virtual env -p python3 venv
source venv/bin/activate
pip install -r requeriments.txt
python setup.py install

from energyOptimal.monitor import monitorProcess
from energyOptimal.cpufreq import cpuFreq
try:
    process= monitorProcess(program_name_= 'example/program.elf', sensor_type_='ipmi')
    args= [['arg1','arg2'],['arg3','arg4'],...]
    thrs= [1]+list(range(2,33,2))

    process.run_dvfs(list_threads= thrs, list_args= args, idle_time= 30,
    verbose=2, save_name='dvfs/{}_complet.pkl'.format(p.program_name))

    # Power and Performance model
    process.run(list_threads= thrs, list_args= args, idle_time= 30,
    verbose=2, save_name='performance/{}_complet.pkl'.format(p.program_name))
except KeyboardInterrupt:
    cpu= cpuFreq()
    cpu.reset()

from energyOptimal.powerModel import powerModel

pw_model= powerModel()
# Fit power model
pw_model.loadData(filename='power_model/ipmi_2-32_cpuload.pkl',load_sensors=False)
pw_model.fit()
pw_model.save('ipmi_2-32_cpuload.pw')
# Load power model
pw_model.load('ipmi_2-32_cpuload.pw')

error= pw_model.error()
print("Error, ", error)
print("Power model constants, ", pw_model.power_model_x)
res= pw_model.estimate(pw_model.frequencies, pw_model.threads)

from energyOptimal.performanceModel import performanceModel

en_model= performanceModel()
# Fit model
df= en_model.loadData(filename='performance_model/program.pkl', arg_num=idx)
en_model.fit(C_=10e3,gamma_=0.1)
scores= en_model.crossValidate(method='mpe')
print("CrossValidation ", np.mean(scores)*100, scores)
en_model.saveDataframe('dataframes/program'+l)
en_model.saveSVR('svr/program'+l)

# Load model
df= en_model.loadDataFrame('dataframes/program')
en_model.loadSVR('svr/program')

# Predict
preds= en_model.estimate(df[['freq', 'thr', 'in_cat']].values)

from energyOptimal.energyModel import energyModel

pw_model= powerModel('data/ipmi_2-32_cpuload.pw')
perf_model= performanceModel('dataframes/program', 'svr/program')
en_model= energyModel(pw_model,perf_model)

print(en_model.minimalEnergy())