I love Python

Any project ideas for practicing Python? I already know C, Java and some of C++ so I am not a total beginner, also no machine learning meme projects pls

Aside from Meme Learning, Python's best use is shell scripting.

Some good ways to get started:
>Jow Forums post and/or image scraper.
>Simple web based file server
>Some simple Monte Carlo sims

You could also find something you've done in Java and translate it into Python.

No. There's lots of large companies that are using Python as an Enterprise solution.

>to write async code the libraries you use should also be async
In which languages does this not apply?

>*changes your class conditionally*
Heh, nothin personnel

>There's lots of large companies that are using Python as an Enterprise solution
That sounds like a disaster waiting to happen.

C# or Java are the rule here for good reason. Python barely works in microservice environment with low load let alone enterprise.

Nope, works fine.

>That sounds like a disaster waiting to happen.
It is happening. I'm also waiting for 2020.
Nope, it breaks.

>Nope, it breaks.
Can assure you it doesn't.

it's not like we're cheap on RAM now, what's the matter?

What do you mean?

>lack of types would probably drive me insane
>it's not like we're cheap on RAM now
nani?

Attached: nani the fuck.jpg (400x381, 58K)

I like lua

Attached: 1528631547870.png (573x892, 213K)

I like ruby! (｡>﹏

Attached: Screenshot_2018-08-26_15-33-32.png (299x432, 100K)

That's not even a nice ass.

he said lack of types. to me having many types of variables make for a hassle making sure all get their allowed values while with python you can pretty much be not declarative and more generous. and it's not like you ought to use the right type of variable to save memory.

it has its moments
def get_evens(intArray):
return [x for x in intArray if not x%2]

Can assure you it does.

OP's a faggot that likes programming languages instead of hating them, like you rightfully should.
That is all

Well it's working fine for:
>JewTube
>PayPal
>Ebay
>OpenStack
>Jewgol

And many more.

Go away.
Make you're own thread

Well, it's not working for:
>Google (hence grumpy as a migration strategy)
>Dropbox (hence their shitty attempt at an JIT compiler and them complaining about unobscured code)
(don't know about the other botnet companies you mentioned but it's guaranteed to have shat itself)
>Blender (addon authors complaining about performance)
>Scribus (lack of proper sandboxing)
>the festering bugfest that is scientific tools

And many more.

Sorry brainlet, that's because python is effectively call-by-need. The body of the function isn't "loaded" until the function is called.
Do this:
l=[]
for a in range(10):
l.append(lambda a=a: a)
for f in l:
print(f())

>Well, it's not working for:
If it's not working for them, they'd move away from it. Large companies that you mentioned don't just stick with something if it doesn't work. Clearly it is working for them, they just want some additional features that either haven't been developed yet or aren't being developed to the specifications of the company. Hell it might even be something that started development in some staffers "N% time" that got moved to a full project because of presented clear benefits.

>they'd move away from it.
they do

But they haven't, and they aren't.

They have partially and they are, faggot.

>partially
Really? Jewgol have partially moved away from Python? Last I heard the core of YouTube was Python. Just face it, Python really does have enterprise level uses, and it's being used in that role on a daily basis.

You personally might not like it, but your dislike doesn't change the fact that it does have a place in a modern IT infrastructure.

I use it at work for data analytics and machine learning...
although we are migrating to golang because it handles concurrency better

python is popular and Jow Forums hates popular stuff.

does lua ever get lewd?

Python is the worst language in existence
It has the worst performance
No type safety
It is a language for brain lets, a horde of noobs using it for everything because it was the first language they learned and they can't let go
Using python for anything but simple scripts is to doom a project
Even Javascript is better than this shit

What the fuck
>the absolute state of python programmers

No way is core of YT python you delusional fuck
Python is so bad at performance, it cannot be used in an enterprise environment in a meaningful way, it simply cannot. Physically impossible.

>No way is core of YT python you delusional fuck
Kek

en.wikipedia.org/wiki/YouTube
>Written in Python (core/API),[2] C (through CPython)

>Even Javascript is better than this shit
funny enough, tensorflow.js and numjs are a thing now too. more python-y are probably coming to browsers/node.js soon.

so it's written in C, gotcha
Thanks for proving my point

partly C Via Cython, also this is interesting. Looks like Google (at least in 2017) was looking at having Go run Python, anyway
opensource.googleblog.com/2017/01/grumpy-go-running-python.html

>Google runs millions of lines of Python code. The front-end server that drives youtube.com and YouTube’s APIs is primarily written in Python, and it serves millions of requests per second! YouTube’s front-end runs on CPython 2.7

the citation on the "Python" usage suggests they transpile it to Go, so I'm not sure it's accurate to say their usage is CPython-esqe, or even python.

dude if you know those langs python is easy as fuck to learn

>, also no machine learning meme projects pls
its actually quite easy using pandas and scikit-learn
import pandas as pd
from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_absolute_error
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeRegressor

def get_mae(max_leaf_nodes, train_X, val_X, train_y, val_y):
model = DecisionTreeRegressor(max_leaf_nodes=max_leaf_nodes, random_state=0)
model.fit(train_X, train_y)
preds_val = model.predict(val_X)
mae = mean_absolute_error(val_y, preds_val)
return(mae)

def getRFmae(forestSize, train_X, val_X, train_y, val_y):
rf_model = RandomForestRegressor(forestSize,random_state=1)
rf_model.fit(train_X, train_y)
rf_val_predictions = rf_model.predict(val_X)
return mean_absolute_error(rf_val_predictions, val_y)

data = pd.read_csv("csv/HomeInfo.csv")
y=data.Price
features=[u'Beds', u'Baths', u'Garage', u'LotSize', u'Quality',u'Sqft']
X=data[features]
train_X, val_X, train_y, val_y = train_test_split(X, y, random_state=1)

print "determining optimum leaf nodes"
for each in [10,25,50,75,100,125,150,175,200,225,250]:
print "%d : :%.2f" % (each,get_mae(each,train_X, val_X, train_y, val_y))

print "using random forest model"
for each in range(1,101,10):
print "%d : %.2f" % (each,getRFmae(each,train_X, val_X, train_y, val_y))

The citation is their to show that it's written in Python, it's true that Google were experimenting with transpiling to Go in order to get around Pythons concurrency problems. But that project hasn't had any commits to it Github since November last year. So I'm not sure if Grumpy was ever actually used at scale or if it stayed in the experimental phase.

The scipy stack is what makes Python really worth it. You could try learning yourself some actual AI and then making it with numpy, you don't have to include machine learning

>tfw new to programming and learning python
Is it really that bad?

no, it's not
All the hate comes from indian java babbies

Yes it is, do not listen to Python will irreparably fuck up your understanding of programming. It will take years to undo the damage it taught you. If you want to future proof your career, do not start with a language that is used in bootcamps to teach brainlet morons how make CRUD shit webapps or meme ML models

>irreparably fuck up your understanding of programming
In what way?

If not python then what, and in what way will it ruin your ability to comprehend more complex languages?

python programming in 90% of cases is gluing libraries together written by smarter people than you, usually actually just c libraries with a python wrapper

Doing this approach means you learn nothing yourself because you are just a monkey with a glue stick

Secondly python has poor support for parallelism and functional programming concepts, both of which are key to modern programming in a post-moore world.

Thirdly, python is dynamically typed, you won't learn some important OOP concepts with it.

It's also slow as fuck so software you write will be bad by default (unless you drop down to C for performance critical code, which you won't because you only know python)

Well I'm halfway through a chunky book on python so won't be stopping learning it but it's good to know what the flaws are.

>scipy stack
ML/DS stacks and pipelines exist in just about every major programming language now.

Python doesn't have the monopoly on that anymore, it's just the easiest for DIY programmers to pick up.

I love Python too!

Attached: 152710154418.gif (400x480, 124K)

>python programming in 90% of cases is gluing libraries together written by smarter people than you, usually actually just c libraries with a python wrapper

Good for productivity. Why would I re-write an efficient k-means algorithm if I can just use it for my task?

>Doing this approach means you learn nothing yourself because you are just a monkey with a glue stick

Sort of. But it also means than I can get an answer to a question within the hour, so I can move on to the next task; not needlessly spending ages on an implementation that might not even work.

>Secondly python has poor support for parallelism and functional programming concepts, both of which are key to modern programming in a post-moore world.

True, but who the fuck would use Python for parallelism purposes? It's like saying Python is a poor choice because it doesn't teach you how to drive a lorry properly.

>Thirdly, python is dynamically typed, you won't learn some important OOP concepts with it.
I'd argue that static types aren't the most important concept of OOP, Python does classes better than e.g. Julia.

Sure you'd be highly inefficient if you'd know Python and only Python, but surely that's not Python's fault. Like I'm not going to slag off the English language for the fact that most Brits know half a second language at best..

>I love Python
That 10+yo Zoomer

Attached: 10+yo Zoomer.png (820x324, 244K)

Companies rarely wants some data scientist with a bunch of fluffy ideas about how to build an algorithm and more someone who can build a microservice or autonomous product which can be sold or deployed. Python is the worst possible language for these kinds of applications.

>quick answer
Whatever you build as a POC will need to be productionised within a couple of months in a business setting (typically while juggling multiple other projects). Building something in Python might not be transferrable to a production language like C#, Java or anything else.

Often data scientist code needs to be rebuilt by production guys or wrapped somehow which is a job most people hate doing.

Realistically, what ML and Data Science needs is a language which doesn't require a deep knowledge of programming like R or Python but can be easily packed into a class, optimized and integrated within a production stack. Python doesn't accommodate that, ever.

Why would you choose a pure OO language and then bitch about it's lack of support for functional programming? That might be one of the most retarded things I've read in years.

Are you serious? It's a generator. It's supposed to work that way unless you explicitly tell it not to. This way you don't just load all the items into memory unless you need them.
g = range(10) # generator
l = list(range(10)) # list of ints

you're retarded my dude.

>got nothing to contribute
>post anyway
Why are you even in this thread?

>>Broken for loops
Use numba
>>shitty lambdas
Why are you doing FP is an OOP lang?
>>having to do multiprocessing to use multiple cores
numpy, pandas, tensorflow, etc. Bypass the GIL so this is a nonissue