first commit

DayCutter.py

@@ -0,0 +1,36 @@
+# -*- coding: utf-8 -*-
+"""
+Spyder Editor
+
+This is a temporary script file.
+"""
+
+import os
+import pandas as pd
+
+WeekPath = 'D:\AcA Mike Dijkhof\cwa files\Pt203_csv' # path to directory with .csv file of entire week
+ScriptPath = 'D:\AcA Mike Dijkhof\Scripts' # path to directory of formules.py
+
+os.chdir(ScriptPath)
+
+import formules
+
+os.chdir(WeekPath)
+
+#%%
+
+for subdir, dirs, files in os.walk(WeekPath):
+    print(subdir)
+    os.chdir(subdir)
+    for i in files:
+        print(i)
+        if i.startswith('._'): #skip ._ files that may be present in the folder
+            continue
+         
+        df = pd.read_csv(i, header=0, names=['Datetime', 'Acc X','Acc Y', 'Acc Z'], dtype={"Datetime": str, "Acc X": 'float32', "Acc Y": 'float32', "Acc Z": 'float32'}, infer_datetime_format=True)
+        df['Datetime'] = pd.to_datetime(df['Datetime'])
+        df['Date'] = [d.date() for d in df['Datetime']]
+        df = df.reindex(columns=['Datetime','Date','Time','Acc X','Acc Y', 'Acc Z'])      
+            
+        formules.CreateDays(df, i, subdir)
+        os.chdir(subdir)

Totalplotter.py

@@ -0,0 +1,61 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Sat Jun 26 13:27:00 2021
+
+@author: D_Mik
+"""
+
+import os
+import pandas as pd
+import formules
+import numpy as np
+import matplotlib.pyplot as plt
+
+# provide directory for formula parameters
+rootdir = 'D:\AcA Mike Dijkhof\cwa files\Pt304_csv\Pt304_week_formula'
+
+d = {}
+X = list()
+
+# import function parameters from previous scripts
+
+for subdir, dirs, files in os.walk(rootdir):
+    for file in files:
+        print(file)
+        os.chdir(subdir)
+        
+        name = file.replace('Formula_','')
+        name = name.replace('.csv','')
+        
+        df = pd.DataFrame(pd.read_csv(file))
+        df = df.set_index('Name')
+        
+        scale = df['ENMOmax'].max()
+        
+        key = name
+        
+        d[key] = df.loc['Mean']
+        X.append(scale)
+
+# plot curves 
+plt.figure(dpi=720)
+
+for key in d:
+    
+    formula = d[key]
+    Xscale = np.arange(0, max(X), 5)
+    
+    Y = formules.func(Xscale, formula.loc['a'], formula.loc['b'], formula.loc['c'] )
+    
+   
+    plt.ylim(0,1440)
+    plt.xlim(0,(max(X)+10))
+    plt.plot(Xscale, Y, label=key)
+
+plt.legend()
+plt.title('All 6 weeks plotter for 304') #Change to pt-number
+plt.show()
+
+plt.savefig(fname='Total_weeks_304.png') 
+    
+        

formules.py

@@ -0,0 +1,135 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Wed Jun 16 09:34:35 2021
+
+@author: -
+"""
+import osa
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+from datetime import timedelta
+from scipy.optimize import curve_fit
+
+# Formulas
+
+def ReadCSV(filename):
+    df = pd.read_csv(filename, names=['Datetime', 'Acc X','Acc Y', 'Acc Z'], infer_datetime_format=True)
+    df['Datetime'] = pd.to_datetime(df['Datetime'])
+    df['Date'] = [d.date() for d in df['Datetime']]
+    df = df.reindex(columns=['Datetime','Date','Time','Acc X','Acc Y', 'Acc Z'])
+    return df
+
+def CreateDays(x, filename, path):
+    
+    savename = filename.replace('.csv','')
+    savepath = path + savename
+    os.makedirs(savepath)
+    os.chdir(savepath)
+    
+    startdate = x['Date'].iloc[0]
+    week = range(1,8)
+    
+    for i in week:
+        weekdayindex = i-1
+        
+        day = startdate + timedelta(days=weekdayindex)
+        daydate = x['Date'] == startdate + timedelta(days=weekdayindex)
+        dataday = x[daydate]
+        totalweek = {day:dataday}
+        
+        savefile = totalweek[day]
+        varname = filename.replace('.csv','-') + str(day) + '.csv'
+        savefile.to_csv(varname)
+        
+        print(varname +' saved')
+    
+    return(totalweek)
+
+
+def SVMEpoch(DF,ResampRate, ResampData):
+    newDF = pd.DataFrame(DF)
+    newDF['X2'] = np.power(newDF['Acc X'], 2)
+    newDF['Y2'] = np.power(newDF['Acc Y'], 2)
+    newDF['Z2'] = np.power(newDF['Acc Z'], 2)
+    newDF['SVM'] = np.sqrt(newDF[['X2', 'Y2', 'Z2']].sum(axis=1))
+    newDF['Datetime'] = pd.to_datetime(newDF['Datetime'])
+   
+    EpochSVM = newDF.resample(ResampRate, on = ResampData).mean()
+    return(newDF, EpochSVM)
+
+def func(x, a, b, c):
+    return a * np.exp(-b*x) + c
+
+def SlopeWeeker(Keylist, Dict):
+    try:    
+        SlopeWeek = pd.DataFrame(columns=['a','b', 'c', 'Name'])
+        SlopeWeek = SlopeWeek.set_index('Name')
+            
+        for key in Keylist:
+            newDF, EpochSVM = SVMEpoch(Dict[key], '60S', 'Datetime')
+           
+            ENMO = EpochSVM['SVM']-1
+            ENMO = ENMO*1000
+           
+            for value in ENMO:
+                if value < 0:
+                    value = 0
+           
+            BinSize = 5 
+           
+            ENMOmax = int(ENMO.max())
+            
+            if ENMOmax % BinSize == 0:
+                ENMOmax = ENMOmax+1 #to make sure that interference with binsize is impossible
+            
+            MaxBin = int(ENMOmax/BinSize)+1
+            ENMO = ENMO.astype(int)
+           
+            Counter = pd.DataFrame(np.zeros((1,MaxBin)))
+           
+            for x in Counter:
+                Count = (x+1)*BinSize
+                Start = Count - BinSize
+                Number = ENMO.between(Start, Count).sum()
+                Counter[x] = Number
+            
+            Counter = Counter.to_numpy()
+            Counter = Counter.astype(float)
+            Counter = Counter.flatten()
+            
+            Xscale = np.arange(0,ENMOmax, BinSize)
+            Xscale = Xscale.astype(float)
+            
+            popt, _ = curve_fit(func, Xscale, Counter, p0=None) # fit curve through points 
+            a, b, c = popt
+            
+            Trendline = func(Xscale, a, b, c)
+           
+            SlopeWeek.loc[key, 'a'] = a
+            SlopeWeek.loc[key, 'b'] = b
+            SlopeWeek.loc[key, 'c'] = c
+            SlopeWeek.loc[key, 'ENMOmax'] = ENMOmax
+           
+            PtName = key.replace('35694_00000', '')
+            PtName = PtName.replace('resampled-','') 
+            PtName = PtName.replace('.csv','')
+            
+            plt.figure()
+            plt.ylim(0,1440)
+            plt.xlim(0,(ENMOmax+10))
+            plt.title('Intensity plot ' + PtName)
+            plt.xlabel('Movement intensity [bins of ' + str(BinSize) + ' mg]')
+            plt.ylabel('Amount of time spend at intensity [min]')
+            plt.grid()
+            plt.scatter(Xscale, y=Counter)
+            plt.plot(Xscale, Trendline, 'r--')
+            plt.show()
+            
+            PtName = (PtName + '.png')
+            plt.savefig(fname=PtName)
+            
+    except:
+        print(PtName + ' could not be used')
+
+    return SlopeWeek

plotter.py

@@ -0,0 +1,74 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Wed Jun 16 13:03:03 2021
+
+@author: -
+"""
+
+import os
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+
+os.chdir('D:\AcA Mike Dijkhof\Scripts') # set path to folder of formules.py
+
+import formules
+    
+rootdir = 'D:\AcA Mike Dijkhof\cwa files\Pt204_csv\Pt204_csv35694_0000020406.resampled' # provide path to .csv files of the cut weekdays 
+d = {}
+
+for subdir, dirs, files in os.walk(rootdir):
+    print(subdir)
+    for file in files:
+        print(file)
+        os.chdir(subdir)
+        
+        d[file] = pd.read_csv(file, infer_datetime_format=True)
+
+Keys = d.keys()
+
+#%%
+
+Check = formules.SlopeWeeker(Keys, d)
+Worklist = Check
+Length = pd.DataFrame(np.zeros(((1,int(len(Check)/7)))))
+
+# Create PtName, this depends on subdir string so has to be changed for each patient
+PtName = subdir.replace('35694_00000', '')
+PtName = PtName.replace('D:\AcA Mike Dijkhof\cwa files\Pt204_csv\Pt204_','') 
+PtName = PtName.replace('resampled-','')
+PtName = PtName.replace('.resampled','')
+PtName = PtName.replace('.csv','')
+
+for i in Length:
+   
+    CheckWeek = Worklist[:7]
+   
+    Worklist = Worklist.drop(CheckWeek.index, axis=0)
+   
+    CheckWeek.loc['Mean','a'] = CheckWeek['a'].median()
+    CheckWeek.loc['Mean','b'] = CheckWeek['b'].median()
+    CheckWeek.loc['Mean','c'] = CheckWeek['c'].median()
+    
+    CheckWeek.to_csv('Formula_Week6.csv')
+   
+    Xscale = np.arange(0,CheckWeek['ENMOmax'].max(), 5)
+   
+    plt.figure()
+    plt.ylim(0,1440)
+    plt.xlim(0,CheckWeek['ENMOmax'].max())
+    plt.title('All weekdays and average plotted ' + PtName)
+    plt.xlabel('Movement intensity [bins of 5 mg]')
+    plt.ylabel('Amount of time spend at intensity [min]')
+    plt.grid()
+   
+    for i, r in CheckWeek.iterrows():
+        Y = formules.func(Xscale, CheckWeek.loc[i,'a'],CheckWeek.loc[i,'b'], CheckWeek.loc[i,'c'] )
+        if i != 'Mean':
+            plt.plot(Xscale, Y, 'grey')
+        else:
+            plt.plot(Xscale, Y, 'k--')
+            
+    plt.show()
+    plt.savefig(fname=('Weekplot ' + PtName+ '.png'))
+