def Peak_Picking_function(acc,Fs,new_f,filtering,PSDfangfa,m,if_log,draw,percent,minpeakdist):
engine.cd(r'/Data/SHZX_output/Matlab_Functions')
[Frequency,PSD,Locs,Peaks] = engine.ANPSD_function_py(
matlab.double(np.array(acc).tolist()),
float(Fs),
'filtering', matlab.double(filtering),
'PSDfangfa', float(PSDfangfa),
'm', float(m),
'if_log', float(if_log),
'draw', float(draw),
'percent', float(percent),
'minpeakdist', float(minpeakdist),
'new_f', float(new_f),
nargout=4)
ANPSD = pd.DataFrame([]); Peak = pd.DataFrame([])
ANPSD['Frequency'] = list(Frequency[0])
ANPSD['ANPSD'] = list(PSD[0])
try:
Peak['f'] = list(Locs[0])
Peak['ANPSD'] = list(Peaks[0])
except:
Peak['f'] = [Locs]
Peak['ANPSD'] = [Peaks]
return(Peak, ANPSD)
def execute(lines):
if type(engine)!=matlab.engine.matlabengine.MatlabEngine:
print "I do not find a matlab engine"
return
oldworkingdir = engine.pwd()
workingdir = vim.eval("expand('%:p:h')")
fscript = workingdir+'/tmpscriptAZWKL.m'
with open(fscript,'w') as script:
for line in lines:
script.write(line+'\n')
result = StringIO.StringIO()
engine.cd(workingdir,nargout=0)
try:
engine.tmpscriptAZWKL(nargout=0,stdout=result,stderr=result)
except matlab.engine.MatlabExecutionError:
pass
engine.cd(oldworkingdir,nargout=0)
remove(fscript)
print result.getvalue()
def SSICOV_function(acc,Ts,Fs,new_f,filtering,if_log,draw,draw_matlab,Xrange,eps_freq):
SSI = pd.DataFrame([])
engine.cd(r'/Data/SHZX_output/Matlab_Functions')
[fn,zeta,phi,plotdata] = engine.SSICOV_function_py(
matlab.double(np.array(acc).tolist()),
float(Fs),
'Ts', Ts, # 这个越长越精准，但速度会受影响
'new_f', float(new_f), # 降采样之后反而效果更好了，可能是因为所受的干扰信息少了
'methodCOV', 1, # 方法1精度更高
'Nmin', float(1), 'Nmax', float(50),
'filtering', matlab.double(filtering),
'draw', draw,
'draw_matlab', draw_matlab,
'Xrange', matlab.double(Xrange),
'if_log', if_log,
'eps_freq', eps_freq,
nargout=4)
if draw:
# stable pole : 频率、阵型、阻尼同时满足精度要求
# stable freq.& MAC : 频率、阵型满足精度要求
# stable freq.& damp.: 频率、阻尼满足精度要求
# stable freq.: 频率满足精度要求
all_pole = pd.DataFrame([[list(x)[0] for x in list(plotdata[0][0])],[list(x)[0] for x in list(plotdata[1][0])]]).T
stable_pole = pd.DataFrame([[list(x)[0] for x in list(plotdata[0][1])],[list(x)[0] for x in list(plotdata[1][1])]]).T
stable_freq_MAC = pd.DataFrame([[list(x)[0] for x in list(plotdata[0][2])],[list(x)[0] for x in list(plotdata[1][2])]]).T
stable_freq_damp = pd.DataFrame([[list(x)[0] for x in list(plotdata[0][3])],[list(x)[0] for x in list(plotdata[1][3])]]).T
stable_freq = pd.DataFrame([[list(x)[0] for x in list(plotdata[0][4])],[list(x)[0] for x in list(plotdata[1][4])]]).T
all_pole['mark'] = 'all pole'; stable_pole['mark'] = 'stable pole'; stable_freq_MAC['mark'] = 'stable freq. & MAC'; stable_freq_damp['mark'] = 'stable freq. & damp.'; stable_freq['mark'] = 'stable freq.'
SSI_data = pd.concat([all_pole,stable_pole,stable_freq_MAC,stable_freq_damp,stable_freq], ignore_index =True); SSI_data.columns = ['frequency', 'mode', 'mark']
PP_data = pd.DataFrame([list(plotdata[2][0]), list(plotdata[3][0])]).T; PP_data.columns = ['frequency', 'mode']
SSI_data = SSI_data[SSI_data['frequency']<=Xrange[1]].drop_duplicates()
PP_data = PP_data[PP_data['frequency']<=Xrange[1]]
else:
SSI_data = ''; PP_data = ''
try:
SSI['fn'] = list(fn[0])
SSI['zeta'] = list(zeta[0])
except:
SSI['fn'] = [fn]
SSI['zeta'] = [zeta]
return(SSI, SSI_data, PP_data)
def get_efms(N, reversibility, verbose=True, efmtool_path=os.getcwd()):
"""
Uses Matlab and EFMtool (https://csb.ethz.ch/tools/software/efmtool.html) to calculate the EFMs of a model.
:param N: stoichiometry matrix
:param reversibility: list of booleans indicating if reaction is reversible
:param verbose:
:param efmtool_path: Path to where EFM software is installed on machine
:return: matrix with efms
"""
import matlab.engine
engine = matlab.engine.start_matlab()
engine.cd(efmtool_path)
result = engine.CalculateFluxModes(matlab.double([list(row) for row in N]), matlab.logical(reversibility))
if verbose:
print('Fetching calculated EFMs')
size = result['efms'].size
shape = size[1], size[0] # _data is in transposed form w.r.t. the result matrix
efms = np.reshape(np.array(result['efms']._data), shape)
if verbose:
print('Finishing fetching calculated EFMs')
return efms
def execute(lines):
if type(engine) != matlab.engine.matlabengine.MatlabEngine:
print("I do not find a matlab engine")
return
oldworkingdir = engine.pwd()
workingdir = vim.eval("expand('%:p:h')")
fscript = workingdir + '/tmpscriptAZWKL.m'
with open(fscript, 'w') as script:
for line in lines:
script.write(line + '\n')
result = io.StringIO()
engine.cd(workingdir, nargout=0)
try:
engine.tmpscriptAZWKL(nargout=0, stdout=result, stderr=result)
except matlab.engine.MatlabExecutionError:
pass
engine.cd(oldworkingdir, nargout=0)
remove(fscript)
message = result.getvalue()
maxOutputChars = int(vim.eval("g:matvim_max_outputchars"))
if maxOutputChars > 0 and len(message) > maxOutputChars:
message = message[0:maxOutputChars-1]+ \
" ...\nWarning: Message is longer than g:matvim_max_outputchars"
print(message)
def run_m(m_path):
main_path = '\\'.join(m_path.split('\\')[:-1])
name = m_path.split('\\')[-1].split('.')[0]
engine.cd('E:\\【论文】\\【小论文】\\宁波南站\\脚本控制器\\操控Matlab')
engine.Control_Matlab(main_path, name)
import os import matlab.engine # TODO: Life would be easier if we used an existing MATLAB engine # so we didn't have to restart MATLAB each time. # https://www.mathworks.com/help/matlab/matlab_external/connect-python-to-running-matlab-session.html engine = matlab.engine.start_matlab() engine.cd(os.path.dirname(os.path.realpath(__file__))) engine.triarea(nargout=0)
model.compile(optimizer=SGD(lr=learning_rate), loss=loss_func)
# load pretrain model
if model_pretrain != '':
model.load_weights(model_pretrain, by_name=True)
# neumf_pretrain = 'Pretrain/gene_NeuMF_8_[64]_0.050.h5'
# model.load_weights(neumf_pretrain,by_name=True)
# print("Load pretrained NeuMF (%s) models done. " %(neumf_pretrain))
# Load pretrain model
# Init performance
engine = matlab.engine.start_matlab()
cur_path = os.getcwd()
engine.cd(cur_path, nargout=0)
cdf_t, recall_t = -1, -1
print('Init: cdf = %.6f, recall = %.6f' % (
cdf_t,
recall_t,
))
best_cdf, best_recall, best_iter = cdf_t, recall_t, -1
# Training model
plot_model(model, to_file='%s.png' % training_model.lower())
model.summary()
for epoch in xrange(1, num_epochs + 1):
t1 = time()
# Generate training instances
user_input, item_input, labels = get_train_instances(
temp_data.iloc[:, 1].mean()).abs() < 3 *
temp_data.iloc[:, 1].std()]
new_data = pd.concat([new_data, temp_data], ignore_index=True)
data = new_data
# 是否取对数
if if_log:
data.iloc[:, 1] = np.log(data.iloc[:, 1])
time_series = np.array(range(1, 1441))
time_stamp = np.array(data.iloc[:, 0])
RMS = np.array(data.iloc[:, 1])
if if_Matlab:
# 高斯过程回归Matlab
engine.cd('E:\\【论文】\\【小论文】\\宁波南站\\Matlab脚本\\GPR')
engine.GaussianProcessRegression(
matlab.double(time_stamp.tolist())[0],
matlab.double(RMS.tolist())[0],
matlab.double(time_series.tolist())[0])
else:
# 高斯过程回归Python
# 抽取子集
data_sample = data.sample(3000)
time_stamp_sample = np.array(data_sample.iloc[:, 0])
RMS_sample = np.array(data_sample.iloc[:, 1])
# 高斯回归
from sklearn.gaussian_process import GaussianProcessRegressor
from sklearn.gaussian_process.kernels import Matern
gp_kernel = 1.0 * Matern(
length_scale=1, length_scale_bounds=(1e-2, 1e2), nu=1.5)
def render_diagram(app, node, docname):
global engine
uri = node['uri']
# do not regenerate
if os.path.exists(uri):
pass
ensuredir(os.path.dirname(uri))
try:
# reuse last engine to save loading time
if engine == None:
engine = matlab.engine.start_matlab()
else:
# clean up used engines
engine.restoredefaultpath(nargout=0)
engine.close('all', nargout=0)
engine.bdclose('all', nargout=0)
engine.clear('classes', nargout=0)
# start engine from document directory
engine.cd(os.path.dirname(app.env.doc2path(docname)))
# then, support changing directory (relative to document)
dir = node.get('dir')
if dir:
engine.cd(dir)
# finally, add the MATLAB paths relative to the changed directory
pathlist = node.get('addpath')
if pathlist:
for path in pathlist:
engine.addpath(path)
# preload script
preload = node.get('preload')
if preload:
engine.eval(preload + ';', nargout=0)
# load system
system = node.get('system')
if system:
engine.load_system(system)
# if subsystem specified, print from this layer
subsystem = node.get('subsystem')
if subsystem:
system = "/".join([system, subsystem])
# print from Simulink handle to .png
engine.eval(
"print( get_param( '{}', 'Handle' ), '-dpng', '{}' )".format(
system, uri),
nargout=0)
except matlab.engine.MatlabExecutionError as err:
raise SimulinkDiagramError('Unable to render Simulink diagram due ' +
'to MATLAB execution error')
from waitress import serve
from flask import request
from flask import Flask
import time
import json
if __name__ == '__main__':
app = Flask(__name__)
# initialize MATLAB
engine = matlab.engine.start_matlab()
engine.cd('./matlab')
engine.addpath('./utils')
@app.route('/api/vrptw', methods=['POST'])
def callVrptw():
req_json = request.json
if req_json is None or req_json == 'null':
return 'Error parsing request! Should be in JSON format!'
incidence_mat = req_json['incidenceMat']
cost_mat = req_json['costMat']
edge_time_vec = req_json['edgeTimeV']
start_vec = req_json['startV']
end_vec = req_json['endV']
distr_vec = req_json['nodeDistributionV']
def render_diagram(app, node, docname):
global engine
uri = node['uri']
# do not regenerate
if os.path.exists( uri ):
pass
ensuredir( os.path.dirname( uri ) )
try:
# reuse last engine to save loading time
if engine == None:
engine = matlab.engine.start_matlab()
else:
# clean up used engines
engine.restoredefaultpath(nargout=0)
engine.close('all', nargout=0)
engine.bdclose('all', nargout=0)
engine.clear('classes', nargout=0)
# start engine from document directory
engine.cd( os.path.dirname( app.env.doc2path( docname ) ) )
# then, support changing directory (relative to document)
dir = node.get('dir')
if dir:
engine.cd( dir )
# finally, add the MATLAB paths relative to the changed directory
pathlist = node.get('addpath')
if pathlist:
for path in pathlist:
engine.addpath( path )
# preload script
preload = node.get('preload')
if preload:
engine.eval( preload + ';', nargout=0)
# load system
system = node.get('system')
if system:
engine.load_system( system );
# if subsystem specified, print from this layer
subsystem = node.get('subsystem')
if subsystem:
system = "/".join( [ system, subsystem ] )
# print from Simulink handle to .png
engine.eval(
"print( get_param( '{}', 'Handle' ), '-dpng', '{}' )".
format( system, uri ),
nargout=0
)
except matlab.engine.MatlabExecutionError as err:
raise SimulinkDiagramError('Unable to render Simulink diagram due ' +
'to MATLAB execution error'
)
