🔒 how to graph binary 2d array of pixels? | Python | Page 1

queen vine Mar 27, 2024, 7:40 PM

#

array like this (but a lot bigger)```
[[1,0,0],
[1,0,1],
[1,1,1]]

it represents each pixel and 1 being black and 0 being white
how do I graph it?```

sturdy sableBOT Mar 27, 2024, 7:40 PM

#

queen vine array like this (but a lot bigger)``` [[1,0,0], [1,0,1], [1,1,1]] ``` it repre...

#

@queen vine

Python help channel opened

Remember to:

Ask your Python question, not if you can ask or if there's an expert who can help.
Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

strong prairie Mar 27, 2024, 7:45 PM

#

queen vine array like this (but a lot bigger)``` [[1,0,0], [1,0,1], [1,1,1]] ``` it repre...

what library are you using?

#

for graphing

queen vine Mar 27, 2024, 7:45 PM

#

strong prairie what library are you using?

matplotlib

strong prairie Mar 27, 2024, 7:45 PM

#

are you using jupter notebooks?

queen vine Mar 27, 2024, 7:45 PM

#

strong prairie are you using jupter notebooks?

nope

strong prairie Mar 27, 2024, 7:46 PM

#

are you using data science based libraries?

queen vine Mar 27, 2024, 7:46 PM

#

I can use them sure

#

numpy

#

I am not using scipy but I can

strong prairie Mar 27, 2024, 7:46 PM

#

alright, you can simply convert your pixels to numpy array

queen vine Mar 27, 2024, 7:46 PM

#

strong prairie alright, you can simply convert your pixels to numpy array

they are in one

whole viper Mar 27, 2024, 7:46 PM

#

You'll probably want to use matplotlib.pyplot.imshow.

queen vine Mar 27, 2024, 7:47 PM

#

whole viper You'll probably want to use matplotlib.pyplot.imshow.

the colormap is seperate tho

whole viper Mar 27, 2024, 7:47 PM

#

It's by default normalised.

#

and you can set the vmin and vmax manually

#

and the cmap

queen vine Mar 27, 2024, 7:48 PM

#

how do I fix this py img = ax.imshow( img, extent=[*big_range, *big_range], aspect="equal", interpolation="gaussian", origin="lower", )

#

#

I have to run it 5 times for it to work

#

img = network.calculate_outputs(pixels)
    #ax.pcolor(img, cmap='binary')
    #img = ax.pcolormesh(img, cmap='binary')
    custom_cmap = matplotlib.colors.ListedColormap(["#ADD8E6", "white"])

    img = ax.imshow(
        img,
        cmap=custom_cmap,
        extent=[*big_range, *big_range],
        aspect="equal",
        interpolation="gaussian",
        origin="lower",
    )```

whole viper Mar 27, 2024, 7:50 PM

#

You look like you know more about matplotlib than I do.

queen vine Mar 27, 2024, 7:50 PM

#

also half of the weights act weird with the graph so I dont think that graphs it correctly

queen vine Mar 27, 2024, 7:51 PM

#

whole viper You look like you know more about matplotlib than I do.

idk what that stuff even does

#

just combined and copied stuff from stackoverflow

queen vine Mar 27, 2024, 7:52 PM

#

queen vine also half of the weights act weird with the graph so I dont think that graphs it...

i know it is how I am graphing it since if I scatter plot it, it works

#

I cant make this shape with the imshow

whole viper Mar 27, 2024, 7:56 PM

#

from matplotlib import pyplot as plt

...

plt.imshow(arr, cmap='gray')
plt.show()```This is all I'd end up doing.

#

But yeah, if it's not liking the lists, convert to a numpy array and it should work fine.

queen vine Mar 27, 2024, 7:58 PM

#

#

 points = np.column_stack((x, y, colors.astype(int)))
    xs, ys = np.meshgrid(np.linspace(*big_range), np.linspace(*big_range))
    pixels = np.stack((xs.T, ys.T), axis=2)

    from scipy.spatial import ConvexHull
    
    graph = None

    def accuracy():
        nonlocal graph
        if graph:
            graph.remove()

        pointsX = []
        pointsY = []
        colors = []

        total = 0
        for point in points:
            px, py, value = point
            output = network.calculate_output((px, py))
            pointsX.append(px)
            pointsY.append(py)
            colors.append(output == value)
            
            total += np.array(output > 0).astype(int) == value

        img = network.calculate_outputs(pixels)
        
        graph = ax.scatter(pointsX, pointsY, np.array(colors) * 100, c="black")
        
    img = network.calculate_outputs(pixels)
    custom_cmap = matplotlib.colors.ListedColormap(["#ADD8E6", "white"])

    img = ax.imshow(
        img,
        cmap=custom_cmap,
        extent=[*big_range, *big_range],
        aspect="equal",
        interpolation="gaussian",
        origin="lower",
    )

#

dont question the nonlocal

whole viper Mar 27, 2024, 8:00 PM

#

If you know enough to use it and that it even exists, I'm going to assume you know when and where to use it.

queen vine Mar 27, 2024, 8:00 PM

#

i just am using it cause I am testing

#

@whole viper maybe it is the way I am getting the pixels? py points = np.column_stack((x, y, colors.astype(int))) xs, ys = np.meshgrid(np.linspace(*big_range), np.linspace(*big_range)) pixels = np.stack((xs.T, ys.T), axis=2)

#

I want pixels to be [(1, 2), (4, 3) ....]

#

and then after pasted through my neural network to be [[1,1,1],[0,0,1],...]

whole viper Mar 27, 2024, 8:09 PM

#

Oh. That's not what I thought your data looked like.

queen vine Mar 27, 2024, 8:12 PM

#

is there a way to flatten everything except the inner most (1,2),(1,2)

whole viper Mar 27, 2024, 8:12 PM

#

queen vine I want pixels to be [(1, 2), (4, 3) ....]

If you have a y by x grid of bools, and you want the location of all the 1s, you can use numpy.where(arr) that'll give you ((y, y, y, y), (x, x, x, x)) style data, which you can feed into an array then use .T on it.

queen vine Mar 27, 2024, 8:12 PM

#

cause pixels ... [[-2.85714286e-02 -2.00000000e-01] [-2.85714286e-02 -1.71428571e-01] [-2.85714286e-02 -1.42857143e-01] [-2.85714286e-02 -1.14285714e-01] [-2.85714286e-02 -8.57142857e-02] [-2.85714286e-02 -5.71428571e-02] [-2.85714286e-02 -2.85714286e-02] [-2.85714286e-02 -2.77555756e-17] [-2.85714286e-02 2.85714286e-02] [-2.85714286e-02 5.71428571e-02] [-2.85714286e-02 8.57142857e-02] [-2.85714286e-02 1.14285714e-01] [-2.85714286e-02], [[-2.85714286e-02 -2.00000000e-01] [-2.85714286e-02 -1.71428571e-01] [-2.85714286e-02 -1.42857143e-01] [-2.85714286e-02 -1.14285714e-01] [-2.85714286e-02 -8.57142857e-02] [-2.85714286e-02 -5.71428571e-02] [-2.85714286e-02 -2.85714286e-02] [-2.85714286e-02 -2.77555756e-17] [-2.85714286e-02 2.85714286e-02] [-2.85714286e-02 5.71428571e-02] [-2.85714286e-02 8.57142857e-02] [-2.85714286e-02 1.14285714e-01] [-2.85714286e-02

queen vine Mar 27, 2024, 8:13 PM

#

whole viper If you have a y by x grid of bools, and you want the location of all the 1s, you...

i think the problem is calculating all the pixels

whole viper Mar 27, 2024, 8:14 PM

#

I'm lost, sorry.

#

You started simple.

#

Now it is not.

queen vine Mar 27, 2024, 8:15 PM

#

whole viper I'm lost, sorry.

if I have a range from 0->1 (through linspace) how do I get all diferent pairs like (0.01, 0.01), (0.02, 0.01), (0.03, 0.01), ... (0.01, 0.02),,,

whole viper Mar 27, 2024, 8:16 PM

#

What's the population?

queen vine Mar 27, 2024, 8:16 PM

#

1 person

#

jk wdym

whole viper Mar 27, 2024, 8:16 PM

#

!d numpy.linspace

sturdy sableBOT Mar 27, 2024, 8:16 PM

#

numpy.linspace


numpy.linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None, axis=0, *, device=None)```
Return evenly spaced numbers over a specified interval.

Returns *num* evenly spaced samples, calculated over the interval [*start*, *stop*].

The endpoint of the interval can optionally be excluded.

Changed in version 1.16.0: Non-scalar *start* and *stop* are now supported.

Changed in version 1.20.0: Values are rounded towards `-inf` instead of `0` when an integer `dtype` is specified. The old behavior can still be obtained with `np.linspace(start, stop, num).astype(int)`

whole viper Mar 27, 2024, 8:16 PM

#

num

queen vine Mar 27, 2024, 8:16 PM

#

anything I want

#

how slow I want it to be

whole viper Mar 27, 2024, 8:16 PM

#

Give me an idea.

queen vine Mar 27, 2024, 8:17 PM

#

100

#

start 0 , stop 1

whole viper Mar 27, 2024, 8:17 PM

#

Is the order of the pairs significant?

queen vine Mar 27, 2024, 8:18 PM

#

no

whole viper Mar 27, 2024, 8:18 PM

#

Worst case scenario, how large is num?

queen vine Mar 27, 2024, 8:18 PM

#

150

whole viper Mar 27, 2024, 8:18 PM

#

Because you could use itertools.combinations.

queen vine Mar 27, 2024, 8:18 PM

#

if it is too slow I would lower num

queen vine Mar 27, 2024, 8:19 PM

#

whole viper Because you could use itertools.combinations.

is that like faster?

whole viper Mar 27, 2024, 8:19 PM

#

Eh, I think it'd just be num ** 2 to calculate.

#

That's not terrible.

queen vine Mar 27, 2024, 8:19 PM

#

how do I use it

queen vine Mar 27, 2024, 8:20 PM

#

whole viper Eh, I think it'd just be num ** 2 to calculate.

only have to calculate it once

#

my code didnt realize I only needed to do it once

whole viper Mar 27, 2024, 8:20 PM

#

!d itertools.combinations

sturdy sableBOT Mar 27, 2024, 8:21 PM

#

itertools.combinations


itertools.combinations(iterable, r)```
Return *r* length subsequences of elements from the input *iterable*.

The combination tuples are emitted in lexicographic ordering according to the order of the input *iterable*. So, if the input *iterable* is sorted, the output tuples will be produced in sorted order.

Elements are treated as unique based on their position, not on their value. So if the input elements are unique, there will be no repeated values in each combination.

Roughly equivalent to:

queen vine Mar 27, 2024, 8:21 PM

#

itertools.combinations(np.linspace(0,1,100), 2)?

#

that prob doesnt work

whole viper Mar 27, 2024, 8:21 PM

#

!e py import itertools data = [1, 2, 3] result = [*itertools.combinations(data, 2)] print(result)

sturdy sableBOT Mar 27, 2024, 8:21 PM

#

@whole viper :white_check_mark: Your 3.12 eval job has completed with return code 0.

[(1, 2), (1, 3), (2, 3)]

whole viper Mar 27, 2024, 8:23 PM

#

Hang on..

queen vine Mar 27, 2024, 8:23 PM

#

ok I willtry that

#

while you hang on

whole viper Mar 27, 2024, 8:23 PM

#

product

#

!d itertools.product

sturdy sableBOT Mar 27, 2024, 8:23 PM

#

itertools.product


itertools.product(*iterables, repeat=1)```
Cartesian product of input iterables.

Roughly equivalent to nested for-loops in a generator expression. For example, `product(A, B)` returns the same as `((x,y) for x in A for y in B)`.

The nested loops cycle like an odometer with the rightmost element advancing on every iteration. This pattern creates a lexicographic ordering so that if the input’s iterables are sorted, the product tuples are emitted in sorted order.

To compute the product of an iterable with itself, specify the number of repetitions with the optional *repeat* keyword argument. For example, `product(A, repeat=4)` means the same as `product(A, A, A, A)`.

whole viper Mar 27, 2024, 8:25 PM

#

I've used this before, but I'm having a stupid brain moment.

#

Oh, right.

#

!e py import itertools data = [1, 2, 3] result = [*itertools.product(data, repeat=2)] print(result)

queen vine Mar 27, 2024, 8:26 PM

#

so itertools.product(np.linespace?

sturdy sableBOT Mar 27, 2024, 8:26 PM

#

@whole viper :white_check_mark: Your 3.12 eval job has completed with return code 0.

[(1, 1), (1, 2), (1, 3), (2, 1), (2, 2), (2, 3), (3, 1), (3, 2), (3, 3)]

queen vine Mar 27, 2024, 8:26 PM

#

ohh

#

that looks more right

whole viper Mar 27, 2024, 8:27 PM

#

There we go. It needed it as a keyword argument, because *iterables.

#

for a in data:
    for b in data:
        (a, b)```Basically.

queen vine Mar 27, 2024, 8:28 PM

#

oh darn it File "c:\Users\jacob\AppData\Local\Programs\Python\Python311\Lib\site-packages\matplotlib\axes_axes.py", line 5665, in imshow
im.set_data(X)
File "c:\Users\jacob\AppData\Local\Programs\Python\Python311\Lib\site-packages\matplotlib\image.py", line 710, in set_data
raise TypeError("Invalid shape {} for image data"
TypeError: Invalid shape (10000,) for image data

#

thats why I had ugly meshgrid

#

but then the input to my neural network is broken

#

okay so I have a bunch of pixels [(1,2),(2,2),(3,2)...] and I have whether to color the pixel white or black [0,1,0,...] @whole viper

#

how do I graph that now

#

wait what if every 100 index it becomes a new array so that the black and white are 2d
[[0,1,0,0,0,0,0,... (100 times)],
[0,1,0,0,0,0,0,... (100 times)],

#

then maybe imshow can graph that

#

but that breaks the order....

#

nvm the pixels do need order

#

!d numpy.meshgrid

sturdy sableBOT Mar 27, 2024, 8:36 PM

#

numpy.meshgrid


numpy.meshgrid(*xi, copy=True, sparse=False, indexing='xy')```
Return a tuple of coordinate matrices from coordinate vectors.

Make N-D coordinate arrays for vectorized evaluations of N-D scalar/vector fields over N-D grids, given one-dimensional coordinate arrays x1, x2,…, xn.

Changed in version 1.9: 1-D and 0-D cases are allowed.

queen vine Mar 27, 2024, 8:38 PM

#

queen vine wait what if every 100 index it becomes a new array so that the black and white ...

how do I do this @whole viper ?

whole viper Mar 27, 2024, 8:39 PM

#

I don't understand how to help you at the present time. I don't understand the problem as it is being presented.

queen vine Mar 27, 2024, 8:42 PM

#

let me test some stuff

queen vine Mar 27, 2024, 8:46 PM

#

queen vine wait what if every 100 index it becomes a new array so that the black and white ...

.reshape(100, -1)

sturdy sableBOT Mar 27, 2024, 9:46 PM

#

Python help channel closed

This help channel has been closed and it's no longer possible to send messages here. If your question wasn't answered, feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.

#🔒 how to graph binary 2d array of pixels?