def compute_outputs(self):
for i in range(len(self.nodes)):
output = self.totals[i]
self.outputs[i] = output
outputs = sort(self.outputs)
size = int(len(outputs) * self.psize)
for i in range(len(outputs)):
n = outputs[i]
if i < size: self.outputs[n] = 1
else: self.outputs[n] = 0
def home():
data = request.get_json()
#logging.info("Request received: {}".format(data))
command = data['command']
res = {}
if(command=='ping'):
res = functions.ping_pong()
elif(command=='sample'):
res = functions.sample(data['size'], data['begin'], data['end'])
elif(command=='sort'):
res = functions.sort(data)
logging.info("Response: {}".format(res))
print(jsonify(res))
return jsonify(res)
parents[i] = range(n_cities)
np.random.shuffle(parents[i])
best = [] #this boy picks the best of each simulation
strd = 0 #this guy is for standard error measurements
means = []
#Action
for j in range(n_sims):
for gen in range(generations):
offspring = partially_mapped(parents, distances, n_cities, N_pop)
new_population = np.append([offspring], [parents], axis=1)
path_lengths, population = sort(new_population, distances)
removables = len(path_lengths) - N_pop
for i in range(removables):
path_lengths = np.delete(path_lengths, len(path_lengths) - 1)
parents = []
for i in range(N_pop):
parents.append(population[i])
for populant in range(N_pop):
parents[populant], path_lengths[populant] = hillclimb2(
parents[populant], distances)
means.append(np.mean(path_lengths))
best.append(path_lengths[0]) #collecting shit for standard deviation
def run_functions(curr_spreadsheet, query_string, query_parameters):
print("Query string: " + query_string)
# ADD <ENTRY_1> <ENTRY_2>
# <ENTRY_1>: row 4
commands = query_string.split(".")
sheet = curr_spreadsheet.sheet
for i in range(len(commands)):
args_and_comm = commands[i].split(" ")
command_name = args_and_comm[0]
args = []
for k in range(1, len(args_and_comm)):
result = args.append(query_parameters[args_and_comm[k]])
result = None
if command_name == "ADD":
result = functions.add(sheet, args[0], args[1])
elif command_name == "INSERTAFT":
if args[0].rowOrCol == 'col':
result = functions.insert_entry(sheet, chr(ord(args[0].value) + 1), args[0].rowOrCol)
else:
result = functions.insert_entry(sheet, int(args[0].value) + 1, args[0].rowOrCol)
elif command_name == "INSERTBEF":
if args[0].rowOrCol == 'col':
result = functions.insert_entry(sheet, chr(ord(args[0].value)), args[0].rowOrCol)
else:
result = functions.insert_entry(sheet, int(args[0].value), args[0].rowOrCol)
elif command_name == "SET":
if(isinstance(args[0], float)):
result = functions.cell_update(sheet, args[1].cell_str, args[0])
else:
result = functions.cell_update(sheet, args[1].cell_str, args[0].value)
elif command_name == "AVG":
result = functions.average_entry(sheet, args[0].value, args[0].rowOrCol)
elif command_name == "BOLD":
if(isinstance(args[0], Cell)):
result = functions.format_bold(sheet, str(args[0].cell_str) + ":" + str(args[0].cell_str))
else:
result = functions.format_bold_entry(sheet, args[0])
elif command_name == "SET_BG":
if(len(args) == 2):
if(isinstance(args[1], Cell)):
result = functions.set_background(sheet, args[1].cell_str + ":" + args[1].cell_str, args[0].color_str)
else:
result = functions.set_background_entry(sheet, args[1], args[0].color_str)
else:
result = functions.set_background(sheet, str(args[1].cell_str) + ":" + str(args[2].cell_str), args[0].color_str)
elif command_name == "MULTIPLY":
if(isinstance(args[1], Entry)):
result = functions.multiply_entry(sheet, int(args[0].value), args[1].value, args[1].rowOrCol == "row")
else:
result = functions.multiply_cell(sheet, int(args[0].value), args[1].cell_str)
elif command_name == "SIN":
if(len(args) == 1):
if (isinstance(args[0], Cell)):
result = functions.sin_cell(sheet, args[0].cell_str)
else:
# Don't know if this works
result = functions.sin_entry(sheet, ord(args[0].value) - ord("A") + 1, args[0].rowOrCol == "col")
else:
result = functions.sin_range(sheet, str(args[0].cell_str) + ":" + str(args[1].cell_str))
elif command_name == "COS":
if(len(args) == 1):
if (isinstance(args[0], Cell)):
result = functions.cos_cell(sheet, args[0].cell_str)
else:
result = functions.cos_entry(sheet, args[0].value, args[0].rowOrCol == "col")
else:
result = functions.cos_range(sheet, str(args[0].cell_str) + ":" + str(args[1].cell_str))
elif command_name == "SORT":
# result = functions.sort(sheet, args[0])
if(args[0].rowOrCol == 'col'):
result = functions.sort(sheet, (ord(args[0].value) - ord("A") + 1, 'asc'))
# elif command_name == "FILTER_EVEN":
# result = functions.filter_even(sheet, args[0])
# elif command_name == "FILTER_ODD":
# result = functions.filter_odd(sheet, args[0])
elif command_name == "FILTER_PRIME":
result = functions.filter_by_prime(sheet)
# elif command_name == "MAX_VAL":
# result = functions.max_val(sheet, args[0])
# elif command_name == "NORMALIZE":
# result = functions.normalize(sheet, args[0], args[1])
query_parameters["<RES_" + str(i + 1) + ">"] = result
buttons.grid_columnconfigure(0, weight=1)
buttons.grid_rowconfigure(0, weight=1)
# Treeview
container = ttk.Frame()
container.pack(fill='both', expand=True)
treeColumns = ("Item", "ID", "Price", "Available", "Checked out",
"Description")
tree = ttk.Treeview(columns=treeColumns, show="headings")
for column in treeColumns:
tree.heading(column,
text=column,
command=lambda c=column: functions.sort(tree, c, 0))
for i in treeColumns:
tree.column(i, width=130, minwidth=30)
vsb = ttk.Scrollbar(orient="vertical",
command=tree.yview) # Vertical scroll bar
hsb = ttk.Scrollbar(orient="horizontal",
command=tree.xview) # Horizontal scroll bar
tree.configure(yscrollcommand=vsb.set, xscrollcommand=hsb.set)
vsb.grid(column=1, row=0, sticky='ns', in_=container)
hsb.grid(column=0, row=1, sticky='ew', in_=container)
tree.grid(column=0, row=0, sticky='nsew', in_=container)
from math import sqrt
N = 1000
chars = ['avr', 'med', 'zr', 'zq', 'ztr']
arr = [10, 100, 1000]
print("Лабораторная работа №2; Выполнила: Фомина Дарья\n")
norm = dst.Normal(1, 0)
print("/************************************************/\n")
print("Normal distribution\n")
for n in arr:
print(f'n={n}')
res = {}
for i in range(5):
res[i] = [0, 0]
for i in range(N):
s = f.sort(norm, n)
z = f.avr(s)
res[0][0] += z / N
res[0][1] += z * z / N
z = f.med(s)
res[1][0] += z / N
res[1][1] += z * z / N
z = f.zr(s)
res[2][0] += z / N
res[2][1] += z * z / N
z = f.zq(s)
res[3][0] += z / N
res[3][1] += z * z / N
z = f.ztr(s)
res[4][0] += z / N
res[4][1] += z * z / N
userInput = input("Choose a index: ")
index = int(userInput)
# Get key by value
if index in headers.values(
) and index not in nonNumericIndexes and index is not 0:
index = int(userInput)
for key, value in headers.items():
if value == index:
header = key
# Sort and show all data, functions in external file 'functions.py'
dataset = copy.copy(data[index])
print(dataset)
sortedData = functions.sort(data[index])
print(header)
print(f"Average: \n{functions.average(dataset)}")
_modus = functions.modus(dataset)
print(f"Modus: \n{_modus[0]}: {_modus[1]} time(s)")
print(f"Median: \n{functions.median(sortedData)}")
print(f"Standard deviaton: \n{functions.standardDeviation(dataset)}")
# Show histogram with confindence interval
fig, ax = plt.subplots()
binwidth = (max(dataset) - min(dataset)) / 40
plt.hist(dataset,
bins=np.arange(min(dataset),
max(dataset) + binwidth, binwidth))
def launch_functions_menu():
print("Functions Section - Enter the exercise number you want to run")
selection = 0
while selection != 13:
print("exercise #1")
print("exercise #2")
print("exercise #3")
print("exercise #4")
print("exercise #5")
print("exercise #6")
print("exercise #7")
print("exercise #8")
print("exercise #9")
print("exercise #10")
print("exercise #11")
print("exercise #12")
print("exit #13")
selection = int(input("Insert Selection = "))
if selection == 1:
number = int(input("Enter a number to see if it's prime "))
print(functions.is_prime(number))
elif selection == 2:
number = int(input("Get all prime numbers smaller than "))
functions.prime_numbers_smaller_than(number)
elif selection == 3:
number = int(input("Fibonacci sequence with elements count = "))
functions.fibonacci_sequence(number)
elif selection == 4:
print("my_map function")
functions.exercise_4()
elif selection == 5:
print("my_filter function")
functions.exercise_5()
elif selection == 6:
print("my_reduce function")
functions.exercise_6()
elif selection == 7:
number = int(input("Special Sum of numbers smaller than "))
print(functions.sum_of_special_numbers_smaller_than(number))
elif selection == 8:
print("sort [4, 2, 3, 1] {0}".format(functions.sort([4, 2, 3, 1])))
elif selection == 9:
print("Check if 1 in list [1, 4, 5] ")
print(functions.element_in_list(1, [1, 4, 5]))
elif selection == 10:
print("call stateful function")
print(functions.stateful_func())
print(functions.stateful_func())
print(functions.stateful_func())
print("----- new state -----")
new_state = functions.State()
print(functions.stateful_func(new_state))
print(functions.stateful_func(new_state))
print("----- old state -----")
print(functions.stateful_func())
elif selection == 11:
print("measure the duration of a function")
print(functions.my_func())
elif selection == 12:
print("Total execution time for multiple functions")
print("----- Total Execution Time -------")
print(functions.my_func())
print("----- Total Execution Time -------")
print(functions.my_func2())
elif selection == 13:
print("Exit")
def action_statistics(action_path, outputpath):
cm_data_raw = pd.read_table(action_path, sep=',', encoding='utf-8')
#对user action data 进行统计处理。
names = ['userid', 'actionType', 'actionTime']
newCust = pd.DataFrame(columns=[
"userid", 'totalstep', 'step1N', 'step2N', 'step3N', 'step4N',
'step5N', 'step6N', 'step7N', 'step8N', 'step9N', 'step1P', 'step2P',
'step3P', 'step4P', 'step5P', 'step6P', 'step7P', 'step8P', 'step9P',
'viewProductN', 'last1time', 'last2time', 'last3time', 'last4time',
'last5time', 'last6time', 'last7time', 'last8time', 'last9time',
'buy/viewProductN', 'buy/n5', '(2-4)/(1)', 'tail1', 'tail2', 'tail3',
'tail4', 'head1', 'maxtime', 'mintime', 'averagetime', 'vartime',
'mediantime', 'tailtime1', 'tailtime2', 'tailtime3', 'tailtime4',
'tail3ave', 'tail3var'
])
idlist = []
for i in range(0, len(cm_data_raw)):
record = cm_data_raw.iloc[i]
uid = record['userid']
if uid not in idlist:
idlist.append(uid)
udata = cm_data_raw[cm_data_raw['userid'] == uid]
udata = sort(udata, ["actionType", "actionTime"], ascending=False)
#总步数为
totalStep = 0
#记录每一步击中数量
n1 = 0
n2 = 0
n3 = 0
n4 = 0
n5 = 0
n6 = 0
n7 = 0
n8 = 0
n9 = 0
#记录每一步击中数量比例
p1 = 0
p2 = 0
p3 = 0
p4 = 0
p5 = 0
p6 = 0
p7 = 0
p8 = 0
p9 = 0
#第一次击中步骤时间标签。用于识别最近一次到达此步骤。
f1 = 0
f2 = 0
f3 = 0
f4 = 0
f5 = 0
f6 = 0
f7 = 0
f8 = 0
f9 = 0
#第一次击中某一步骤的时间
t1 = 0
t2 = 0
t3 = 0
t4 = 0
t5 = 0
t6 = 0
t7 = 0
t8 = 0
t9 = 0
#倒数几步为何总总type
tail1 = 0
tail2 = 0
tail3 = 0
tail4 = 0
#第一步为何type
head1 = 0
#时间间隔初始化
timeslot = []
times1 = 0
ptime = 0
ctime = 0
#时间间隔相关统计
maxtime = 0
mintime = 0
averagetime = 0
vartime = 0
mediantime = 0
#倒数实践间隔
tail3a = []
tailtime1 = 0
tailtime2 = 0
tailtime3 = 0
tailtime4 = 0
tail3ave = 0
tail3var = 0
#对一个用户行为数据信息进行统计
for j in range(0, len(udata)):
oudata = udata.iloc[j]
#倒数的type
if j == 0:
tail1 = oudata['actionType']
tailtime1 = oudata['actionTime']
if j == 1:
tail2 = oudata['actionType']
tailtime2 = oudata['actionTime']
if j == 2:
tail3 = oudata['actionType']
tailtime3 = oudata['actionTime']
if j == 3:
tail4 = oudata['actionType']
tailtime4 = oudata['actionTime']
#每个行为每个用户使用过几次
if oudata['actionType'] == 1:
n1 = n1 + 1
if f1 == 0:
f1 = 2
t1 = oudata['actionTime']
if oudata['actionType'] == 2:
n2 = n2 + 1
if f2 == 0:
f2 = 2
t2 = oudata['actionTime']
if oudata['actionType'] == 3:
n3 = n3 + 1
if f3 == 0:
f3 = 2
t3 = oudata['actionTime']
if oudata['actionType'] == 4:
n4 = n4 + 1
if f4 == 0:
f4 = 2
t4 = oudata['actionTime']
if oudata['actionType'] == 5:
n5 = n5 + 1
if f5 == 0:
f5 = 2
t5 = oudata['actionTime']
if oudata['actionType'] == 9:
n9 = n9 + 1
if f9 == 0:
f9 = 2
t9 = oudata['actionTime']
if oudata['actionType'] == 8:
n8 = n8 + 1
if f8 == 0:
f8 = 2
t8 = oudata['actionTime']
if oudata['actionType'] == 7:
n7 = n7 + 1
if f7 == 0:
f7 = 2
t7 = oudata['actionTime']
if oudata['actionType'] == 6:
n6 = n6 + 1
if f6 == 0:
f6 = 2
t6 = oudata['actionTime']
if oudata['actionType'] == 5:
n5 = n5 + 1
if f5 == 0:
f5 = 2
t5 = oudata['actionTime']
head1 = oudata['actionType']
#进行时间间隔计算
if j != 0:
ctime = oudata['actionTime']
times1 = ctime - ptime
ptime = ctime
timeslot.append(times1)
else:
ptime = oudata['actionTime']
if len(timeslot) != 0:
maxtime = np.max(timeslot)
mintime = np.min(timeslot)
averagetime = np.average(timeslot)
mediantime = np.median(timeslot)
vartime = np.var(timeslot)
tail3a = timeslot[-3:]
tail3ave = np.average(tail3a)
tail3var = np.var(tail3a)
if n2 + n3 + n4 != 0:
buy_viewProductN = n9 / (n2 + n3 + n4)
elif n9 == 0:
buy_viewProductN = 0
elif n9 != 0 and n2 + n3 + n4 == 0:
buy_viewProductN = 1
if n5 != 0:
buy_5 = n9 / n5
else:
buy_5 = 0
if n2 + n3 + n4 != 0:
p59_24 = (n5 + n6 + n7 + n8 + n9) / (n2 + n3 + n4)
else:
p59_24 = 1
if p59_24 > 1:
p59_24 = 1
if n1 != 0:
p24_1 = (n2 + n3 + n4) / n1
else:
p24_1 = 0
totalStep = n1 + n2 + n3 + n4 + n5 + n6 + n7 + n8 + n9
p1 = n1 / totalStep
p2 = n2 / totalStep
p3 = n3 / totalStep
p4 = n4 / totalStep
p5 = n5 / totalStep
p6 = n6 / totalStep
p7 = n7 / totalStep
p8 = n8 / totalStep
p9 = n9 / totalStep
finalud = {
"userid": uid,
'totalStep': totalStep,
'step1N': n1,
'step2N': n2,
'step3N': n3,
'step4N': n4,
'step5N': n5,
'step6N': n6,
'step7N': n7,
'step8N': n8,
'step9N': n9,
'step1P': p1,
'step2P': p2,
'step3P': p3,
'step4P': p4,
'step5P': p5,
'step6P': p6,
'step7P': p7,
'step8P': p8,
'step9P': p9,
'viewProductN': n2 + n3 + n4,
'last1time': t1,
'last2time': t2,
'last3time': t3,
'last4time': t4,
'last5time': t5,
'last6time': t6,
'last7time': t7,
'last8time': t8,
'last9time': t9,
'buy/viewProductN': buy_viewProductN,
'buy/n5': buy_5,
'(2-4)/(1)': p24_1,
'tail1': tail1,
'tail2': tail2,
'tail3': tail3,
'tail4': tail4,
'head1': head1,
'maxtime': maxtime,
'mintime': mintime,
'averagetime': averagetime,
'vartime': vartime,
'mediantime': mediantime,
'tailtime1': tailtime1,
'tailtime2': tailtime2,
'tailtime3': tailtime3,
'tailtime4': tailtime4,
'tail3ave': tail3ave,
'tail3var': tail3var
}
newCust = newCust.append(finalud, ignore_index=True)
newCust.to_csv(outputpath)
return
def action_statistics(action_path,outputpath):
cm_data_raw=pd.read_table(action_path,sep=',',encoding='utf-8')
#对user action data 进行统计处理。
names=['userid','actionType','actionTime']
newCust=pd.DataFrame(columns=["userid",'step1N','step2N','step3N','step4N',
'step5N','step6N','step7N','step8N',
'step9N','viewProductN','last1time','last2time','last3time','last4time',
'last5time','last6time','last7time','last8time','last9time','buy/viewProductN',
'buy/n5','(9)/(1)','(2-4)/(1)'])
idlist=[]
for i in range(0,len(cm_data_raw)):
record=cm_data_raw.iloc[i]
uid=record['userid']
if uid not in idlist:
idlist.append(uid)
udata=cm_data_raw[cm_data_raw['userid']==uid]
udata=sort(udata,["actionType","actionTime"],ascending=False)
#记录每一步击中数量
n1=0
n2=0
n3=0
n4=0
n5=0
n6=0
n7=0
n8=0
n9=0
#第一次击中步骤时间标签。用于识别最近一次到达此步骤。
f1=0
f2=0
f3=0
f4=0
f5=0
f6=0
f7=0
f8=0
f9=0
#第一次击中某一步骤的时间
t1=0
t2=0
t3=0
t4=0
t5=0
t6=0
t7=0
t8=0
t9=0
#对一个用户行为数据信息进行统计
for j in range(0,len(udata)):
oudata=udata.iloc[j]
#每个行为每个用户使用过几次
if oudata['actionType']==1:
n1=n1+1
if f1==0:
f1=2
t1=oudata['actionTime']
if oudata['actionType']==2:
n2=n2+1
if f2==0:
f2=2
t2=oudata['actionTime']
if oudata['actionType']==3:
n3=n3+1
if f3==0:
f3=2
t3=oudata['actionTime']
if oudata['actionType']==4:
n4=n4+1
if f4==0:
f4=2
t4=oudata['actionTime']
if oudata['actionType']==5:
n5=n5+1
if f5==0:
f5=2
t5=oudata['actionTime']
if oudata['actionType']==9:
n9=n9+1
if f9==0:
f9=2
t9=oudata['actionTime']
if oudata['actionType']==8:
n8=n8+1
if f8==0:
f8=2
t8=oudata['actionTime']
if oudata['actionType']==7:
n7=n7+1
if f7==0:
f7=2
t7=oudata['actionTime']
if oudata['actionType']==6:
n6=n6+1
if f6==0:
f6=2
t6=oudata['actionTime']
if oudata['actionType']==5:
n5=n5+1
if f5==0:
f5=2
t5=oudata['actionTime']
#因为数据有缺失，如果5-9有缺失就用下一步数量补充。
if n8<n9:
n8=n9
t8=t9
if n7<n8:
n7=n8
t7=t8
if n6<n7:
n6=n7
t6=t7
if n5<n6:
n5=n6
t5=t6
if n2+n3+n4 !=0:
buy_viewProductN=n9/(n2+n3+n4)
elif n9==0:
buy_viewProductN=0
elif n9!=0 and n2+n3+n4 ==0:
buy_viewProductN=1
if n5 !=0:
buy_5=n9/n5
else:
buy_5=0
if n2+n3+n4 !=0:
p59_24=(n5+n6+n7+n8+n9)/(n2+n3+n4)
else:
p59_24=1
if p59_24>1:
p59_24=1
if n1!=0:
p9_1=n9/n1
else:
p9_1=0
if n1!=0:
p24_1=(n2+n3+n4)/n1
else:
p24_1=0
finalud={"userid":uid,'step1N':n1,'step2N':n2,'step3N':n3,'step4N':n4,
'step5N':n5,'step6N':n6,'step7N':n7,'step8N':n8,
'step9N':n9,'viewProductN':n2+n3+n4,'last1time':t1,'last2time':t2,'last3time':t3,'last4time':t4,
'last5time':t5,'last6time':t6,'last7time':t7,'last8time':t8,'last9time':t9,'buy/viewProductN':buy_viewProductN,
'buy/n5':buy_5,'(9)/(1)':p9_1,'(2-4)/(1)':p24_1}
newCust=newCust.append(finalud,ignore_index=True)
newCust.to_csv(outputpath)
return
def up_oh_statistics(up_path, oh_path, outputpath):
up_op_data = datatonumber(up_path, oh_path)
up_data = up_op_data[1]
oh_data = up_op_data[0]
#对orderhistory data 进行统计处理。
names = [
'userid', 'orderid', 'orderTime', 'orderType', 'contientN', 'cityN',
'countryN'
]
prev_cus_id = '0'
newCust = pd.DataFrame(columns=[
"userid", 'nearestOrderTime', 'nearestCity', 'nearestCountry',
'nearestContinent', 'totalOrder', 'numberOftype0ne', 'typeOnePer',
'numberofCity', 'numberofCountry', 'NumberofU', 'NumberofNA',
'NumberofA', 'NumberofO', 'NumberofAF', 'NumberofSA'
])
idlist = []
for i in range(0, len(oh_data)):
record = oh_data.iloc[i]
uid = record['userid']
if uid not in idlist:
idlist.append(uid)
udata = oh_data[oh_data['userid'] == uid]
udata = sort(udata, ["orderTime", "orderType"], ascending=False)
prv_time = []
totalOrder = 0
numberOftype0ne = 0
numberofCity = 0
numberofCountry = 0
NumberofU = 0
NumberofA = 0
NumberofNA = 0
NumberofSA = 0
NumberofAF = 0
NumberofO = 0
cityList = []
countryList = []
#对一个用户历史信息进行统计
for j in range(0, len(udata)):
oudata = udata.iloc[j]
odtime = oudata['orderTime']
ot = oudata['orderType']
ct = oudata['cityN']
coun = oudata['countryN']
cn = oudata['continentN']
#最近一次去过的城市，国家，以及大洲,因为之前userdata 按照时间大小排序，所以第一个就是最近的订单
if j == 0:
nearestCity = ct
nearestCou = coun
nearestCn = cn
nearestOd = odtime
if odtime not in prv_time:
prv_time.append(odtime)
#总订单数，同一时间订单视为父子订单只计一个订单
totalOrder = totalOrder + 1
#精品游订单数量统计
if ot == 1:
numberOftype0ne = numberOftype0ne + 1
#去过的城市列表，用于计算去过几个城市
if ct not in cityList:
cityList.append(ct)
#去过的国家列表，用于计算去过几个国家
if coun not in countryList:
countryList.append(coun)
#去过各个大洲几次
if cn == 'U':
NumberofU = NumberofU + 1
if cn == 'NA':
NumberofNA = NumberofNA + 1
if cn == 'A':
NumberofA = NumberofA + 1
if cn == 'O':
Numberof0 = NumberofO + 1
if cn == 'AF':
NumberofAF = NumberofAF + 1
if cn == 'SA':
NumberofSA = NumberofSA + 1
finalud = {
'userid': uid,
'nearestOrderTime': nearestOd,
'nearestCity': nearestCity,
'nearestCountry': nearestCou,
'nearestContinent': nearestCn,
'totalOrder': totalOrder,
'numberOftype0ne': numberOftype0ne,
'typeOnePer': numberOftype0ne / totalOrder,
'numberofCity': len(cityList),
'numberofCountry': len(countryList),
'NumberofU': NumberofU,
'NumberofNA': NumberofNA,
'NumberofA': NumberofA,
'NumberofO': NumberofO,
'NumberofAF': NumberofAF,
'NumberofSA': NumberofSA
}
newCust = newCust.append(finalud, ignore_index=True)
#用order history 统计好的表left join user profile表得出一个新表并输出
up_oh_data = pd.merge(up_data, newCust, how='outer', on='userid')
data_1 = up_oh_data[[
'genderN', 'provinceN', 'nearestCity', 'nearestCountry',
'nearestContinent'
]]
data_2 = pd.get_dummies(data_1)
result_1 = pd.concat([up_oh_data, data_2], axis=1)
result_1.drop([
'genderN', 'provinceN', 'nearestCity', 'nearestCountry',
'nearestContinent'
],
axis=1,
inplace=True)
result_1.to_csv(outputpath)
return
def action_statistics(action_path, outputpath):
cm_data_raw = pd.read_table(action_path, sep=',', encoding='utf-8')
#对user action data 进行统计处理。
names = ['userid', 'actionType', 'actionTime']
newCust = pd.DataFrame(columns=[
"userid", 'totalstep', 'step1N', 'step2N', 'step3N', 'step4N',
'step5N', 'step6N', 'step7N', 'step8N', 'step9N', 'step1P', 'step2P',
'step3P', 'step4P', 'step5P', 'step6P', 'step7P', 'step8P', 'step9P',
'viewProductN', 'last1time', 'last2time', 'last3time', 'last4time',
'last5time', 'last6time', 'last7time', 'last8time', 'last9time',
'buy/viewProductN', 'buy/n5', '(2-4)/(1)', 'tail1', 'tail2', 'tail3',
'tail4', 'head1', 'maxtime', 'mintime', 'averagetime', 'vartime',
'mediantime', 'tailtime1', 'tailtime2', 'tailtime3', 'tailtime4',
'tail3ave', 'tail3var', 'near1', 'near2', 'near3', 'near4', 'near5',
'near6', 'near7', 'nsar8', 'near9', 't1a', 't1v', 't1mini', 't1max',
't2a', 't2v', 't2mini', 't2max', 't3a', 't3v', 't3mini', 't3max',
't4a', 't4v', 't4mini', 't4max', 't5a', 't5v', 't5mini', 't5max',
't6a', 't6v', 't6mini', 't6max', 't7a', 't7v', 't7mini', 't7max',
't8a', 't8v', 't8mini', 't8max', 't9a', 't9v', 't9mini', 't9max',
't9av'
])
idlist = []
for i in range(0, len(cm_data_raw)):
record = cm_data_raw.iloc[i]
uid = record['userid']
if uid not in idlist:
idlist.append(uid)
udata = cm_data_raw[cm_data_raw['userid'] == uid]
udata = sort(udata, ["actionTime", "actionType"], ascending=False)
#总步数为
totalStep = 0
#记录每一步击中数量
n1 = 0
n2 = 0
n3 = 0
n4 = 0
n5 = 0
n6 = 0
n7 = 0
n8 = 0
n9 = 0
#记录每一步击中数量比例
p1 = 0
p2 = 0
p3 = 0
p4 = 0
p5 = 0
p6 = 0
p7 = 0
p8 = 0
p9 = 0
#第一次击中步骤时间标签。用于识别最近一次到达此步骤。
f1 = 0
f2 = 0
f3 = 0
f4 = 0
f5 = 0
f6 = 0
f7 = 0
f8 = 0
f9 = 0
#第一次击中某一步骤的时间
t1 = 0
t2 = 0
t3 = 0
t4 = 0
t5 = 0
t6 = 0
t7 = 0
t8 = 0
t9 = 0
#倒数几步为何总总type
tail1 = 0
tail2 = 0
tail3 = 0
tail4 = 0
#第一步为何type
head1 = 0
#时间间隔初始化
timeslot = []
times1 = 0
ptime = 0
ctime = 0
#时间间隔相关统计
maxtime = 0
mintime = 0
averagetime = 0
vartime = 0
mediantime = 0
#倒数实践间隔
tail3a = []
tailtime1 = 0
tailtime2 = 0
tailtime3 = 0
tailtime4 = 0
tail3ave = 0
tail3var = 0
#距离每个行为最近的行为和时间
near1 = 0
near2 = 0
near3 = 0
near4 = 0
near5 = 0
near6 = 0
near7 = 0
near8 = 0
near9 = 0
#距离各个节点最近时间间隔
ne1 = []
ne2 = []
ne3 = []
ne4 = []
ne5 = []
ne6 = []
ne7 = []
ne8 = []
ne9 = []
#对一个用户行为数据信息进行统计
for j in range(0, len(udata)):
oudata = udata.iloc[j]
#倒数的type
if j == 0:
tail1 = oudata['actionType']
tailtime1 = oudata['actionTime']
if j == 1:
tail2 = oudata['actionType']
tailtime2 = oudata['actionTime']
if j == 2:
tail3 = oudata['actionType']
tailtime3 = oudata['actionTime']
if j == 3:
tail4 = oudata['actionType']
tailtime4 = oudata['actionTime']
#每个行为每个用户使用过几次
if oudata['actionType'] == 1:
n1 = n1 + 1
if f1 == 0:
f1 = 2
t1 = oudata['actionTime']
near1 = j
if j > 0:
for i in range(1, j):
ne1.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 2:
n2 = n2 + 1
if f2 == 0:
f2 = 2
t2 = oudata['actionTime']
near2 = j
if j > 0:
for i in range(1, j):
ne2.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 3:
n3 = n3 + 1
if f3 == 0:
f3 = 2
t3 = oudata['actionTime']
near3 = j
if j > 0:
for i in range(1, j):
ne3.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 4:
n4 = n4 + 1
if f4 == 0:
f4 = 2
t4 = oudata['actionTime']
near4 = j
if j > 0:
for i in range(1, j):
ne4.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 5:
n5 = n5 + 1
if f5 == 0:
f5 = 2
t5 = oudata['actionTime']
near5 = j
if j > 0:
for i in range(1, j):
ne5.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 9:
n9 = n9 + 1
if f9 == 0:
f9 = 2
t9 = oudata['actionTime']
near9 = j
if j > 0:
for i in range(1, j):
ne9.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 8:
n8 = n8 + 1
if f8 == 0:
f8 = 2
t8 = oudata['actionTime']
near8 = j
if j > 0:
for i in range(1, j):
ne8.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 7:
n7 = n7 + 1
if f7 == 0:
f7 = 2
t7 = oudata['actionTime']
near7 = j
if j > 0:
for i in range(1, j):
ne7.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 6:
n6 = n6 + 1
if f6 == 0:
f6 = 2
t6 = oudata['actionTime']
near6 = j
if j > 0:
for i in range(1, j):
ne6.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
if oudata['actionType'] == 5:
n5 = n5 + 1
if f5 == 0:
f5 = 2
t5 = oudata['actionTime']
near5 = j
if j > 0:
for i in range(1, j):
ne7.append(udata.iloc[i]['actionTime'] -
udata.iloc[i - 1]['actionTime'])
head1 = oudata['actionType']
#进行时间间隔计算
if j != 0:
ctime = oudata['actionTime']
times1 = ctime - ptime
ptime = ctime
timeslot.append(times1)
else:
ptime = oudata['actionTime']
if len(timeslot) != 0:
maxtime = np.max(timeslot)
mintime = np.min(timeslot)
averagetime = np.average(timeslot)
mediantime = np.median(timeslot)
vartime = np.var(timeslot)
tail3a = timeslot[-3:]
tail3ave = np.average(tail3a)
tail3var = np.var(tail3a)
if n2 + n3 + n4 != 0:
buy_viewProductN = n9 / (n2 + n3 + n4)
elif n9 == 0:
buy_viewProductN = 0
elif n9 != 0 and n2 + n3 + n4 == 0:
buy_viewProductN = 1
if n5 != 0:
buy_5 = n9 / n5
else:
buy_5 = 0
if n2 + n3 + n4 != 0:
p59_24 = (n5 + n6 + n7 + n8 + n9) / (n2 + n3 + n4)
else:
p59_24 = 1
if p59_24 > 1:
p59_24 = 1
if n1 != 0:
p24_1 = (n2 + n3 + n4) / n1
else:
p24_1 = 0
totalStep = n1 + n2 + n3 + n4 + n5 + n6 + n7 + n8 + n9
p1 = n1 / totalStep
p2 = n2 / totalStep
p3 = n3 / totalStep
p4 = n4 / totalStep
p5 = n5 / totalStep
p6 = n6 / totalStep
p7 = n7 / totalStep
p8 = n8 / totalStep
p9 = n9 / totalStep
t8a = 0
t8v = 0
t8min = 0
t8max = 0
t7a = 0
t7v = 0
t7min = 0
t7max = 0
t6a = 0
t6v = 0
t6min = 0
t6max = 0
t5a = 0
t5v = 0
t5min = 0
t5max = 0
t4a = 0
t4v = 0
t4min = 0
t4max = 0
t3a = 0
t3v = 0
t3min = 0
t3max = 0
t2a = 0
t2v = 0
t2min = 0
t2max = 0
t1a = 0
t1v = 0
t1min = 0
t1max = 0
t9a = 0
t9v = 0
t9min = 0
t9max = 0
t9av = 0
if len(ne9) != 0:
t9a = np.average(ne9)
t9v = np.var(ne9)
t9min = np.min(ne9)
t9max = np.max(ne9)
t9av = t9a * t9v
if len(ne8) != 0:
t8a = np.average(ne8)
t8v = np.var(ne8)
t8min = np.min(ne8)
t8max = np.max(ne8)
if len(ne7) != 0:
t7a = np.average(ne7)
t7v = np.var(ne7)
t7min = np.min(ne7)
t7max = np.max(ne7)
if len(ne6) != 0:
t6a = np.average(ne6)
t6v = np.var(ne6)
t6min = np.min(ne6)
t6max = np.max(ne6)
if len(ne5) != 0:
t5a = np.average(ne5)
t5v = np.var(ne5)
t5min = np.min(ne5)
t5max = np.max(ne5)
if len(ne4) != 0:
t4a = np.average(ne4)
t4v = np.var(ne4)
t4min = np.min(ne4)
t4max = np.max(ne4)
if len(ne3) != 0:
t3a = np.average(ne3)
t3v = np.var(ne3)
t3min = np.min(ne3)
t3max = np.max(ne3)
if len(ne2) != 0:
t2a = np.average(ne2)
t2v = np.var(ne2)
t2min = np.min(ne2)
t2max = np.max(ne2)
if len(ne1) != 0:
t1a = np.average(ne1)
t1v = np.var(ne1)
t1min = np.min(ne1)
t1max = np.max(ne1)
finalud = {
"userid": uid,
'totalStep': totalStep,
'step1N': n1,
'step2N': n2,
'step3N': n3,
'step4N': n4,
'step5N': n5,
'step6N': n6,
'step7N': n7,
'step8N': n8,
'step9N': n9,
'step1P': p1,
'step2P': p2,
'step3P': p3,
'step4P': p4,
'step5P': p5,
'step6P': p6,
'step7P': p7,
'step8P': p8,
'step9P': p9,
'viewProductN': n2 + n3 + n4,
'last1time': t1,
'last2time': t2,
'last3time': t3,
'last4time': t4,
'last5time': t5,
'last6time': t6,
'last7time': t7,
'last8time': t8,
'last9time': t9,
'buy/viewProductN': buy_viewProductN,
'buy/n5': buy_5,
'(2-4)/(1)': p24_1,
'tail1': tail1,
'tail2': tail2,
'tail3': tail3,
'tail4': tail4,
'head1': head1,
'maxtime': maxtime,
'mintime': mintime,
'averagetime': averagetime,
'vartime': vartime,
'mediantime': mediantime,
'tailtime1': tailtime1,
'tailtime2': tailtime2,
'tailtime3': tailtime3,
'tailtime4': tailtime4,
'tail3ave': tail3ave,
'tail3var': tail3var,
't9a': t9a,
't9v': t9v,
't9mini': t9min,
't9max': t9max,
't9av': t9av,
't8a': t8a,
't8v': t8v,
't8mini': t8min,
't8max': t8max,
't7a': t7a,
't7v': t7v,
't7mini': t7min,
't7max': t7max,
't6a': t6a,
't6v': t6v,
't6mini': t6min,
't6max': t6max,
't5a': t5a,
't5v': t5v,
't5mini': t5min,
't5max': t5max,
't4a': t4a,
't4v': t4v,
't4mini': t4min,
't4max': t4max,
't3a': t3a,
't3v': t3v,
't3mini': t3min,
't3max': t3max,
't2a': t2a,
't2v': t2v,
't2mini': t2min,
't2max': t2max,
't1a': t1a,
't1v': t1v,
't1mini': t1min,
't1max': t1max
}
newCust = newCust.append(finalud, ignore_index=True)
newCust.to_csv(outputpath)
return
def rating_stantistics(inputpath, outputpath):
cm_data_raw = pd.read_table(inputpath, sep=',', encoding='utf-8')
#对user action data 进行统计处理。
newCust = pd.DataFrame(columns=[
"userid", 'totalrate', 'totalnumber', 'averate', 'numberof1',
'numberof2', 'numberof3', 'numberof367', 'numberof433', 'numberof4',
'numberof5', 'lowrate', 'highrate'
])
idlist = []
for i in range(0, len(cm_data_raw)):
record = cm_data_raw.iloc[i]
uid = record['userid']
if uid not in idlist:
idlist.append(uid)
udata = cm_data_raw[cm_data_raw['userid'] == uid]
udata = sort(udata, ["orderid"], ascending=False)
#记录总分数
totalrate = 0
totalnumber = 0
averate = 0
numberof5 = 0
numberof1 = 0
numberof2 = 0
numberof3 = 0
numberof367 = 0
numberof433 = 0
numberof4 = 0
#3分及其以下为low rate,求其数量
lowrate = 0
#4分及其以上为高分
highrate = 0
#对一个用户评分信息进行统计
for j in range(0, len(udata)):
oudata = udata.iloc[j]
totalrate = totalrate + oudata['rating']
totalnumber = totalnumber + 1
if j == 0:
nr = oudata['rating']
#统计用户评分
if oudata['rating'] == 1:
numberof1 = numberof1 + 1
lowrate = lowrate + 1
if oudata['rating'] == 2:
numberof2 = numberof2 + 1
lowrate = lowrate + 1
if oudata['rating'] == 3:
numberof3 = numberof3 + 1
lowrate = lowrate + 1
if oudata['rating'] == 3.67:
numberof367 = numberof367 + 1
if oudata['rating'] == 4.33:
numberof367 = numberof433 + 1
if oudata['rating'] == 4:
numberof4 = numberof4 + 1
if oudata['rating'] == 5:
numberof5 = numberof5 + 1
averate = totalrate / totalnumber
lowrate = numberof1 + numberof2 + numberof3
highrate = numberof4 + numberof5 + numberof433
finalud = {
"userid": uid,
'totalrate': totalrate,
'totalnumber': totalnumber,
'averate': averate,
'numberof1': numberof1,
'numberof2': numberof2,
'numberof3': numberof3,
'numberof367': numberof367,
'numberof433': numberof433,
'numberof4': numberof4,
'numberof5': numberof5,
'lowrate': lowrate,
'highrate': highrate
}
newCust = newCust.append(finalud, ignore_index=True)
newCust.to_csv(outputpath)
return
# Generates list of files in path
csv_files = os.listdir(path)
# Define variables
frequency = functions.namesList(csv_files) # Independent variable
rangeList1 = [] # Going to be sorted from first to last
rangeList2 = [] # Going to be sorted from first to last
meanList1 = [] # Going to be sorted from first to last
meanList2 = [] # Going to be sorted from first to last
#print(frequency)
#print(csv_files)
# Take the directory and append the ranges from their respective columns
for file in functions.sort(csv_files):
analog1Range = functions.getRange(
functions.convertToFloat(functions.columnList(path + file, 0)))
analog2Range = functions.getRange(
functions.convertToFloat(functions.columnList(path + file, 1)))
rangeList1.append(analog1Range)
rangeList2.append(analog2Range)
analog1Mean = functions.mean(
functions.convertToFloat(functions.columnList(path + file, 0)))
analog2Mean = functions.mean(
functions.convertToFloat(functions.columnList(path + file, 1)))
meanList1.append(analog1Mean)
meanList2.append(analog2Mean)
print("File: " + file)
#print(functions.sort(csv_files))
