extract multiple ranges from a list

[pi.arctan]

Hi guys,

One of my many project involves working with YUV-files, where I need
to reduce
the vertical resolution with a factor of two, i.e. remove every other
scan line.
Today I'm using two for-loops in the fashion shown below

y = []
for i in range(0, width*height, width*2):
for j in range(0,width):
y.append(Y[i+j])

This approach doesn't feel very pythonic but I can't come up with a
better idea to do it.
I've tried list comprehension and map together with lambda but I can't
get a flattened list
of every other scan-line...

CIF = 352x288 items for luminance and the aim is to have the list
below:
y = [0:352 704:1056 ... ]

//Fredrik

[Paul Hankin]

On Mar 8, 3:28*pm, [e-mail] wrote:

Hi guys,

One of my many project involves working with YUV-files, where I need
to reduce
the vertical resolution with a factor of two, i.e. remove every other
scan line.
Today I'm using two for-loops in the fashion shown below

y = []
for i in range(0, width*height, width*2):
* * for j in range(0,width):
* * * * y.append(Y[i+j])

There's nothing really wrong with your code. Maybe it's a little nicer
written like this:

y = []
for i in range(0, height, 2):
y.extend(Y[i * width + j] for j in range(width))

--
Paul Hankin

[Peter Otten]

pi*******@gmail.com wrote:

One of my many project involves working with YUV-files, where I need
to reduce
the vertical resolution with a factor of two, i.e. remove every other
scan line.
Today I'm using two for-loops in the fashion shown below

y = []
for i in range(0, width*height, width*2):
for j in range(0,width):
y.append(Y[i+j])

This approach doesn't feel very pythonic but I can't come up with a
better idea to do it.
I've tried list comprehension and map together with lambda but I can't
get a flattened list
of every other scan-line...

CIF = 352x288 items for luminance and the aim is to have the list
below:
y = [0:352 704:1056 ... ]

>>width = 3; height = 5
Y = range(width*height)
y = []
for i in range(0, width*height, 2*width):

.... y.extend(Y[i:i+width])
....

>>y

[0, 1, 2, 6, 7, 8, 12, 13, 14]

Probably more efficient, but needs numpy:

>>import numpy
width = 3
height = 5
Y = range(width*height)
a = numpy.array(Y).reshape(height, width)
a

array([[ 0, 1, 2],
[ 3, 4, 5],
[ 6, 7, 8],
[ 9, 10, 11],
[12, 13, 14]])

>>b = a[::2]
b

array([[ 0, 1, 2],
[ 6, 7, 8],
[12, 13, 14]])

>>list(b.reshape(len(b)*width))

[0, 1, 2, 6, 7, 8, 12, 13, 14]

Peter

[pi.arctan]

On 8 Mar, 17:32, Peter Otten <[e-mail]:

[e-mail] wrote:

One of my many project involves working with YUV-files, where I need
to reduce
the vertical resolution with a factor of two, i.e. remove every other
scan line.
Today I'm using two for-loops in the fashion shown below

y = []
for i in range(0, width*height, width*2):
for j in range(0,width):
y.append(Y[i+j])

This approach doesn't feel very pythonic but I can't come up with a
better idea to do it.
I've tried list comprehension and map together with lambda but I can't
get a flattened list
of every other scan-line...

CIF = 352x288 items for luminance and the aim is to have the list
below:
y = [0:352 704:1056 ... ]

>width = 3; height = 5
Y = range(width*height)
y = []
for i in range(0, width*height, 2*width):

... y.extend(Y[i:i+width])
...>>y

[0, 1, 2, 6, 7, 8, 12, 13, 14]

Probably more efficient, but needs numpy:

>import numpy
width = 3
height = 5
Y = range(width*height)
a = numpy.array(Y).reshape(height, width)
a

array([[ 0, 1, 2],
[ 3, 4, 5],
[ 6, 7, 8],
[ 9, 10, 11],
[12, 13, 14]])>>b = a[::2]

>b

array([[ 0, 1, 2],
[ 6, 7, 8],
[12, 13, 14]])>>list(b.reshape(len(b)*width))

[0, 1, 2, 6, 7, 8, 12, 13, 14]

Peter

Thanx guys!