Shouvik, who is allegedly cool and fun, made us look at this on Friday:
import numpy as np
img_arr = np.array([
[1, 2, 3],
[4, 5, 6]
])
At least I think that's what happened, I found it boring and fell asleep instantly.
In my dream, I imaged the impossible... what if numpy was good?
Images
I literally only ever used numpy for one thing ever.
import numpy as np
from PIL import Image as im
I used it with the famous "Python Image Library" to make digital art.
This is an actual assignment in one of my upper division classes: brot.py
Images
It is relatively easy to convert a np array into an image:
A numpy array is a basically a list of rows.
Each row is a list of "pixels" - the tiniest unit our screen can display.
The value in some pixel in some row gives how bright that pixel is.
Image a computer doing pointillism, perhaps.
Pointillism is a technique of painting in which small, distinct dots of color are applied in patterns to form an image.
So we say: take 2 steps over and 1 steps down, and using a relatively dark color, perhaps.
step_over = 2
step_down = 1
brightnes = 0 # relatively dark
img_arr[step_over][step_down] = brightnes
By the way I have no idea if over/down is correct - check it by making things bright or dark.
Images
Once we have set any brightness in our array, we use the following code from the image library:
im.fromarray(img_arr)
# 'im.fromarray' from PIL
# 'img_arr' we just made
I think by default this doesn't work because PIL has very specific tastes in arrays.
Recall: Shouvik told us arrays have different types.
Images
For whatever reason, we have to use "np.uint8" (look it up when you feel like it)
img_arr = np.array([
[1, 2, 3],
[4, 5, 6]
]).astype(np.uint8) # From an error message
im.fromarray(img_arr)
This will make a very small image.
Images
To make something big enough to see...
# let's make a 110 by 120 image of brightness 130
reg_arr = [[130 for i in range(110)] for i in range(120)]
img_arr = np.array(reg_arr).astype(np.uint8)
im.fromarray(img_arr)
This will make a very small image.
Group 1
Group 2
Group 3
Group 4
Question #1
How are you enjoying CC so far?
Think about a possible answer on your own
Discuss your answers with the rest of the group
Record a summary of each group’s discussion
Question #2
Do we like images more than, say, lists or pandas dataframes or numpy arrays?
Think about a possible answer on your own
Discuss your answers with the rest of the group
Record a summary of each group’s discussion
Question #3
What makes something art?
Think about a possible answer on your own
Discuss your answers with the rest of the group
Record a summary of each group’s discussion
Digital Art
Exercise C.1 Create something you find beautiful.
Consider the following:
You can determine maximum brightness by increasing it until something breaks.
Computer monitors don't display only brightness, but color, using 3 brightnesses.
Gradients, changes in color over space, occur by determining color based on the position in an image.
I claim: beauty is created by community - work together to design and build.
# let's make a 110 by 120 image of increasing brightness.
reg_arr = [[i for i in range(110)] for j in range(120)]
img_arr = np.array(reg_arr, dtype=np.uint8)
im.fromarray(img_arr)
Bonus
Exercise C.1 Create something you find beautiful.
My first thing:
reg_arr = [[i for i in range(110)] for j in range(120)]
img_arr = np.array(reg_arr).astype(np.uint8)
im.fromarray(img_arr)
Bonus
Exercise C.1 Create something you find beautiful.
Something beautiful (to me):
im.fromarray(np.array([[[i + j, i + -j,-i + j] for i in range(255)] for j in range(255)]).astype(np.uint8))
Cramming this all on one line is admittedly obnoxious, but I hope will make you thing for yourself while still learning something (let me know).
Question #4
What do you do when given a hard problem?
Think about a possible answer on your own
Discuss your answers with the rest of the group
Record a summary of each group’s discussion
Homework
Homework 3 Electoral Votes
Using multiple line strings, the new .split() method, and what you know about indices, write code that takes the information from Archives.gov and creates a list of states and their number of electoral votes.
This is a special list that will help us latter.
You may need many lines of code or many cells.
Whatever you submit you should test using "run all" or "ctrl+F9" to make sure that when I run your code, or when you run the code in the future, that it does what you expect.
You may make a copy of this Notebook and share it with me and anyone with whom you collaborate.
Give Shouvik and I edit permissions.
Due Fri @ 12 Noon (so we're able to discuss it in class)
You may only give your peers view or comment permissions.
It is more important to try things that interest you than to complete the assignment.
