mailsys_server/kpi.py

import os
import time
import math
import sys
import numpy as np
from numpy.lib.function_base import append
import pandas as pd
from pandas import DataFrame as df
from sklearn.cluster import DBSCAN
import pymssql
#import matplotlib.pyplot as plt


ssd_dir = str(sys.argv[1])
file_path = '/hdd/' + ssd_dir[5:]
count = 1

def put_kpi_to_db(kpi):
    conn = pymssql.connect(server='52.231.39.219',port=1433, user='mc365', password='tkadbrdhmc1!', database='master')
    
    cursor = conn.cursor()
    cursor.execute("INSERT INTO tsfmc_mailsystem.dbo.KPI(surgeryID, totalStrokeCount, totalStrokeVelocity, totalStrokeDepth, upDownTotalStrokeCount, upDownTotalStrokeVelocity, upDownTotalStrokeDepth, leftRightTotalStrokeCount, leftRightTotalStrokeVelocity, leftRightTotalStrokeDepth, quadrant1TotalStrokeCount, quadrant1TotalStrokeVelocity, quadrant1TotalStrokeDepth, quadrant2TotalStrokeCount, quadrant2TotalStrokeVelocity, quadrant2TotalStrokeDepth, quadrant3TotalStrokeCount, quadrant3TotalStrokeVelocity, quadrant3TotalStrokeDepth, quadrant4TotalStrokeCount, quadrant4TotalStrokeVelocity, quadrant4TotalStrokeDepth) VALUES ("+str(kpi[0])+", "+kpi[1]+", "+kpi[2]+", "+kpi[3]+", "+kpi[4]+", "+kpi[5]+", "+kpi[6]+", "+kpi[7]+", "+kpi[8]+", "+kpi[9]+", "+kpi[10]+", "+kpi[11]+", "+kpi[12]+", "+kpi[13]+", "+kpi[14]+", "+kpi[15]+", "+kpi[16]+", "+kpi[17]+", "+kpi[18]+", "+kpi[19]+", "+kpi[20]+", "+kpi[21]+")")  
    conn.commit()
    conn.close()


def get_quad(df):
    if df['x'] >= 0 and df['y'] >= 0:
        return 1
    elif df['x'] >= 0 and df['y'] < 0:
        return 2
    elif df['x'] < 0 and df['y'] < 0:
        return 3
    elif df['x'] < 0 and df['y'] >= 0:
        return 4
def get_inc_num(df):
    global count
    prev_count = 0
    if abs(df['diff_diff']) > 0.0008:                
        prev_count = df['inc_num']
        df['inc_num'] = count
        #print(prev_count)
        return count        
    else:
        if prev_count != 0:            
            count += 1
            df['inc_num'] = 0
            prev_count = df['inc_num']
            return 0
        else:
            df['inc_num'] = 0
            prev_count = df['inc_num']
            return 0


def count_inc(df):
    e= 0
    count  = 1
    prev_count = 0
    for x in df['diff_diff']:
        if abs(x) > 8:
            df['inc_num'][e] = count
            prev_count = count
            e +=1
        else:
            if prev_count != 0:
                count += 1
                df['inc_num'][e] = 0
                prev_count = 0
                e +=1
            else:
                df['inc_num'][e] = 0
                prev_count = 0
                e +=1

def stroke_count(df):
    e = 0
    count = 1    
    for x in df['diff_diff']:
        if x > 1.2:
            df['stroke_count'][e] = count
            e +=1
        else:
            df['stroke_count'][e] = 0
            e+=1
            
def get_sqrt(df):
    e =0
    count = 1
    prev_x = 0.0
    prev_y = 0.0
    prev_z = 0.0
    for x in range(len(df)):
        df['sqrt'][e] = np.sqrt((df['x'][e] - prev_x) **2 + (df['y'][e] - prev_y) ** 2 +(df['z'][e] - prev_z) ** 2)
        prev_x = df['x'][e]
        prev_y = df['y'][e]
        prev_z = df['z'][e]
        e += 1

        


def dist():
    file = os.path.join(file_path, 'coordinate.csv')
    norm = pd.read_csv(file)
    norm = norm.drop(norm.columns[0], axis=1)
    norm = norm.fillna(0)
    data = norm[['x', 'y', 'z']]
    norm['sqrt'] = 0.0

    get_sqrt(norm)

    #norm = norm.apply(lambda x:np.sqrt(x) if x.name in ['sqrt'] else x, axis =1)
    norm['diff'] = norm.diff()['sqrt']
    norm['diff_diff'] = norm.diff()['diff']


    norm['quad'] = norm.apply(get_quad, axis =1)
    
    
    norm = norm.fillna(0)

    norm['inc_num'] = 0
    #norm['inc_num'] = norm.apply(get_inc_num, axis = 1)
    count_inc(norm)
    norm['stroke_count'] = 0
    stroke_count(norm)
    
    #print('num of inc')
    #print(norm['inc_num'].unique())

    #print('total stroke')
    #print(norm['stroke_count'].unique())
    #print(norm)

    

    #plt.plot(norm['frame'],norm['diff_diff'])
    #plt.savefig('/sources/diff_diff.png')



    kpi = pd.DataFrame(index=range(0,1))

    quad_count = norm[norm['stroke_count'] != 0]
    
    quad1_c = quad_count[quad_count['quad'] == 1]
    quad2_c = quad_count[quad_count['quad'] == 2]
    quad3_c = quad_count[quad_count['quad'] == 3]
    quad4_c = quad_count[quad_count['quad'] == 4]
    #print(len(quad1_c))
    #print(len(quad2_c))
    #print(len(quad3_c))
    #print(len(quad4_c))
    #print(quad1_c)
    kpi['PK'] = ssd_dir[5:] 
    kpi['qaud1_c'] = len(quad1_c)
    kpi['qaud2_c'] = len(quad2_c)
    kpi['qaud3_c'] = len(quad3_c)
    kpi['qaud4_c'] = len(quad4_c)
    #FIXME: this values are huristic value. please change this values after research
    kpi['qaud1_s'] = quad1_c['sqrt'].mean() / 30
    kpi['qaud2_s'] = quad2_c['sqrt'].mean() / 30
    kpi['qaud3_s'] = quad3_c['sqrt'].mean() / 30
    kpi['qaud4_s'] = quad4_c['sqrt'].mean() / 30
    kpi['qaud1_d'] = quad1_c['sqrt'].mean() / 45
    kpi['qaud2_d'] = quad2_c['sqrt'].mean() / 45
    kpi['qaud3_d'] = quad3_c['sqrt'].mean() / 45
    kpi['qaud4_d'] = quad4_c['sqrt'].mean() / 45


    kpi_1 = {
        "stroke": [len(quad1_c),len(quad2_c),len(quad3_c),len(quad4_c)],
        "velocity": [quad1_c['sqrt'].mean() / 30, quad2_c['sqrt'].mean() / 30, quad3_c['sqrt'].mean() / 30, quad4_c['sqrt'].mean() / 30],
        "depth": [quad1_c['sqrt'].mean() / 45, quad2_c['sqrt'].mean() / 45, quad3_c['sqrt'].mean() / 45, quad4_c['sqrt'].mean() / 45]
    }


    #print(kpi)
    val_list = []

    #PK =
    val_list.append(str(ssd_dir[5:-1]))

    #t_stroke = 
    val_list.append(sum(kpi_1['stroke']))
    #t_vel
    val_list.append("{:.4f}".format(sum(kpi_1['velocity'])/4)) 
    #t_dep
    val_list.append("{:.4f}".format(sum(kpi_1['depth'])/4)) 
    #ud_s_rate =
    val_list.append("{:.4f}".format((kpi_1['stroke'][0] + kpi_1['stroke'][3] - kpi_1['stroke'][1] - kpi_1['stroke'][2]) / sum(kpi_1['stroke'])*100))
    #ud_v_rate =
    val_list.append("{:.4f}".format((kpi_1['velocity'][0] + kpi_1['velocity'][3] - kpi_1['velocity'][1] - kpi_1['velocity'][2])/2))
    #ud_d_rate =
    val_list.append("{:.4f}".format((kpi_1['depth'][0] + kpi_1['depth'][3] - kpi_1['depth'][1] - kpi_1['depth'][2])/2))

    #lr_s_rate =
    val_list.append("{:.4f}".format((kpi_1['stroke'][2] + kpi_1['stroke'][3] - kpi_1['stroke'][0] - kpi_1['stroke'][1]) / sum(kpi_1['stroke'])*100))
    #lr_v_rate =
    val_list.append("{:.4f}".format((kpi_1['velocity'][2] + kpi_1['velocity'][3] - kpi_1['velocity'][0] - kpi_1['velocity'][1])/2))
    #lr_d_rate =
    val_list.append("{:.4f}".format((kpi_1['depth'][2] + kpi_1['depth'][3] - kpi_1['depth'][0] - kpi_1['depth'][1])/2))

    #q1 s, v, d
    val_list.append(len(quad1_c))
    val_list.append(quad1_c['sqrt'].mean() / 30)
    val_list.append(quad1_c['sqrt'].mean() / 45)
    val_list.append(len(quad2_c))
    val_list.append(quad2_c['sqrt'].mean() / 30)
    val_list.append(quad2_c['sqrt'].mean() / 45)
    val_list.append(len(quad3_c))
    val_list.append(quad3_c['sqrt'].mean() / 30)
    val_list.append(quad3_c['sqrt'].mean() / 45)
    val_list.append(len(quad4_c))
    val_list.append(quad4_c['sqrt'].mean() / 30)
    val_list.append(quad4_c['sqrt'].mean() / 45)




    return val_list

    #print(val_list)



if __name__ == "__main__":
    kpi = dist()
    put_kpi_to_db(kpi)