In [1]:

    

import nibabel as nib


    

In [2]:

    

nii = nib.load('862ss2c500regcUKFCC.nii')


    

In [3]:

    

data = nii.get_data()


    

In [4]:

    

data.shape


    

    
Out[4]:





(42, 32, 15, 2671)

In [5]:

    

import numpy as np


    

In [6]:

    

D = np.reshape(data,(42*32*15,2671))


    

In [7]:

    

np.save('centraltest.npy', D)


    

In [10]:

    

d = tsc.loadSeries('centraltest.npy', inputFormat='npy', minPartitions=5)
d.first()


    

    
Out[10]:





(0, array([-0.20409763, -0.15216248, -0.15467691, ...,  0.11422426,
         0.07105274,  0.03869029], dtype=float32))

In [30]:

    

d


    

    
Out[30]:





Series
nrecords: None (inspect to compute)
dtype: float32
dims: None (inspect to compute)
index: [   0    1    2 ..., 2668 2669 2670]

In [12]:

    

data2 = d.toTimeSeries().normalize()


    

In [13]:

    

data2.cache()


    

    
Out[13]:





TimeSeries
nrecords: None (inspect to compute)
dtype: None (inspect to compute)
dims: None (inspect to compute)
index: [   0    1    2 ..., 2668 2669 2670]

In [14]:

    

from thunder import PCA


    

In [15]:

    

model = PCA(k=40).fit(data2)


    

In [25]:

    

imgs = model.scores


    

In [20]:

    

%matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_context('notebook')
sns.set_style('darkgrid')
from thunder import Colorize
image = Colorize.image
plt.plot(model.comps.T);


    

In [21]:

    

model


    

    
Out[21]:





<thunder.factorization.pca.PCA at 0x7f8cac7b93d0>

In [22]:

    

A=model.comps


    

In [24]:

    

A.shape


    

    
Out[24]:





(40, 2671)

In [26]:

    

imgs.shape


    

    




---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-26-d95c9f202c6a> in <module>()
----> 1 imgs.shape

AttributeError: 'RowMatrix' object has no attribute 'shape'

In [28]:

    

svdtry = thunder.SVD(k=40).calc(data2)


    

In [36]:

    

svdu=svdtry.u


    

In [32]:

    

np.save('svds.npy',svdtry.s)


    

In [33]:

    

np.save('svdv.npy',svdtry.v)


    

In [37]:

    

svdu.saveAsBinarySeries('svdu')


    

    




---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-37-bdc2969400dd> in <module>()
----> 1 svdu.saveAsBinarySeries('svdu')

/root/thunder/python/thunder/rdds/series.pyc in saveAsBinarySeries(self, outputDirPath, overwrite)
    867 
    868         if not overwrite:
--> 869             self._checkOverwrite(outputDirPath)
    870             overwrite = True  # prevent additional downstream checks for this path
    871 

/root/thunder/python/thunder/rdds/data.pyc in _checkOverwrite(self, outputDirPath)
    115         from thunder.utils.common import AWSCredentials, raiseErrorIfPathExists
    116         awsCredentialOverride = AWSCredentials.fromContext(self.rdd.ctx)
--> 117         raiseErrorIfPathExists(outputDirPath, awsCredentialsOverride=awsCredentialOverride)
    118 
    119     def first(self):

/root/thunder/python/thunder/utils/common.pyc in raiseErrorIfPathExists(path, awsCredentialsOverride)
    221     if existing:
    222         raise ValueError("Path %s appears to already exist. Specify a new directory, or call " % path +
--> 223                          "with overwrite=True to overwrite.")
    224 
    225 

ValueError: Path svdu appears to already exist. Specify a new directory, or call with overwrite=True to overwrite.

In [ ]: