def extractSingleDailyActDiet(subjectID):
file_location = 'subject_template_' + subjectID + '.xlsx'
workbook = xlrd.open_workbook(file_location)
sheet = workbook.sheet_by_index(3)
duration = dietActInfoRetrv.getDuration(subjectID)
for n in range(1, duration + 1):
f_act = open(
'activity/activityFromExcel/activity_' + subjectID + '_' + str(n) +
'.txt', 'w')
f_diet = open(
'diet/dietFromExcel/diet_' + subjectID + '_' + str(n) + '.txt',
'w')
f_act.close()
f_diet.close()
count = 0
for row in range(8, sheet.nrows):
if sheet.cell_value(row, 0):
count += 1
if sheet.cell_value(row, 3):
temp = str(sheet.cell_value(row, 3).encode('utf-8'))
f_act = open(
'activity/activityFromExcel/activity_' + subjectID + '_' +
str(count) + '.txt', 'a')
f_act.write(temp)
f_act.write('\n')
f_act.close()
if sheet.cell_value(row, 4):
temp = str(sheet.cell_value(row, 4))
f_diet = open(
'diet/dietFromExcel/diet_' + subjectID + '_' + str(count) +
'.txt', 'a')
f_diet.write(temp)
f_diet.write('\n')
f_diet.close()
def buildDailyItemFreqTXTFile():
'''
write daily diet and activity index of each subject into txt files
'''
print 'in buildDailyItemFreqTXTFile()'
for subjectID in available_list:
# print subjectID
duration = dietActInfoRetrv.getDuration(subjectID)
# print duration
for n in range(1,duration+1):
f_act = open('activity/activityItemFreq/activity_frequency_'+subjectID+'_'+str(n)+'.txt','w')
f_diet = open('diet/dietItemFreq/diet_frequency_'+subjectID+'_'+str(n)+'.txt','w')
index_act = buildItemIndex.build_daily_single_activity_index(subjectID,n)
index_diet = buildItemIndex.build_daily_single_diet_index(subjectID,n)
# print index_act
# print index_diet
for key in index_act:
f_act.write("%-25s%-10s"%(key,index_act[key]))
f_act.write('\n')
for key in index_diet:
f_diet.write("%-25s%-10s"%(key,index_diet[key]))
f_diet.write('\n')
f_act.close()
f_diet.close()
def buildDailySingleActTypeFreqFile():
for subjectID in available_list:
duration = dietActInfoRetrv.getDuration(subjectID)
for n in range(1,duration+1):
f_act = open('activity/activityTypeFreq/activityType_frequency_'+subjectID+'_'+str(n)+'.txt','w')
singleActType_dict = buildTypeIndex.build_daily_single_activity_index(subjectID,n)
for key in singleActType_dict:
f_act.write("%-25s%-10s"%(key,singleActType_dict[key]))
f_act.write('\n')
f_act.close()
def genDailyActTypeDataSet(): dataset = [] for subjectid in available_list: duration = dietActInfoRetrv.getDuration(subjectid) for i in range(duration): indexDict = buildTypeIndex.build_daily_single_activity_index(subjectid,i+1) temp = tuple(indexDict) dataset.append(temp) dataset = tuple(dataset) print len(dataset) return dataset
def genDailyDietTypeDataSet(): dataset = [] for subjectid in available_list: duration = dietActInfoRetrv.getDuration(subjectid) for i in range(duration): indexDict = buildTypeIndex.build_daily_single_diet_index(subjectid,i+1) if 'compositeP' in indexDict: del indexDict['compositeP'] temp = tuple(indexDict) dataset.append(temp) dataset = tuple(dataset) print len(dataset) return dataset
def preprocessDailyDiaryWithTime():
print 'in preprocessDailyDiaryWithTime()'
for subjectID in available_list:
print subjectID
duration = dietActInfoRetrv.getDuration(subjectID)
for n in range(1, duration + 1):
preprocessingWithTime(
'activity/activityFromExcel/activity_' + subjectID + '_' +
str(n) + '_with_time.txt',
'activity/activityProcessed/processed_activity_' + subjectID +
'_' + str(n) + '_with_time.txt')
preprocessingWithTime(
'diet/dietFromExcel/diet_' + subjectID + '_' + str(n) +
'_with_time.txt', 'diet/dietProcessed/processed_diet_' +
subjectID + '_' + str(n) + '_with_time.txt')
def genDailySingleActTypeTFArray(subjectID):
item_dict = genActTypeDict()
duration = dietActInfoRetrv.getDuration(subjectID)
n = len(item_dict)
dims = (duration, n)
array = np.zeros(dims)
for i in range(duration):
ItemIndex = buildTypeIndex.build_daily_single_activity_index(
subjectID, i + 1)
for key in item_dict:
if item_dict[key] in ItemIndex:
# print item_dict[key]
array[i, key] = ItemIndex[item_dict[key]]
return array
def getDietTypeTFArray4DC():
type_dict = dataGen4DietAct.genDietTypeDict()
x = len(available_list)
n = len(type_dict)
array = np.zeros((x,n))
i = 0
for subjectID in available_list:
duration = dietActInfoRetrv.getDuration(subjectID)
for n in range(1,duration+1):
dictWithTime = buildTypeIndex.build_daily_single_diet_index_with_time4DC(subjectID,n)
for time in dictWithTime:
for key in type_dict:
if type_dict[key] in dictWithTime[time]:
array[i,key] += dictWithTime[time][type_dict[key]]
i += 1
return array
def genDailyActDietTypeDataSet():
dataset = []
for subjectid in available_list:
duration = dietActInfoRetrv.getDuration(subjectid)
for i in range(duration):
indexDict1 = buildTypeIndex.build_daily_single_activity_index(subjectid,i+1)
if 'others' in indexDict1: del indexDict1['others']
temp1 = tuple(indexDict1)
indexDict2 = buildTypeIndex.build_daily_single_diet_index(subjectid,i+1)
if 'compositeP' in indexDict2: del indexDict2['compositeP']
temp2 = tuple(indexDict2)
temp = temp1+temp2
dataset.append(temp)
dataset = tuple(dataset)
print len(dataset)
return dataset
def singleSubjectDailyArray(domain, subjectID):
'''
build daily item TFIDF normalization array
'''
if domain == 'ActItem':
item_dict = dataGen4DietAct.genActItemDict()
elif domain == 'DietItem':
item_dict = dataGen4DietAct.genDietItemDict()
duration = dietActInfoRetrv.getDuration(subjectID)
x = duration
n = len(item_dict)
dims = (x, n)
array = np.zeros(dims)
if domain == 'ActItem':
for i in range(duration):
ItemIndex = buildItemIndex.build_daily_single_activity_index(
subjectID, i + 1)
for key in item_dict:
if "'" + item_dict[key] + "'" in ItemIndex:
array[i, key] = ItemIndex["'" + item_dict[key] + "'"]
if domain == 'DietItem':
for i in range(duration):
ItemIndex = buildItemIndex.build_daily_single_diet_index(
subjectID, i + 1)
for key in item_dict:
if "'" + item_dict[key] + "'" in ItemIndex:
array[i, key] = ItemIndex["'" + item_dict[key] + "'"]
transformer = TfidfTransformer(norm=None)
tfidf = transformer.fit_transform(array)
aa = tfidf.toarray()
tfidfNorm = utilise.normArray(aa)
print tfidfNorm.shape
return tfidfNorm
def singleSubjectDailyArray(domain,subjectID):
'''
build daily item TF array
'''
if domain == 'DietType':
item_dict = dataGen4DietAct.genDietTypeDict()
elif domain == 'ActType':
item_dict = dataGen4DietAct.genActTypeDict()
# print item_dict
duration = dietActInfoRetrv.getDuration(subjectID)
x = duration
n = len(item_dict)
dims = (x,n)
array = np.zeros(dims)
if domain == 'ActItem':
for i in range(duration):
ItemIndex = buildItemIndex.build_daily_single_activity_index(subjectID,i+1)
for key in item_dict:
if item_dict[key] in ItemIndex:
array[i,key] = ItemIndex[item_dict[key]]
else:
array[i,key] = 0.0
if domain == 'DietItem':
for i in range(duration):
ItemIndex = buildItemIndex.build_daily_single_diet_index(subjectID,i+1)
# print ItemIndex
for key in item_dict:
if item_dict[key] in ItemIndex:
array[i,key] = ItemIndex[item_dict[key]]
else:
array[i,key] = 0.0
if domain == 'DietType':
for i in range(duration):
ItemIndex = buildTypeIndex.build_daily_single_diet_index(subjectID,i+1)
# print ItemIndex
for key in item_dict:
if item_dict[key] in ItemIndex:
array[i,key] = ItemIndex[item_dict[key]]
else:
array[i,key] = 0.0
if domain == 'ActType':
for i in range(duration):
ItemIndex = buildTypeIndex.build_daily_single_activity_index(subjectID,i+1)
for key in item_dict:
if item_dict[key] in ItemIndex:
array[i,key] = ItemIndex[item_dict[key]]
else:
array[i,key] = 0.0
'''
change the TF array to TFIDF array. But the DF here is not equal to the one we use for mean Vector
'''
# transformer = TfidfTransformer(norm=None)
# tfidf = transformer.fit_transform(array)
# aa = tfidf.toarray()
# tfidfNorm = utilise.normArray(aa)
# result = utilise.normArray(array)
# print array
return array