In [1]:
import numpy as np
In [2]:
a = np.array( [1, 2, 3, 4, 5], float)
a
Out[2]:
In [6]:
b = np.array( [9,8,9], int)
print(b.dtype)
In [7]:
a = np.array([1,2,3])
b = np.array([4,5,6])
a+b
Out[7]:
In [8]:
# np.arange is the numpy equivalnt to range
a = np.arange(12)
# access to single values
a[1]
# slicing
a[2:5]
a[2:]
a[:3]
# access from the back
a[-1]
# acessing every n-th element
a[2:10:2]
a[::3]
# reassignment
a[0] = 4
a
Out[8]:
In [10]:
# iteration
for elem in a:
print(elem)
In [11]:
a==2
a>5
# index using a comparison
a[a > 5]
Out[11]:
In [12]:
A = np.array( [[1,2,3], [4,5,6], [7,8,9]])
A
Out[12]:
In [14]:
print(A[0,0])
print(A[2,:])
print(A[:,0])
print(A[::2,1])
In [15]:
print(len(A)) # length of 1st dimension
print(A.shape) # shape of the arrays
In [16]:
print(2 in A)
print(13 in A)
In [17]:
c = np.array([1,2,3,4,5,6])
print(c)
c.reshape( (3,2) ) # reshape takes a dimension tuple as input (#rows,#cols)
Out[17]:
In [18]:
a = np.array([1,2,3])
b = a
c = a.copy()
a[0]=6
print(a)
print(b)
print(c)
In [19]:
# convert to python list
a.tolist()
# fill up with new value
a.fill(12)
print(A)
print(A.transpose()) # transpose matrix
print(A.T) # same as A.transpose()
In [20]:
# concatenate arrays
B = A.transpose()
np.concatenate( (A,B) )
np.concatenate( (A,B), axis=1 )
Out[20]:
In [21]:
# equally spaced values within an interval: arange( start, stop, step )
np.arange(0, 5, 0.1)
Out[21]:
In [23]:
# generate 1 or 0-arrays
print(np.ones( (2,3) ))
print(np.zeros( (3,4) ))
In [24]:
# construct an array with a similar shape to another one
np.zeros_like(A)
Out[24]:
In [25]:
# identity matrix
Id = np.identity(4)
Id
Out[25]:
In [26]:
A = np.array([[1,2],[3,4]])
B = np.array([[0,1],[1,0]])
A + B
A - B
A * B
B / A
A ** B
Out[26]:
In [27]:
A = np.array( [[0,1], [1,0]] )
v = np.array( [6,7] )
np.dot( A,v ) # matrix product
Out[27]:
In [28]:
A = np.ones( (3,3) )
c = np.array([1,2,3])
# per default arrays are added row-wise
A + c
# like this col-wise
A + c[:,np.newaxis]
Out[28]:
In [33]:
# element-wise functions
np.sqrt(a) # square root
np.sign(a) # sign
np.log(a) # natural logarithm
np.log10(a) # decadic logarithm
np.exp(a) #exponential
np.sin(a) # trigonometric (also cos, tan, arcsin, arccos, arctan)
# non-element-wise
a.sum() # sum of all elements
a.prod() # product of all elements
a.mean() # mean
a.var() # variance
a.std() # standard deviation
a.max() # maximum
a.min() # minimum
a.sort()
# matrix
np.unique( [1,1,3,3,5] ) # get unique elements
Out[33]:
In [44]:
A = np.array( [[1,2,3], [4,5,6], [7,8,9]])
np.linalg.norm(A)
np.linalg.det(A) # determinant
np.linalg.eig(A) # eigenvalues and eigenvectors
np.linalg.inv(A) # matrix inverse
np.linalg.svd(A) # singular value decomposition
Out[44]:
In [45]:
a = np.array( [1,6,3,4,9,6,7,3,2,4,5] )
a[ a>4 ] # get all elements larger than 4
a[ np.logical_and(a>4, a<12) ]
# acessing via index
indices = [1,3]
a[indices]
Out[45]:
Scipy is a collection of useful scientific algorithms
In [46]:
import scipy as sp
import scipy.optimize
In [47]:
from matplotlib import pyplot as plt
#plt.style.use('ggplot')
%matplotlib inline
In [48]:
# define a function
def myfunc( x ):
return ( 3 + x ) * (8 + x)
# plot the function
x = np.arange( -10, 10, 1 )
plt.plot(x, myfunc(x) )
Out[48]:
In [49]:
# minimize using scipy
sp.optimize.minimize_scalar( myfunc )
Out[49]:
In [65]:
p = scipy.poly1d([1,11,24])
print(p)
# derivative
print(p.deriv())
# integration
print(p.integ())
In [ ]: