def getMeanMedSDFreMaxMinSound(data): #create special for only sound no sound #get the mean median sd and frequency in the set of data (x,y) = data.shape mean = np.zeros((2)) median = np.zeros((2)) std = np.zeros((2)) fre = np.zeros((2)) maxi = np.zeros((2)) mini = np.zeros((2)) desData = (mean,median,std,fre,maxi,mini) setOfAction = ATS.countAction(data,[1]) desData = ATS.updateMeanMedSDFreMaxMin(desData,0,setOfAction) setOfAction = ATS.countAction(data,[3]) desData = ATS.updateMeanMedSDFreMaxMin(desData,1,setOfAction) return desData
def updateDiscriptiveData(ori,data,GoT,nameFile,dyad,section): start = 0 end = 0 i = 0 (x,y) = data.shape ####test### # data[x-7][2] = 2 # data[x-6][2] = 2 # data[x-5][2] = 2 count = 0 while i < x: if i == x-1 or data[i+1][1] != data[start][1] : end = i line = [] line.append(int(nameFile)) line.append(dyad) line.append(section) line.append(count) line.append(data[start][0]) #"end-1" is because to prevent out of bound for "end = x" #+1 at the is because we want to take into acount of the last number. Ex, from 0 to 20, there is 21 # line.append(data[end][0]) line.append(end-start+1) line.append(data[start][1]) #get the actions of either gaze or talk of the line action action1 = data[start:end+1,[0,2]] action2 = data[start:end+1,[0,3]] if GoT == 0: line = ATS.updateDiscriptiveData(line,action1,1) else: line = ATS.updateDiscriptiveData(line,action1,0) line = ATS.updateDiscriptiveData(line,action2,1) start = end + 1 count = count + 1 ori.append(line) i = i + 1 return ori
def getReliability(inpath, Start, End, GoTSize):
BoC = np.zeros((End, GoTSize), dtype=np.float64) #best of coder
ori = []
for i in ["52", "65", "79"]:
for j in ["C1", "C2"]:
for root, dirs, files in walk(inpath):
BoC = np.zeros((End, GoTSize),
dtype=np.float64) #best of coder
for file in files:
if file.endswith(".csv"):
if i in file and j in file:
print(file)
nameOutput = file
data = np.genfromtxt(join(root, file),
dtype=np.float64,
delimiter=',')
dyad = int(splitext(file)[0][2:4])
conv = int(splitext(file)[0][6:7])
coder = int(getCoderName(splitext(file)[0][15:17]))
beg = getDyadCovCoderInfor(dyad, conv, coder,
End) #beg is begining
data = AT.shortenData(data, Start, End)
combData = AT.combineData(beg, data, End)
ori.extend(combData)
BoC = bestOfCoderTag(BoC, data, coder, End)
data = getBestOfCoderTag(BoC, End)
beg = getDyadCovCoderInfor(dyad, conv, 105, End)
combData = np.c_[beg, data]
ori.extend(combData)
time = np.arange(End).reshape([End, 1])
output = np.c_[time, data]
np.savetxt(nameOutput + "_combined.csv",
output,
fmt='%1.3f',
delimiter=',')
return ori
def getDiscriptiveWithWidth(inpath, Start, End, nameFiles, DynEnd, Width, GazeType):
# get discription information of the data
# width mean how long do you want to check the characteristics
# DynEnd dynamic ending
# GazeType
ori = []
for nameFile in nameFiles:
print (nameFile)
for i in [1, 2]:
print (i)
(dataTHroot, dataTSroot, dataGroot) = ATS.readInput(inpath, nameFile, i)
if DynEnd: # using dynamic ending
(yTH, xTH) = (dataTHroot.shape)
(yTS, xTS) = (dataTSroot.shape)
(yG, xG) = (dataGroot.shape)
End = np.amin(np.asarray([yTH, yTS, yG])) # get the actual end time
dataTHroot = ATS.shortenData(dataTHroot, Start, End)
dataTSroot = ATS.shortenData(dataTSroot, Start, End)
dataGroot = ATS.shortenData(dataGroot, Start, End)
j = 0
while j < End:
startSub = j
if j + Width > End:
endSub = End
else:
endSub = j + Width
dataTH = ATS.shortenData(dataTHroot, startSub, endSub)
dataTS = ATS.shortenData(dataTSroot, startSub, endSub)
dataG = ATS.shortenData(dataGroot, startSub, endSub)
dataGTH = ATS.combineData(dataG, dataTH, endSub)
dataGTHS = ATS.combineData(dataGTH, dataTS, endSub)
# print (dataGTHS)
j = j + Width
# for i in ori:
# print (i)
return ori
def getDiscriptiveTwoTalkWithWidth(inpath, Start,End, nameFiles, DynEnd, Width,GoT): #get discription information of the data #width mean how long do you want to check the characteristics #DynEnd dynamic ending ori = [] for nameFile in nameFiles: print (nameFile) for i in [1,2]: print (i) (dataTHroot, dataTSroot, dataGroot) = ATS.readInput(inpath, nameFile,i) if DynEnd: #using dynamic ending (yTH, xTH) = (dataTHroot.shape) (yTS, xTS) = (dataTSroot.shape) (yG, xG) = (dataGroot.shape) End = np.amin(np.asarray([yTH,yTS,yG])) #get the actual end time dataTHroot = ATS.shortenData(dataTHroot,Start,End) dataTSroot = ATS.shortenData(dataTSroot,Start,End) dataGroot = ATS.shortenData(dataGroot,Start,End) dataTHSroot = combinedTwoTalkTogether(dataTHroot,dataTSroot) j = 0 while j < End: startSub = j if j+Width > End: endSub = End else: endSub = j+Width dataG = ATS.shortenData(dataGroot,startSub,endSub) dataTHS = ATS.shortenData(dataTHSroot,startSub,endSub) if GoT == 0: dataGTHS = ATS.combineData(dataG,dataTHS,endSub) else: dataGTHS = ATS.combineData(dataTHS,dataG,endSub) ori = updateDiscriptiveDataForBothTalks(ori,dataGTHS,GoT,nameFile,i,j) j = j + Width return ori
def updateDiscriptiveDataForBothTalks(ori,data,GoT,nameFile,dyad,section): # start = 0 end = 0 i = 0 (x,y) = data.shape count = 0 while i < x: # let i == x-1 first because it will check this first and skip the second one if it reach the end # you i+1 because we are counting forward if i == x-1 or data[i+1][1] != data[start][1] : end = i line = [] line.append(int(nameFile)) line.append(dyad) line.append(section) line.append(count) line.append(data[start][0]) #"end-1" is because to prevent out of bound for "end = x" #+1 at the is because we want to take into acount of the last number. Ex, from 0 to 20, there is 21 # line.append(data[end][0]) line.append(end-start+1) line.append(data[start][1]) #get the actions of either gaze or talk of the line action gazeAction = data[start:end+1,[0,2]] if GoT == 0: line = updateDiscriptiveDataSound(line,gazeAction) else : line = ATS.updateDiscriptiveData(line,gazeAction,0) count = count + 1 start = end +1 ori.append(line) i = i + 1 # for i in ori: # print (i) return ori
import numpy as np
import AnalysisToolForSingleSet as AT
if __name__ == "__main__":
# should have more varriable REMEMBER: DON'T CHANGE THIS CODE FIRST, CHANGE THE CODE FOR CONVERTING DATA FROM SRT TO SVN FIRST IF YOU DECIDE TO COMBINING CATEGORIES
Start = 0
End = 9000
inpath = "Input/"
nameFiles = np.loadtxt("NameFile.txt", dtype='S')
print nameFiles
title = ["Dyad", "Cov", "Time", "Gaze", "Talking Hawking", "Talking Shake"]
ori1 = AT.getOverallResult(inpath, Start, End, nameFiles, False)
name = "Output/OverallData_EndTime_5min.csv"
AT.printResult(ori1, title, name)
ori1 = AT.getOverallResult(inpath, Start, End, nameFiles, True)
name = "Output/OverallData_EndTime_Actual.csv"
AT.printResult(ori1, title, name)
inpath = "Input/"
nameFiles = np.loadtxt("NameFile.txt", dtype='S')
title = [
"Dyad", "Cov", "SectionStartTime", "TalkID", "Start", "End",
"Duration", "TalkType"
]
TalkType = (1, 2, 3, 4, 5)
Person = "H"
PersonTile = "S_"
nameOutput1 = "Output/CombinedOneLayer/Talk/DiscriptiveData_5talks_H_5mins.csv"
title1 = getTitle(title, GazeType, SpeakType, PersonTile, Character)
ori1 = AMS.getDiscriptiveTalkWithWidth(inpath, Start, End, nameFiles,
False, 9000, TalkType, Person)
ATS.printResult(ori1, title1, nameOutput1)
nameOutput2 = "Output/CombinedOneLayer/Talk/DiscriptiveData_5talks_H_100sec.csv"
title2 = getTitle(title, GazeType, SpeakType, PersonTile, Character)
ori2 = AMS.getDiscriptiveTalkWithWidth(inpath, Start, End, nameFiles,
False, 3000, TalkType, Person)
ATS.printResult(ori2, title2, nameOutput2)
Person = "S"
PersonTile = "H_"
nameOutput3 = "Output/CombinedOneLayer/Talk/DiscriptiveData_S_5mins.csv"
title3 = getTitle(title, GazeType, SpeakType, PersonTile, Character)
ori3 = AMS.getDiscriptiveTalkWithWidth(inpath, Start, End, nameFiles,
False, 9000, TalkType, Person)
ATS.printResult(ori3, title3, nameOutput3)
GazeType = [
"Around_", "Monitor_", "Keyboard_", "Face_", "NoFace_", "CompOnly_"
]
SpeakType = [
"NoSpeak_", "Typing_", "Speaking_", "SSS_", "Hovering_", "Sound_",
"NoSound_"
]
Person = ["H_", "S_"]
Character = ["Mean", "Median", "PopularSD", "Frequency", "Max", "Min"]
inpath = "Input/"
nameFiles = np.loadtxt("NameFile.txt", dtype='S')
print nameFiles
nameOutput1 = "Output/TTestData_5mins.csv"
title1 = getTitle(GazeType, SpeakType, Person, Character)
ori1 = AST.getDiscriptiveData(inpath, Start, End, nameFiles, False)
checkIndex = 1
groupCompaired = (1, 2)
startIndex = 2
endIndex = len(ori1[0])
result1 = ST.TTestPaired(ori1, checkIndex, groupCompaired, startIndex,
endIndex).tolist()
data = []
data.append(title1)
data.append(result1)
with open(nameOutput1, "wb") as t:
writer = csv.writer(t)
"BodyOnly_"
]
SpeakType = [
"NoSpeak_", "Typing_", "Speaking_", "SSS_", "Hovering_", "Sound_",
"NoSound_"
]
Person = ["H_", "S_"]
Character = ["Mean", "Median", "PopularSD", "Frequency", "Max", "Min"]
inpath = "Input/"
nameFiles = np.loadtxt("NameFile.txt", dtype='S')
print nameFiles
title1 = ["Dyad", "Cov"]
nameOutput1 = "Output/DiscriptiveData_5mins.csv"
title1 = getTitle(title1, GazeType, SpeakType, Person, Character)
ori1 = AT.getDiscriptiveData(inpath, Start, End, nameFiles, False)
AT.printResult(ori1, title1, nameOutput1)
title2 = ["Dyad", "Cov", "StartFrame"]
nameOutput2 = "Output/DiscriptiveData_100secs.csv"
title2 = getTitle(title2, GazeType, SpeakType, Person, Character)
ori2 = AT.getDiscriptiveDataWithWidth(inpath, Start, End, nameFiles, False,
3000)
AT.printResult(ori2, title2, nameOutput2)
title3 = ["Dyad", "Cov", "StartFrame"]
nameOutput3 = "Output/DiscriptiveData_60secs.csv"
title3 = getTitle(title3, GazeType, SpeakType, Person, Character)
ori3 = AT.getDiscriptiveDataWithWidth(inpath, Start, End, nameFiles, False,
1800)
AT.printResult(ori3, title3, nameOutput3)
import AnalysisToolForSingleSet as AT
import CoderAgreement as CA
if __name__ == "__main__":
inpath = "InRe/Talking/Hawking/" #double slash because codec is difference
Start = 0
End = 9000
GoTSize = 5
ori = CA.getReliability(inpath, Start, End, GoTSize)
title = ["Dyad", "Cov", "Coder", "Time", "talking"]
name = "Output/Reliability_Data_Talking_Hawking_5min.csv"
AT.printResult(ori, title, name)
inpath = "InRe/Talking/Shake/" #double slash because codec is difference
ori = CA.getReliability(inpath, Start, End, GoTSize)
name = "Output/Reliability_Data_Talking_Shakespear_5min.csv"
AT.printResult(ori, title, name)
End = 9000
SpeakType = ["NoSound_", "Sound_"]
GazeType = [
"Around_", "Monitor_", "Keyboard_", "Face_", "NoFace_", "CompOnly_",
"BodyOnly_"
]
Character = ["Mean", "Median", "PopularSD", "Frequency", "Max", "Min"]
inpath = "Input/"
nameFiles = np.loadtxt("NameFile.txt", dtype='S')
GoT = 0
nameOutput1 = "Output/CombinedTwoLayer/DiscriptiveData_Gaze_Sound_5mins.csv"
title1 = getTitle(GoT, GazeType, SpeakType, Character)
ori1 = AMS.getDiscriptiveTwoTalkWithWidth(inpath, Start, End, nameFiles,
False, 9000, GoT)
ATS.printResult(ori1, title1, nameOutput1)
nameOutput2 = "Output/CombinedTwoLayer/DiscriptiveData_Gaze_Sound_100secs.csv"
title2 = getTitle(GoT, GazeType, SpeakType, Character)
ori2 = AMS.getDiscriptiveTwoTalkWithWidth(inpath, Start, End, nameFiles,
False, 3000, GoT)
ATS.printResult(ori2, title2, nameOutput2)
GoT = 1
nameOutput3 = "Output/CombinedTwoLayer/DiscriptiveData_Sound_Gaze_5mins.csv"
title3 = getTitle(GoT, GazeType, SpeakType, Character)
ori3 = AMS.getDiscriptiveTwoTalkWithWidth(inpath, Start, End, nameFiles,
False, 9000, GoT)
ATS.printResult(ori3, title3, nameOutput3)
nameOutput4 = "Output/CombinedTwoLayer/DiscriptiveData_Sound_Gaze_100secs.csv"
