Demo returns: ValueError: too many values to unpack (expected 2)

[PrettyJerry]

hi, when i exec the command in the url: https://github.com/logpai/loglizer/blob/master/docs/demo.md

python PCA_demo_without_labels.py

it returns:

(base) localhost:demo guojie$ python PCA_demo_without_labels.py
====== Input data summary ======
Loading ../data/HDFS/HDFS_100k.log_structured.csv
Total: 7940 instances, train: 3970 instances, test: 3970 instances
Traceback (most recent call last):
File "PCA_demo_without_labels.py", line 24, in
split_type='sequential', save_csv=True)
ValueError: too many values to unpack (expected 2)

[MouadAouni]

Hey there Jerry,

I hope that you are okay and that youa re doing well. I would like to ask your help on this.
I had the same error as you here, and i could not figure out how to fix it. Did you find the solution ? If yes could you share it with me please ?

Thanks !

[errorhandlerst]

@MouadAouni you only have to remove data_df
from dataloader.py-->return (x_train, None), (x_test, None), data_df
That's it

[MouadAouni]

@errorhandlerst thank you for the reply. It helped me avoid this problem to fall in another one lol.

There it is :

Train phase:
====== Model summary ======
Traceback (most recent call last):
File "PCA_demo_without_labels.py", line 34, in
model.fit(x_train)
File "..\loglizer\models\PCA.py", line 67, in fit
n_components = i + 1
UnboundLocalError: local variable 'i' referenced before assignment

What do you think ?
Thanks again.

EDIT : I'm using here my own data, it isn't the same as theirs, and i have just 4 columns :

Content | EventId | EventTemplate | ParameterList

[errorhandlerst]

@MouadAouni here you have to apply one space make n_components = i + 1 inside for loop:-- in PCA.py

[MouadAouni]

@errorhandlerst I already tried this but got this error :

Train phase:
====== Model summary ======
Traceback (most recent call last):
File "PCA_demo_without_labels.py", line 34, in
model.fit(x_train)
File "..\loglizer\models\PCA.py", line 69, in fit
P = U[:, :n_components]
TypeError: slice indices must be integers or None or have an index method

and then made this change: n_components = num_events + 1 and got this :

====== Input data summary ======
Loading ../data/HDFS/mover.txt_structured.csv
Total: 0 instances, train: 0 instances, test: 0 instances
====== Transformed train data summary ======
..\loglizer\preprocessing.py:102: RuntimeWarning: Mean of empty slice.
mean_vec = X.mean(axis=0)
Train data shape: 0-by-0

Train phase:
====== Model summary ======
n_components: 1
Project matrix shape: 0-by-0
..\loglizer\models\PCA.py:83: RuntimeWarning: invalid value encountered in double_scalars
h0 = 1.0 - 2 * phi[0] * phi[2] / (3.0 * phi[1] * phi[1])
SPE threshold: nan

Test phase:
====== Input data summary ======
Loading ../data/HDFS/mover.txt_structured.csv
Total: 0 instances, train: 0 instances, test: 0 instances
====== Transformed test data summary ======
Test data shape: 0-by-0

[errorhandlerst]

@MouadAouni Did you change the load_hdfs data according your data? if not then you should change because load_hdfs method used session window .You might require fixed window (First check PCA working with default hdfs data If that working fine then might be problem id related to your data)

[pavansai-k]

@errorhandlerst
Do we have any resolution for this error yet ?

ValueError: too many values to unpack (expected 2)

[huhui]

hi, when i exec the command in the url: https://github.com/logpai/loglizer/blob/master/docs/demo.md

python PCA_demo_without_labels.py

it returns:

(base) localhost:demo guojie$ python PCA_demo_without_labels.py
====== Input data summary ======
Loading ../data/HDFS/HDFS_100k.log_structured.csv
Total: 7940 instances, train: 3970 instances, test: 3970 instances
Traceback (most recent call last):
File "PCA_demo_without_labels.py", line 24, in
split_type='sequential', save_csv=True)
ValueError: too many values to unpack (expected 2)

Solution:
train_test_tuple = dataloader.load_HDFS(..........)
(x_train, y_train), (x_test, y_test) = train_test_tuple[0], train_test_tuple[1]

[pavansai-k]

@huhui , will this demo without label codes work on Android Log file datasets ?