["creator"])
# TODO this can become multi processored
cherrypy.log.error("Pulling timeseries computed data")
for file_name in clusters_file:
file_cluster_data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/file_cluster_%s.json" % (
workdir, uid, file_name))
file_cluster_attr = json.loads(file_cluster_data)
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_cluster_%s.out" % (
workdir, uid, file_name),
file_cluster_attr["aid"], file_cluster_attr["parser"], client=model["creator"])
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset(database, model, "preview-%s" % file_name)
waveform_dimensions_data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/waveform_%s_dimensions"
".pickle" % (
workdir, uid, file_name))
waveform_dimensions_array = pickle.loads(waveform_dimensions_data)
waveform_attributes_data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/waveform_%s_attributes"
".pickle" % (
workdir, uid, file_name))
waveform_attributes_array = pickle.loads(waveform_attributes_data)
waveform_times_data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/waveform_%s_times"
".pickle" % (
workdir, uid, file_name))
waveform_times_array = pickle.loads(waveform_times_data)
waveform_values_data = get_remote_file_server(model["creator"], sid,
"%s/slycat_timeseries_%s/waveform_%s_values"
".pickle" % (
workdir, uid, file_name))
waveform_values_array = pickle.loads(waveform_values_data)
# cherrypy.log.error("timeseries_count=%s" % timeseries_count)
# TODO this can become multi processored
for index in range(int(timeseries_count)):
try:
slycat.web.server.put_model_array(database, model, "preview-%s" % file_name, index,
waveform_attributes_array[index],
waveform_dimensions_array[index])
slycat.web.server.put_model_arrayset_data(database, model, "preview-%s" % file_name,
"%s/0/...;%s/1/..." % (index, index),
[waveform_times_array[index],
waveform_values_array[index]])
except:
cherrypy.log.error("failed on index: %s" % index)
pass
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
model["computing"] = False
slycat.web.server.update_model(database, model)
computing = False
cherrypy.log.error("finnished Pulling timeseries computed data")
# TODO add finished to the model state
# TODO add remove dir command by uncommenting below
# payload = {
# "action": "run_remote_command",
# "command": ("rm -rf %s" % workdir_raw)
# }
except cherrypy._cperror.HTTPError as e:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", False)
cherrypy.log.error("Timeseries model compute exception type: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" % sys.exc_info()[1])
cherrypy.log.error("Timeseries model compute exception traceback: %s" % sys.exc_info()[2])
raise Exception(e.message)
except Exception as e:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", False)
cherrypy.log.error("Timeseries model compute exception type: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" % sys.exc_info()[1])
cherrypy.log.error("Timeseries model compute exception traceback: %s" % sys.exc_info()[2])
# fail_model(model_id, "Timeseries model compute exception: %s" % sys.exc_info()[0])
raise Exception(e.message)
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.delete_model_parameter(database, model, "computing")
def compute(database, model, sid, uid, workdir, hostname, username,
password):
"""
Computes the Time Series model. It fetches the necessary files from a
remote server that were computed by the slycat-agent-compute-timeseries.py
script.
:param database:
:param model:
:param sid: session ID
:param uid: user ID
:param workdir:
:param hostname:
:param username:
:param password:
"""
workdir += "/slycat/pickle" # route to the slycat directory
try:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
model["model_compute_time"] = datetime.datetime.utcnow().isoformat(
)
slycat.web.server.update_model(database, model)
sid, inputs = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/arrayset_inputs.pickle" %
(workdir, uid))
inputs = pickle.loads(inputs)
slycat.web.server.put_model_arrayset(database, model,
inputs["aid"])
attributes = inputs["attributes"]
slycat.web.server.put_model_array(database, model, inputs["aid"],
0, attributes,
inputs["dimensions"])
sid, data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/inputs_attributes_data.pickle" %
(workdir, uid))
attributes_data = pickle.loads(data)
# TODO this can become multi processored
for attribute in range(len(attributes)):
slycat.web.server.put_model_arrayset_data(
database, model, inputs["aid"], "0/%s/..." % attribute,
[attributes_data[attribute]])
clusters = json.loads(
slycat.web.server.get_remote_file(
sid, "%s/slycat_timeseries_%s/file_clusters.json" %
(workdir, uid)))
clusters_file = json.JSONDecoder().decode(clusters["file"])
timeseries_count = json.JSONDecoder().decode(
clusters["timeseries_count"])
slycat.web.server.post_model_file(
model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_clusters.out" % (workdir, uid),
clusters["aid"], clusters["parser"])
# TODO this can become multi processored
for file_name in clusters_file:
sid, file_cluster_data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/file_cluster_%s.json" %
(workdir, uid, file_name))
file_cluster_attr = json.loads(file_cluster_data)
slycat.web.server.post_model_file(
model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_cluster_%s.out" %
(workdir, uid, file_name), file_cluster_attr["aid"],
file_cluster_attr["parser"])
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset(database, model,
"preview-%s" % file_name)
sid, waveform_dimensions_data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_dimensions.pickle" %
(workdir, uid, file_name))
waveform_dimensions_array = pickle.loads(
waveform_dimensions_data)
sid, waveform_attributes_data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_attributes.pickle" %
(workdir, uid, file_name))
waveform_attributes_array = pickle.loads(
waveform_attributes_data)
sid, waveform_times_data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_times.pickle" %
(workdir, uid, file_name))
waveform_times_array = pickle.loads(waveform_times_data)
sid, waveform_values_data = get_remote_file(
sid, hostname, username, password,
"%s/slycat_timeseries_%s/waveform_%s_values.pickle" %
(workdir, uid, file_name))
waveform_values_array = pickle.loads(waveform_values_data)
cherrypy.log.error("timeseries_count=%s" % timeseries_count)
# TODO this can become multi processored
for index in range(int(timeseries_count)):
try:
slycat.web.server.put_model_array(
database, model, "preview-%s" % file_name, index,
waveform_attributes_array[index],
waveform_dimensions_array[index])
slycat.web.server.put_model_arrayset_data(
database, model, "preview-%s" % file_name,
"%s/0/...;%s/1/..." % (index, index), [
waveform_times_array[index],
waveform_values_array[index]
])
except:
cherrypy.log.error("failed on index: %s" % index)
pass
except:
fail_model(
model["_id"],
"Timeseries model compute exception: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception type: %s" %
sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" %
sys.exc_info()[1])
cherrypy.log.error(
"Timeseries model compute exception traceback: %s" %
sys.exc_info()[2])
def ExecuteMultilevelMonteCarloAnalisys_Task(
current_MLMC_level, pickled_coarse_model, pickled_coarse_parameters,
size_meshes, pickled_settings_metric_refinement,
pickled_settings_remesh_refinement, sample, current_level,
mlmc_results_class):
'''unpickle model and build Kratos Model object'''
serialized_model = pickle.loads(pickled_coarse_model)
current_model = KratosMultiphysics.Model()
serialized_model.Load("ModelSerialization", current_model)
del (serialized_model)
'''unpickle parameters and build Kratos Parameters object'''
serialized_parameters = pickle.loads(pickled_coarse_parameters)
current_parameters = KratosMultiphysics.Parameters()
serialized_parameters.Load("ParametersSerialization", current_parameters)
del (serialized_parameters)
'''start time'''
start_MLMC_time = time.time()
'''refine if current current_level > 0, adaptive refinement based on the solution of previous level'''
if (current_level > 0):
'''unpickle metric and remesh refinement parameters and build Kratos Parameters objects'''
settings_metric_refinement_serializer = pickle.loads(
pickled_settings_metric_refinement)
settings_remesh_refinement_serializer = pickle.loads(
pickled_settings_remesh_refinement)
current_settings_metric_refinement = KratosMultiphysics.Parameters()
current_settings_remesh_refinement = KratosMultiphysics.Parameters()
settings_metric_refinement_serializer.Load(
"MetricRefinementParametersSerialization",
current_settings_metric_refinement)
settings_remesh_refinement_serializer.Load(
"RemeshRefinementParametersSerialization",
current_settings_remesh_refinement)
del (settings_metric_refinement_serializer,
settings_remesh_refinement_serializer)
'''refine the model Kratos object'''
refined_model, refined_parameters = refinement.compute_refinement_hessian_metric(
current_model, current_parameters, size_meshes[current_level],
size_meshes[current_level - 1], current_settings_metric_refinement,
current_settings_remesh_refinement)
'''initialize the model Kratos object'''
simulation = MultilevelMonteCarloAnalysis(refined_model,
refined_parameters, sample)
simulation.Initialize()
'''update model Kratos object'''
current_model = simulation.model
current_parameters = simulation.project_parameters
del (simulation)
simulation = MultilevelMonteCarloAnalysis(current_model,
current_parameters, sample)
simulation.Run()
QoI = EvaluateQuantityOfInterest(simulation)
'''save model and parameters as StreamSerializer Kratos objects'''
serialized_finer_model = KratosMultiphysics.StreamSerializer()
serialized_finer_model.Save("ModelSerialization", simulation.model)
serialized_finer_parameters = KratosMultiphysics.StreamSerializer()
serialized_finer_parameters.Save("ParametersSerialization",
simulation.project_parameters)
'''pickle model and parameters'''
pickled_finer_model = pickle.dumps(
serialized_finer_model, 2
) # second argument is the protocol and is NECESSARY (according to pybind11 docs)
pickled_finer_parameters = pickle.dumps(
serialized_finer_parameters, 2
) # second argument is the protocol and is NECESSARY (according to pybind11 docs)
del (simulation)
end_MLMC_time = time.time()
'''register results of the current level in the MultilevelMonteCarloResults class'''
mlmc_results_class.time_ML[current_level].append(
end_MLMC_time - start_MLMC_time
) # saving each result in the corresponding list in order to ensure the correctness of the results order and the levels
mlmc_results_class.QoI[current_level].append(
QoI
) # saving each result in the corresponding list in order to ensure the correctness of the results order and the levels
return mlmc_results_class, pickled_finer_model, pickled_finer_parameters
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# -----------------------------------------------------
# FileName : pickle_string.py
# Author : linuxme@
# Project : monitor_maintain
# Date : 2013-09-01 10:32
# Description :
# -----------------------------------------------------
try:
import cpickle as pickle
except:
import pickle
import pprint
data1 = [{'a': 'A', 'b': 2, 'c': 3.0}]
print 'BEFORE: ',
pprint.pprint(data1)
data1_string = pickle.dumps(data1)
data2 = pickle.loads(data1_string)
print 'AFTER : ',
pprint.pprint(data2)
print 'SAME? :', (data1 is data2)
print 'EQUAL?:', (data1 == data2)
def compute(model_id, stop_event, calling_client):
"""
Computes the Time Series model. It fetches the necessary files from a
remote server that were computed by the slycat-agent-compute-timeseries.py
script.
:param model_id: uid for the model in the database
:param sid: session ID
:param uid: user ID
:param workdir:
:param hostname:
:param username:
"""
try:
total_file_delta_time = []
#cherrypy.log.error("in thread")
# workdir += "/slycat/pickle" # route to the slycat directory
start_time = time.time()
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
model["model_compute_time"] = datetime.datetime.utcnow().isoformat()
with slycat.web.server.get_model_lock(model["_id"]):
database.save(model)
slycat.web.server.update_model(database, model, state="waiting", message="starting data pull Timeseries")
model = database.get("model", model_id)
uid = slycat.web.server.get_model_parameter(database, model, "pickle_uid")
workdir_raw = slycat.web.server.get_model_parameter(database, model, "working_directory")
workdir = workdir_raw + "pickle"
hostname = slycat.web.server.get_model_parameter(database, model, "hostname")
username = slycat.web.server.get_model_parameter(database, model, "username")
# get an active session
sid = get_sid(hostname, model)
# load inputs
slycat.web.server.update_model(database, model, progress=50, message="loading inputs")
use_tar = True
# keep this blank unless we need it
pickle_path = ''
input_tar=None
try:
myfiles_tar_gz = get_remote_file_server(hostname, model,
"%s/slycat_timeseries_%s/slycat-timeseries.tar.gz" % (workdir, uid),
total_file_delta_time,
calling_client)
myfiles_tar_gz = io.BytesIO(myfiles_tar_gz)
input_tar = tarfile.open(fileobj=myfiles_tar_gz, mode="r:gz")
except:
# looks like the file is too large lets just grab one file at a time
use_tar = False
pickle_path = "%s/slycat_timeseries_%s/" % (workdir, uid)
inputs = helpGetFile("%sarrayset_inputs.pickle" % (pickle_path),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
inputs = pickle.loads(inputs)
slycat.web.server.put_model_arrayset(database, model, inputs["aid"])
# load attributes
slycat.web.server.update_model(database, model, progress=55, message="loading attributes")
attributes = inputs["attributes"]
slycat.web.server.put_model_array(database, model, inputs["aid"], 0, attributes, inputs["dimensions"])
# load attribute data
data = helpGetFile("%sinputs_attributes_data.pickle" % (pickle_path),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
attributes_data = pickle.loads(data)
# push attribute arraysets
# TODO this can become multi processored
for attribute in range(len(attributes)):
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset_data(database, model, inputs["aid"], "0/%s/..." % attribute,
[attributes_data[attribute]])
# load clusters data
slycat.web.server.update_model(database, model, progress=60, message="loading cluster data")
clusters = helpGetFile("%sfile_clusters.json" % (pickle_path),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
clusters = json.loads(clusters)
clusters_file = json.JSONDecoder().decode(clusters["file"])
timeseries_count = json.JSONDecoder().decode(clusters["timeseries_count"])
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_clusters.out" % (workdir, uid),
clusters["aid"], clusters["parser"], client=calling_client)
# TODO this can become multi processored
cherrypy.log.error("Pulling timeseries computed data")
slycat.web.server.update_model(database, model, progress=65, message="Pulling timeseries computed data for %s cluster files" % len(clusters_file))
progress = 65
progress_part = 30/len(clusters_file)
for file_name in clusters_file:
progress = progress + progress_part
slycat.web.server.update_model(database, model, progress=progress, message="loading %s cluster file" % file_name)
file_cluster_data = helpGetFile("%sfile_cluster_%s.json" % (pickle_path, file_name),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
file_cluster_attr = json.loads(file_cluster_data)
slycat.web.server.post_model_file(model["_id"], True, sid,
"%s/slycat_timeseries_%s/file_cluster_%s.out" % (
workdir, uid, file_name),
file_cluster_attr["aid"], file_cluster_attr["parser"], client=calling_client)
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset(database, model, "preview-%s" % file_name)
waveform_dimensions_data = helpGetFile("%swaveform_%s_dimensions.pickle" % (pickle_path, file_name),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
waveform_dimensions_array = pickle.loads(waveform_dimensions_data)
waveform_attributes_data = helpGetFile("%swaveform_%s_attributes.pickle" % (pickle_path, file_name),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
waveform_attributes_array = pickle.loads(waveform_attributes_data)
waveform_times_data = helpGetFile("%swaveform_%s_times.pickle" % (pickle_path, file_name),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
waveform_times_array = pickle.loads(waveform_times_data)
waveform_values_data = helpGetFile("%swaveform_%s_values.pickle" % (pickle_path, file_name),
use_tar, hostname, model, total_file_delta_time,calling_client, input_tar)
waveform_values_array = pickle.loads(waveform_values_data)
for index in range(int(timeseries_count)):
try:
model = database.get("model", model["_id"])
slycat.web.server.put_model_array(database, model, "preview-%s" % file_name, index,
waveform_attributes_array[index],
waveform_dimensions_array[index])
model = database.get("model", model["_id"])
slycat.web.server.put_model_arrayset_data(database, model, "preview-%s" % file_name,
"%s/0/...;%s/1/..." % (index, index),
[waveform_times_array[index],
waveform_values_array[index]])
except:
cherrypy.log.error("failed on index: %s" % index)
pass
if input_tar:
input_tar.close()
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.update_model(database, model, message="finished loading all data")
slycat.web.server.put_model_parameter(database, model, "computing", False)
cherrypy.log.error("finished Pulling timeseries computed data")
finish_time = time.time()
file_stats = {
"min": min(total_file_delta_time),
"max": max(total_file_delta_time),
"mean": statistics.mean(total_file_delta_time),
"median": statistics.median(total_file_delta_time),
"number_of_files_pulled":len(total_file_delta_time),
"total_time_Pulling_data": sum(total_file_delta_time),
"total_time": (finish_time - start_time)
}
cherrypy.log.error("File Stats %s" % str(file_stats))
total_file_delta_time = []
finish(model["_id"])
stop_event.set()
# TODO add finished to the model state
# TODO add remove dir command by uncommenting below
# payload = {
# "action": "run_remote_command",
# "command": ("rm -rf %s" % workdir_raw)
# }
except cherrypy._cperror.HTTPError as e:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", False)
cherrypy.log.error(traceback.format_exc())
cherrypy.log.error("Timeseries cperror model compute exception type: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" % sys.exc_info()[1])
cherrypy.log.error("Timeseries model compute exception traceback: %s" % sys.exc_info()[2])
stop_event.set()
except:
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model_id)
slycat.web.server.put_model_parameter(database, model, "computing", False)
cherrypy.log.error(traceback.format_exc())
cherrypy.log.error("Timeseries model compute exception type: %s" % sys.exc_info()[0])
cherrypy.log.error("Timeseries model compute exception value: %s" % sys.exc_info()[1])
cherrypy.log.error("Timeseries model compute exception traceback: %s" % sys.exc_info()[2])
stop_event.set()
def compute(database, model, sid, uid, workdir, hostname, username, password):
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
model["model_compute_time"] = datetime.datetime.utcnow().isoformat()
slycat.web.server.update_model(database, model)
sid, inputs = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/arrayset_inputs.pickle" % (workdir, uid))
inputs = pickle.loads(inputs)
slycat.web.server.put_model_arrayset(database, model, inputs["aid"])
attributes = inputs["attributes"]
slycat.web.server.put_model_array(database, model, inputs["aid"], 0, attributes, inputs["dimensions"])
for attribute in range(len(attributes)):
sid, data = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/inputs_attributes_data_%s.pickle" % (workdir, uid, attribute))
data = pickle.loads(data)
slycat.web.server.put_model_arrayset_data(database, model, inputs["aid"], "0/%s/..." % attribute, [data])
clusters = json.loads(slycat.web.server.get_remote_file(sid, "%s/slycat_timeseries_%s/file_clusters.json" % (workdir, uid)))
clusters_file = json.JSONDecoder().decode(clusters["file"])
slycat.web.server.post_model_file(model["_id"], True, sid, "%s/slycat_timeseries_%s/file_clusters.out" % (workdir, uid), clusters["aid"], clusters["parser"])
for f in clusters_file:
cherrypy.log.error("Processing file cluster %s" % f)
sid, file_cluster_data = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/file_cluster_%s.json" % (workdir, uid, f))
cherrypy.log.error("Got remote file cluster %s JSON file" % f)
file_cluster_attr = json.loads(file_cluster_data)
cherrypy.log.error("Loaded cluster data JSON file")
slycat.web.server.post_model_file(model["_id"], True, sid, "%s/slycat_timeseries_%s/file_cluster_%s.out" % (workdir, uid, f), file_cluster_attr["aid"], file_cluster_attr["parser"])
cherrypy.log.error("Posted cluster %s OUT file to model" % f)
sid, waveforms = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/waveforms_%s.pickle" % (workdir, uid, f))
cherrypy.log.error("Got remote file waveforms_%s.pickle" % f)
try:
waveforms = pickle.loads(waveforms)
except Exception as e:
cherrypy.log.error("Loading waveforms exception caught: %s" % e)
cherrypy.log.error("Loaded waveforms from pickle file")
database = slycat.web.server.database.couchdb.connect()
model = database.get("model", model["_id"])
cherrypy.log.error("Putting model arrayset for cluster %s" % f)
slycat.web.server.put_model_arrayset(database, model, "preview-%s" % f)
cherrypy.log.error("Starting to enumerate waveforms for %s" % f)
for index, waveform in enumerate(waveforms):
cherrypy.log.error("Processing waveform %s - %s" % (f, index))
sid, waveform_dimensions = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/waveform_%s_%s_dimensions.pickle" % (workdir, uid, f, index))
cherrypy.log.error("Got remote pickle dimensions file %s - %s" % (f, index))
waveform_dimensions = pickle.loads(waveform_dimensions)
cherrypy.log.error("Loaded pickle dimensions file %s - %s" % (f, index))
sid, waveform_attributes = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/waveform_%s_%s_attributes.pickle" % (workdir, uid, f, index))
cherrypy.log.error("Got remote pickle attributes file %s - %s" % (f, index))
waveform_attributes = pickle.loads(waveform_attributes)
cherrypy.log.error("Loaded pickle attributes file %s - %s" % (f, index))
slycat.web.server.put_model_array(database, model, "preview-%s" % f, index, waveform_attributes, waveform_dimensions)
cherrypy.log.error("Put model array for preview-%s" % f)
sid, waveform_times = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/waveform_%s_%s_times.pickle" % (workdir, uid, f, index))
cherrypy.log.error("Got remote pickle times file %s - %s" % (f, index))
waveform_times = pickle.loads(waveform_times)
cherrypy.log.error("Loaded pickle times file %s - %s" % (f, index))
sid, waveform_values = get_remote_file(sid, hostname, username, password, "%s/slycat_timeseries_%s/waveform_%s_%s_values.pickle" % (workdir, uid, f, index))
cherrypy.log.error("Got remote pickle values file %s - %s" % (f, index))
waveform_values = pickle.loads(waveform_values)
cherrypy.log.error("Loaded pickle values file %s - %s" % (f, index))
slycat.web.server.put_model_arrayset_data(database, model, "preview-%s" % f, "%s/0/...;%s/1/..." % (index, index), [waveform_times, waveform_values])
cherrypy.log.error("Put model arrayset data for preview-%s" % f)
def items(self):
return [(pickle.loads(app_rpc), app)
for app_rpc, app in self.applications
if app_rpc != 'transport']
def ExecuteMultilevelMonteCarloAnalisys_Task(
finest_level, pickled_coarse_model, pickled_coarse_parameters,
size_meshes, pickled_settings_metric_refinement,
pickled_settings_remesh_refinement):
'''unpickle model and build Kratos Model object'''
model_serializer = pickle.loads(pickled_coarse_model)
current_model = KratosMultiphysics.Model()
model_serializer.Load("ModelSerialization", current_model)
del (model_serializer)
'''unpickle parameters and build Kratos Parameters object'''
serialized_parameters = pickle.loads(pickled_coarse_parameters)
current_parameters = KratosMultiphysics.Parameters()
serialized_parameters.Load("ParametersSerialization", current_parameters)
del (serialized_parameters)
'''unpickle metric and remesh refinement parameters and build Kratos Parameters objects'''
settings_metric_refinement_serializer = pickle.loads(
pickled_settings_metric_refinement)
settings_remesh_refinement_serializer = pickle.loads(
pickled_settings_remesh_refinement)
current_settings_metric_refinement = KratosMultiphysics.Parameters()
current_settings_remesh_refinement = KratosMultiphysics.Parameters()
settings_metric_refinement_serializer.Load(
"MetricRefinementParametersSerialization",
current_settings_metric_refinement)
settings_remesh_refinement_serializer.Load(
"RemeshRefinementParametersSerialization",
current_settings_remesh_refinement)
del (settings_metric_refinement_serializer,
settings_remesh_refinement_serializer)
'''generate the sample'''
sample = GenerateSample()
'''initialize the MultilevelMonteCarloResults class and prepare the results'''
mlmc_results_class = mlmc.MultilevelMonteCarloResults()
QoI = []
start_MLMC_time = time.time()
end_MLMC_time = []
if (finest_level == 0):
simulation = MultilevelMonteCarloAnalysis(current_model,
current_parameters, sample)
simulation.Run()
QoI.append(EvaluateQuantityOfInterest(simulation))
del (simulation)
end_MLMC_time.append(time.time())
else:
for lev in range(finest_level + 1):
simulation = MultilevelMonteCarloAnalysis(current_model,
current_parameters,
sample)
simulation.Run()
QoI.append(EvaluateQuantityOfInterest(simulation))
end_MLMC_time.append(time.time())
'''refine if level < finest level exploiting the solution just computed'''
if (lev < finest_level):
'''refine the model Kratos object'''
model_refined = refinement.compute_refinement_hessian_metric(
simulation, size_meshes[lev + 1], size_meshes[lev],
current_settings_metric_refinement,
current_settings_remesh_refinement)
'''initialize the model Kratos object'''
simulation = MultilevelMonteCarloAnalysis(
model_refined, current_parameters, sample)
simulation.Initialize()
'''update model Kratos object'''
current_model = simulation.model
del (simulation)
'''prepare results of the simulation in the MultilevelMonteCarloResults class'''
mlmc_results_class.finer_level = finest_level
for lev in range(finest_level + 1):
mlmc_results_class.time_ML.append(end_MLMC_time[lev] - start_MLMC_time)
mlmc_results_class.QoI.append(QoI[lev])
return mlmc_results_class
def query(key):
result = self.col.find_one()
dump = result[key]
return cpickle.loads(dump)
逆操作
store['obj2']
5.9 pickle python对象序列化======
序列化是指把对象的层级几个转换成字节流的过程
5.9.1 用 cPickle实现python对象序列化
import cpickle as pickle
dumps()函数
pickled_data = pickle.dumps(data)
数据序列化后，再写入文件或用套接字，管道等发送就很简单
反序列化 loads()函数
nframe = pickle.loads(pickled_data)
----5.9.2 用pandas实现对象序列化----- ‘’‘重点====
无需导入cpickle模块，所有的操作都是隐士进行的
frame.to_pickle('frame.pkl')
读取数据
pd.read_pickle('trame.pkl')
--------5.10 对接数据库------
from sqlalchemy import create_engine
try:
import cpickle as pickle
except ImportError:
import pickle
s = pickle.dumps(d)
f = open('dump.txt', 'wb')
pickle.dump(d, f)
f.close()
f = open('dump.txt', 'rb')
h = pickle.load(f)
f.close()
pickle.loads(pickle.dumps(d))
import json
w = json.dumps(d)
i = json.loads(w)
import json
class Student(object):
def __init__(self, name, age, score):
self.name = name
self.age = age
self.score = score
#!/usr/lib/python import urllib2, sys try: import cpickle as pickle except ImportError as e: import pickle url = 'http://www.pythonchallenge.com/pc/def/peak.html' if __name__ == "__main__": banner_url = 'http://www.pythonchallenge.com/pc/def/banner.p' try: raw_data = urllib2.urlopen(banner_url).read() data = pickle.loads(raw_data) except urllib2.URLError as e: print e.reason raise for row in data: for (char, freq) in row: sys.stdout.write(char * freq) print
def query(key):
result = self.col.find_one()
dump = result[key]
return cpickle.loads(dump)
