Basic LiF Sims

Data_Sim_Basic_Lif.py

# -*- coding: utf-8 -*-

import numpy as np
import matplotlib.pyplot as plt
import json
import pandas as pd


pd_csv = pd.read_csv('face_clean_with_demo.csv')

##############################################################################

def matrixpush(m, row):
    if not np.all(np.isfinite(values[:,0])):
        i = np.count_nonzero(np.logical_not(np.isnan(values[:,0])))
        m[i,] = row
    else:
        m = np.vstack([m,row])
        m = m[1:,]
        
    return(m)

def getobs( x, max = 5, err=0 ):
    if (err==0):
        obsr = -pow((x-max),2)
    else:
        obsr = -pow((x-max),2) + np.random.normal(0,err,1)
    return obsr;

##############################################################################

inttime = 150                     # Integration time
amplitude = 8                     # Amlitude LIF
learnrate = 0.0006                # Learnrate
x0 = 1.0                          # Startvalue

##############################################################################

variance = 1                       # variance in observations

##############################################################################

values = np.zeros((inttime,4))
values.fill(np.nan)


track_x0 = []
track_x  = []
track_t  = []
track_y  = []

x = 0.0
t = 0.0
y = 0.0

omega_1 = 2.63
omega_2 = 2.51

omega = (omega_1 + omega_2)/2

##############################################################################

for i in range(0,len(pd_csv.t)):

    t = pd_csv.t[i]

    x = x0 + amplitude * np.cos(omega * t)

    y = amplitude * np.cos(omega * t) * pd_csv.y[i]


    if np.all(np.isfinite(values[:,0])):
        x0 = x0 + learnrate * sum( values[:,2] ) /inttime

    row_to_add = np.array([t, x, y, x0])
    values = matrixpush(values, row_to_add)


    track_t  = np.append(track_t, t)
    track_x  = np.append(track_x, x)
    track_y  = np.append(track_y, y)
    track_x0 = np.append(track_x0, x0)


##############################################################################

# plot some vars

plt.plot(track_x)
plt.show()

plt.plot(track_x0)
plt.show()

# print final x0

print(x0)

#print(values)

print ()

##############################################################################


def _np_nan_fill(rows,columns):
        nan_values = np.zeros((rows,columns))
        nan_values.fill(np.nan)
        nan_values = json.dumps(nan_values.tolist())
        return nan_values

Local_Server_Sim_Basic_Lif.py

# -*- coding: utf-8 -*-

import numpy as np
import matplotlib.pyplot as plt
import json

##############################################################################

def matrixpush(m, row):
    if not np.all(np.isfinite(values[:,0])):
        i = np.count_nonzero(np.logical_not(np.isnan(values[:,0])))
        m[i,] = row
    else:
        m = np.vstack([m,row])
        m = m[1:,]
        
    return(m)

def getobs( x, max = 5, err=0 ):
    if (err==0):
        obsr = -pow((x-max),2)
    else:
        obsr = -pow((x-max),2) + np.random.normal(0,err,1)
    return obsr;

##############################################################################

stream = 400                      # Length of stream
inttime = 100                     # Integration time
amplitude = 1.4                   # Amlitude LIF
learnrate = .004                  # Learnrate
omega = 0.8                       # Omega
x0 = 1.0                          # Startvalue

##############################################################################

variance = 1                       # variance in observations

##############################################################################

values = np.zeros((inttime,4))
values.fill(np.nan)


track_x0 = []
track_x  = []
track_t  = []
track_y  = []

x = 0.0
t = 0.0
y = 0.0

##############################################################################

for i in range(0,stream):

    t = i+1

    x = x0 + amplitude * np.cos(omega * t)

    y = amplitude * np.cos(omega * t) * getobs(x, 5, 1)



    if np.all(np.isfinite(values[:,0])):
        #x0 = np.mean(values[:,1])                # either or
        x0 = x0 + learnrate * sum( values[:,2] ) /inttime  # /inttime

    row_to_add = np.array([t, x, y, x0])
    values = matrixpush(values, row_to_add)


    track_t  = np.append(track_t, t)
    track_x  = np.append(track_x, x)
    track_y  = np.append(track_y, y)
    track_x0 = np.append(track_x0, x0)


##############################################################################

# plot some vars

plt.plot(track_x)
plt.show()

plt.plot(track_x0)
plt.show()

# print final x0

print(x0)

#print(values)

##############################################################################

def _np_nan_fill(rows,columns):
        nan_values = np.zeros((rows,columns))
        nan_values.fill(np.nan)
        nan_values = json.dumps(nan_values.tolist())
        return nan_values

Remote_Server_Sim_Basic_Lif.py

# -*- coding: utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import urllib,json,codecs
import pymongo
import ssl


ssl._create_default_https_context = ssl._create_unverified_context
MONGO_IP = "78.46.212.194"
MONGO_PORT = 27017

def do_chart(qr):

    client = pymongo.MongoClient(MONGO_IP, MONGO_PORT)
    db = client.logs
    
    fig = plt.figure(figsize=(4.8,4))
    ax = fig.add_subplot(1,1,1)
    ax.tick_params(which = 'both', direction = 'out')
    ax.grid(which='both')

    cursor = db.logs.find({"type": "setreward","q":int(qr)}) \
                .sort([("t", pymongo.ASCENDING)])
    result_list = list(cursor)
    client.close();
    t = [ts['t'] for ts in result_list]
    x0 = [xs['x0'] for xs in result_list]
    plt.plot(t,x0)
    plt.show()

def getobs( x, max = 5, err=0 ):
    if (err==0):
        obsr = -1*pow((x-max),2)
    else:
        obsr = -1*pow((x-max),2) + np.random.normal(0,err,1)
    return obsr;

BASE_URL = 'https://strm.mnds.org:8080'
key = "784537c64"    
exp_id = 2
question_nr = 197152212     

QID = question_nr 

##############################################################################

stream = 400                      # Length of stream
inttime = 100                     # Integration time
A = 1.4                           # Amlitude LIF
gamma = .004                      # Learnrate
omega = 0.8                       # Omega
x0 = 1.0                          # Startvalue

##############################################################################

p_return = 0.80                    # Chance that returns
variance = 1                       # variance in observations

##############################################################################


picked = "T"
lifv = 2


track_x0 = []
track_x  = []
track_t  = []
track_y  = []

x = 0.0
t = 0.0
y = 0.0

for i in range(0,stream):  
    
   request =  BASE_URL + "/" + str(exp_id) 
   request += "/getAction.json?key="+key
   request += "&context={\"question\":"+str(question_nr)
   request += ",\"T\":"+str(inttime)
   request += ",\"gamma\":"+str(gamma)
   request += ",\"omega\":"+str(omega)
   request += ",\"lifv\":"+str(lifv)   
   request += ",\"QID\":"+str(QID)+",\"x0\":"+str(x0)+",\"A\":"+str(A)+"}"
   
   print (request)
   
   response = urllib.request.urlopen(request)
   reader = codecs.getreader("utf-8") 
   obj = json.load(reader(response))
   
   t =  (obj["action"]["t"])
   x =  (obj["action"]["x"])
   
   if np.random.binomial(1, p_return, 1)==1: 
  
       y = getobs(x,5,variance)
       
       request =  BASE_URL + "/" + str(exp_id) + "/setReward.json"
       request += "?key="+key
       request += "&context={\"question\":"+str(question_nr)
       
       request += ",\"T\":"+str(inttime)
       request += ",\"gamma\":"+str(gamma)
       request += ",\"omega\":"+str(omega)
       request += ",\"lifv\":"+str(lifv)
  
       request += ",\"dvalue\":\""+str(picked)+"\""
       request += ",\"QID\":"+str(QID)+",\"x0\":"+str(x0)+",\"A\":"+str(A)+"}"

       request += "&action={\"x\":" + str(float(x))
       request += ",\"t\":" + str(float(t)) + "}"
       request += "&reward=" + str(float(y))    
       
       print (request)
              
       response = urllib.request.urlopen(request)
       reader = codecs.getreader("utf-8") 
       obj = json.load(reader(response)) 

   track_x  = np.append(track_x, x)

plt.plot(track_x)
plt.show()

do_chart(question_nr)