I would like to obtain the subtotal (average, min, max...) of a group of data. I have achieve the goal using the code below. How can I use loop to simplify it? Many thanks!
Sub AddSubs()
Worksheets("Summary (3)").Activate
'http://msdn.microsoft.com/en-us/library/office/ff838166(v=office.15).aspx
Selection.Subtotal GroupBy:=14, Function:=xlAverage, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Worksheets("Summary (3)").Activate
Selection.Subtotal GroupBy:=14, Function:=xlStDev, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Worksheets("Summary (3)").Activate
Selection.Subtotal GroupBy:=14, Function:=xlMin, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Worksheets("Summary (3)").Activate
Selection.Subtotal GroupBy:=14, Function:=xlMax, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Worksheets("Summary (3)").Activate
Selection.Subtotal GroupBy:=14, Function:=xlCount, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
End Sub
Further to my comment. This is one way of simplifying your code
Sub AddSubs()
Worksheets("Summary (3)").Activate
Dim constList As Collection
Set constList = New Collection
constList.Add (xlAverage)
constList.Add (xlStDev)
constList.Add (xlMin)
constList.Add (xlMax)
constList.Add (xlCount)
Dim cnst
For Each cnst In constList
Selection.Subtotal GroupBy:=14, Function:=cnst, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Next
End Sub
or even simpler (per #simocos hint)
Sub Main()
Dim cnst
For Each cnst In Array(xlAverage, xlStDev, xlMin, xlMax, xlCount)
Selection.Subtotal GroupBy:=14, Function:=cnst, SummaryBelowData:=False, Replace:=False, PageBreaks:=True, TotalList:=Array(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
Next
End Sub
Related
I have this 2-D tensor:
tmp = torch.tensor([[ 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5,
5, 6, 6, 6, 7, 7, 7, 8, 8, 8, 9, 9, 9, 10, 10, 10, 11, 11,
11, 12, 12, 12, 13, 13, 13, 14, 14, 14, 15, 15, 15, 15, 16, 16, 16, 17,
17, 17, 18, 18, 18, 19, 19, 19, 20, 20, 20, 21, 21, 21, 22, 22, 22, 23,
23, 23, 24, 24, 24, 25, 25, 25, 26, 26, 26, 27, 27, 27, 28, 28, 28, 29,
29, 29, 30, 30, 30, 31, 31, 31, 31],
[ 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5,
5, 6, 6, 6, 7, 7, 7, 8, 8, 8, 9, 9, 9, 10, 10, 10, 11, 11,
11, 12, 12, 12, 13, 13, 13, 14, 14, 14, 15, 15, 15, 15, 0, 16, 16, 17,
17, 17, 18, 18, 18, 19, 19, 19, 20, 20, 20, 21, 21, 21, 22, 22, 22, 23,
23, 23, 24, 24, 24, 25, 25, 25, 26, 26, 26, 27, 27, 27, 28, 28, 28, 29,
29, 29, 30, 30, 30, 31, 31, 31, 31]])
So there is 0 in the 50th column of row 2. When I apply torch.unique along
dim=1, I get:
a,c = torch.unique(tmp,dim=1,return_counts=True)
a
tensor([[ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 16,
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31],
[ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 0, 16,
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]])
It can be seen that the second row of the output has two 0s and the first row has two 16s. Am I doing something wrong here or this is suspicious?
It is because you specified dim=1. PyTorch is thus checking for unique pairs (which it correctly does). Like (0, 0), (1, 1), (16, 0): these are the unique pairs that it generated. In general the pair (temp[0,i], temp[1,i]) is unique for all i.
If you want all the elements to be unique, just throw away the dim: torch.unique(tmp).
If you need to maintain the two list structure, the output cannot be stored as a single tensor because their sizes might not match. You can do something like output1 = torch.unique(tmp[0]) and output2 = torch.unique(tmp[1]).
I want to write a function that goes through all 3 of the lists down below and would print out the numbers that are present on all 3. Like how the number 23 is present on all 3 lists.
list_1 =[27, 20, 22, 21, 17, 12, 24, 23, 19, 14, 11, 26, 25, 13, 15, 21, 18, 28, 29, 10]
list_2 = [14, 25, 26, 21, 22, 17, 11, 23, 27, 18, 24, 28, 12, 29, 16, 19, 13, 10, 20, 15]
list_3 = [19, 21, 11, 24, 16, 17, 18, 22, 26, 10, 23, 29, 27, 13, 25, 14, 15, 20, 28, 12]
As #Heike said, you can use intersect1d
print(reduce(np.intersect1d, (list_1, list_2, list_3)))
Result:
[10 11 12 13 14 15 17 18 19 20 21 22 23 24 25 26 27 28 29]
Code:
import numpy as np
from functools import reduce
list_1 = [27, 20, 22, 21, 17, 12, 24, 23, 19, 14, 11, 26, 25, 13, 15, 21, 18, 28, 29, 10]
list_2 = [14, 25, 26, 21, 22, 17, 11, 23, 27, 18, 24, 28, 12, 29, 16, 19, 13, 10, 20, 15]
list_3 = [19, 21, 11, 24, 16, 17, 18, 22, 26, 10, 23, 29, 27, 13, 25, 14, 15, 20, 28, 12]
print(reduce(np.intersect1d, (list_1, list_2, list_3)))
I have a numpy array atoms.numbers which looks like:
array([27, 27, 27, 27, 27, 27, 57, 57, 57, 57, 57, 57, 57, 57, 27, 27, 8,
8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
8, 8, 8, 8, 8, 8, 27, 27, 27, 27, 27, 27, 57, 57, 57, 57, 57,
57, 57, 57, 27, 27, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8])
I can replace all of the same instance such as every '57' in the array using:
atoms.numbers[atoms.numbers==57]=38
which gives:
array([27, 27, 27, 27, 27, 27, 38, 38, 38, 38, 38, 38, 38, 38, 27, 27, 8,
8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
8, 8, 8, 8, 8, 8, 27, 27, 27, 27, 27, 27, 38, 38, 38, 38, 38,
38, 38, 38, 27, 27, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8])
I would like to be able to replace every nth instance in the array. I have tried:
n=5
atoms.numbers[atoms.numbers==57][::n]=38
Which does not work.
Use np.where to find the indexes of the items of interest. Find every n'th index. Update the items:
locations = np.where(numbers == 57)[0]
numbers[locations[::n]] = 38
I want to update the value of a key in dictionary. This is a snippet of a list that contains over 300 dictionaries
chats = [
{'hour': 10, 'operator': 'john_doe', 'duration': [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'john_doe', 'duration': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'john_doe', 'duration': [18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 11, 'operator': 'john_doe', 'duration': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'joseph_doe', 'duration': [5, 6, 7, 8, 9], 'date': '2019-09-09'}
]
script: I am getting an error on that script. I am looping to know if this dict is already in so that I can update the duration.
chat_list = list()
for chat in chats:
hour = chat.get('hour')
operator = chat.get("operator")
if len(chat_list) == 0:
chat_list.append(chat)
else:
found = False
for i in chat_list:
hour2 = chat.get('hour')
operator2 = chat.get("operator")
if (hour2 == hour) and (operator == operator2):
found = True
#concat both dictionary
i['duration'] = i.get('duration') + chat.get("duration")
if found == True:
found = False
else:
chat_list.append(chat)
My expected output is
chat_list = [
{'hour': 10, 'operator': 'john_doe', 'duration': [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 11, 'operator': 'john_doe', 'duration': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'joseph_doe', 'duration': [5, 6, 7, 8, 9], 'date': '2019-09-09'}
]
or
df = pd.DataFrame(chat_list)
df['duration'] = df['duration'].apply(lambda x: list(set(x)))
To be honest, I didn't tested your algorithm. Instead I took it as a small challenge and I wrote the following algorithm which doesn't need to copy chats in to a new list.
It finds the first occurrence of "similar" chat and concat the duration arrays. Then it deletes the "duplicated" chat. Further explanation in the code itself:
chats = [
{'hour': 10, 'operator': 'john_doe', 'duration': [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'john_doe', 'duration': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'john_doe', 'duration': [18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 11, 'operator': 'john_doe', 'duration': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28], 'date': '2019-09-09'},
{'hour': 10, 'operator': 'joseph_doe', 'duration': [5, 6, 7, 8, 9], 'date': '2019-09-09'}
]
index = 0
while index < len(chats) - 1:
chat = chats[index]
# detect if there is another "similar" chat in the list (before this one)
first_index = next(
i for i, first_chat in enumerate(chats)
if chat.get('hour') == first_chat.get('hour') and chat.get('operator') == first_chat.get('operator')
)
# if the first index found is not this one:
# - concat `duration` arrays
# - delete this (duplicated) chat
if index != first_index:
chats[first_index]['duration'] += chat['duration']
del chats[index]
# otherwise continue and increment the index
else:
index += 1
print(chats)
I'm trying to build my own speech recognition network. I understood how to pre-process audio. But I can't figure out the pre-processing of the text.
I have a alphabet:
alphabet = {'a': 1, 'b': 2, 'c': 3, 'd': 4, 'e': 5, 'f': 6, 'g': 7, 'h': 8, 'i': 9, 'j': 10, 'k': 11, 'l': 12, 'm': 13, 'n': 14,'o': 15, 'p': 16, 'q': 17, 'r': 18, 's': 19, 't': 20, 'u': 21, 'v': 22, 'w': 23, 'x': 24, 'y': 25, 'z': 26}
And I encode each letter of the sentence into a number (27 is a space):
array([list([27, 23, 8, 5, 14, 27, 8, 5, 27, 19, 16, 5, 1, 11, 19, 27, 9, 14, 27, 15, 21, 18, 27, 12, 1, 14, 7, 21, 1, 7, 5, 27, 9, 27, 3, 1, 14, 27, 9, 14, 20, 5, 18, 16, 18, 5, 20, 27, 23, 8, 1, 20, 27, 8, 5, 27, 8, 1, 19, 27, 19, 1, 9, 4, 27]),
list([27, 19, 15, 27, 14, 15, 23, 27, 9, 27, 6, 5, 1, 18, 27, 14, 15, 20, 8, 9, 14, 7, 27, 2, 5, 3, 1, 21, 19, 5, 27, 9, 20, 27, 23, 1, 19, 27, 20, 8, 15, 19, 5, 27, 15, 13, 5, 14, 19, 27, 20, 8, 1, 20, 27, 2, 18, 15, 21, 7, 8, 20, 27, 25, 15, 21, 27, 20, 15, 27, 13, 5, 27]),
list([27, 14, 9, 7, 8, 20, 27, 6, 5, 12, 12, 27, 1, 14, 4, 27, 1, 14, 27, 1, 19, 19, 15, 18, 20, 13, 5, 14, 20, 27, 15, 6, 27, 6, 9, 7, 8, 20, 9, 14, 7, 27, 13, 5, 14, 27, 1, 14, 4, 27, 13, 5, 18, 3, 8, 1, 14, 20, 19, 27, 5, 14, 20, 5, 18, 5, 4, 27, 1, 14, 4, 27, 5, 24, 9, 20, 5, 4, 27, 20, 8, 5, 27, 20, 5, 14, 20, 27]),
list([27, 9, 27, 8, 5, 1, 18, 4, 27, 1, 27, 6, 1, 9, 14, 20, 27, 13, 15, 22, 5, 13, 5, 14, 20, 27, 21, 14, 4, 5, 18, 27, 13, 25, 27, 6, 5, 5, 20, 27]),
list([27, 25, 15, 21, 27, 3, 1, 13, 5, 27, 19, 15, 27, 20, 8, 1, 20, 27, 25, 15, 21, 27, 3, 15, 21, 12, 4, 27, 12, 5, 1, 18, 14, 27, 1, 2, 15, 21, 20, 27, 25, 15, 21, 18, 27, 4, 18, 5, 1, 13, 19, 27, 19, 1, 9, 4, 27, 20, 8, 5, 27, 15, 12, 4, 27, 23, 15, 13, 1, 14, 27])],
dtype=object)
Here are 5 sentences.
I just create one network layer and try to transfer this data there in order to get a number corresponding to the letter.
model = Sequential()
model.add(Dense(27, input_shape=(20,), activation='softmax'))
model.compile(loss='mean_squared_error',optimizer='Adam', metrics=['accuracy'])
for X, y in batch(X_train, y_train, 5):
model.train_on_batch(X, y)
batch() just breaks X_train, y_train into batch.
5 is size of batch.
But when I try to start the network I get an error
Error when checking target: expected dense_25 to have shape (27,) but got array with shape (1,)
UPD:
I'm using MFCC for X
audio, sr = librosa.load(pathTrain+"\\"+str(file), mono=True, sr=None)
fileMFCC = librosa.feature.mfcc(audio)
mean_scale = np.mean(fileMFCC, axis=0)
std_scale = np.std(fileMFCC, axis=0)
fileMFCC = (fileMFCC - mean_scale[np.newaxis, :]) / std_scale[np.newaxis, :]
X is
[array([[-4.35889894, -4.35889894, -4.35455134, ..., -3.95851777,
-3.99308173, -4.05261022],
[ 0.22941573, 0.22941573, 0.31913073, ..., 1.87189324,
1.7987301 , 1.66804349],
[ 0.22941573, 0.22941573, 0.31165866, ..., -0.27962786,
-0.19009062, -0.13788484],
...,
[ 0.22941573, 0.22941573, 0.18657944, ..., 0.14699792,
0.12751924, 0.16724807],
[ 0.22941573, 0.22941573, 0.18478513, ..., 0.00674492,
-0.04570105, 0.01231168],
[ 0.22941573, 0.22941573, 0.18232521, ..., 0.2571599 ,
0.22477036, 0.09153304]])
etc.