Mailman 3 - Edu-sig - python.org

python in the movies...
by kirby urner 16 Jun '20

16 Jun '20

I've taken a deep dive into Blender recently, you'd have thought earlier maybe, but I was slow into CAD, doing my work directly in a ray tracer (POV-Ray, also free open source). https://nbviewer.jupyter.org/github/4dsolutions/School_of_Tomorrow/blob/mas… Why not jump right into something full featured even for plotting XY graphs? Keep doing those parabolas but add texture / materials, lighting. Math as theater (a kind of art). Sure, we can do that in matlabish panda type environments but lets not forget live ray tracers with fully internalized live Python (like ESRI). Just import bpy and you have full scripting control over the GUI interface as well as hooks into all the data, the current context and on and on. The ergonomics drive the economics i.e. if every student is going to need hardware up to Blender's capabilities, with GPU(s), then we're beyond "one laptop per child" to "one (or more) well equipped workspace(s) per child" (adults need PWSs too, and many have them). The idea that school is cramped little desks close together with no room for multiple monitors is so 1900s. Kirby

2 1

Re: back to school... "1/3" --> float question, just eval?
by kirby urner 11 Apr '20

11 Apr '20

I decided to try replying through the HyperKitty archived interface, where I’ve been examining the cosmetics of [our thread](https://mail.python.org/archives/list/edu-sig@python.org/thread/H6K… <https://mail.python.org/archives/list/edu-sig@python.org/thread/H6KKJR57V7G…>) to date with a critical eye. Also, I see I said “construct manager” a few times when I was meaning “context manager” — oh well. Hyperkitty pops me into an Apple mail client I never use. Some default Apple setting somewhere I suppose I could dig up. Lets see if this sends or will I end up cutting and pasting to Gmail in Chrome. Yes Wes, you’re pretty close on your educated guess/memory re that semi-esoteric species of class (the context manager), wherein ```__enter__``` is what actually returns the object of the optional “as” as in: with DB(“users”) as db: # <— optional object returned by ```__enter__``` which takes no args indented block ```__exit__``` triggered on exit, has info in args re any exceptions, I use *oops to gather up the 3-tuple. outdented (db stays available but __exit__ will have likely closed the connection, e.g. hung up the phone on the database) The arguments to class DB() go to the optional ```__init__``` as always, however ```__enter__``` gets to return the “db” (above) which then becomes available to the indented block (suite as Holden calls it) as a potentially useful object. Most typically ```__enter__``` simply returns the self (the instance of the context manager class itself). However in principle it’s free to return any object it likes, similar to ```__iter__``` (which should at least return an iterator, but only self if self is already such, per class generator API where ```__next__``` must also be there). [My pattern](https://github.com/4dsolutions/python_camp/blob/master/ContextMana… <https://github.com/4dsolutions/python_camp/blob/master/ContextManager.ipynb>) is to load up the target object db with a lot of tools for working with the database at hand, much as open(filename) may be use with with to create the filetype object. On ```__exit__``` the database closes. In between, I might even db.zap_tables() and db.create_tables(). Sometimes my DB is very customized to work with a very specific database. Other times I separate the ```__enter__``` ```__exit__``` part, could be a parent helper. The linked Notebook gives more context. Yes, completely agree on Jake Vanderplas as a go to guy for anything both pythonic and data scientific. From my campsite FAQ: Q: Where's a good place to begin with data science? A: Follow Jake Vanderplas <http://vanderplas.com/> and study his work. Wes, your pull requests came in smoothly and my current plan is to exhibit this program as is, as an example of what [super duper fluent Python](https://github.com/4dsolutions/python_camp/blob/master/camper_progr… <https://github.com/4dsolutions/python_camp/blob/master/camper_program.py>) actually looks like, with unit testing and the whole business. I might use git to dig up an old version (say from this last Monday), showing how ancient drafts might still be excavated in a repo, and then roll forward again, to the way it is now. That’s what git’s good for kids! I did the pull request process online for campers to observe, and commented on closing the request, how smoothly it all went. I’m hoping my practice of outlining a course in a hyperlink-rich Jupyter Notebook will continue to spread, not that I invented it or anything. I’m telling my campers this generic Notebook technology (not just Jupyter) is what their high schools might well be using soon, if not already. I’m eager to pack in a lot of serious academic content given these ["covid camps”](https://flic.kr/p/2iLG2qU <https://flic.kr/p/2iLG2qU>) aren’t just “summer camp”. They’re a continuation of their ordinary schooling, in a homeschooling context. Kirby

2 2

back to school... "1/3" --> float question, just eval?
by kirby urner 10 Apr '20

10 Apr '20

Here's a thread I'm starting with my campers taking a PyCamp course. Yesterday we were co-evolving a calculator driven by looping menu. It does squares and square roots (one of the campers supplied this logic) and I was just getting around to introducing the y^x key as some call it, i.e. any number to any power, Python's pow(num, exp). Today we'll pick up where we left off. We used the shared whiteboard to draw radical signs as I intoned about how pow(3, 3) means like ```3 * 3 * 3```. Math Adventures in Python. We might also say ```3 ** 3``` (and not ```3^3``` like in so many languages). sqrt(num) is the same as ```num ** (1/2)``` and we can get the 3rd root of 27 with ```27 ** (1/3)```. [ Sorry for all the markdown by the way, if you're getting a plaintext version -- an option I'd advocate for browser based viewers of the archives someday (to see what it looked like way back when, Before Markdown). ] So here's what we're up against: say I want to raise 10 to the 1/3rd power, same as taking a "cube root" of 10 (tetrahedral whatever). The menu prompts: ```python num = float(input("Your number please: > ")) # yes a bit dangerous exp = float(input("...and your exponent: > ")) # user wants to do 1/3 ``` Of course that last statement raises an exception because float("1/3") doesn't work (isn't supposed to). What shall we do then? I'm thinking we parse anything that looks like a legal fraction and do the floating point conversion from int(p)/int(q). We won't allow input like 2.1/3.6 just int/int. I've already preached against eval() but if it passes my regular expression test first... Anyway, it's a discussion. These are like middle schoolers previewing high school, at a time when kid focused code schools were under some pressure to convert to "everything online". I used to drive to the schools. We don't have to use a regexp! A good excuse to say what these are though. Or test for the pattern but then use split("/") to pry numerator from denominator, once we know that'll work? https://github.com/4dsolutions/python_camp/blob/master/camper_program.py (snapshot -- part of the camp is they watch me git pushing updates to the camp repo). Rather perversely, I may set things back between camps i.e. I'll deliberately revert the code. Comments? Code? Kirby PS: I notice Python 3.8 is being rather more generous in its statistics module. Campers who wander from the base camp fire are likely to find themselves staring at pdfs and cdfs.

3 8

A Quantum 'Hello World'
by Wes Turner 10 Apr '20

10 Apr '20

"Ask HN: What's the Equivalent of 'Hello, World' for a Quantum Computer?" https://news.ycombinator.com/item?id=22707580 https://westurner.github.io/hnlog/#story-22707580 "Getting Started with Qiskit" > https://qiskit.org/documentation/getting_started.html > > Qiskit / qiskit-community-tutorials > "1 Hello, Quantum World with Qiskit" > https://github.com/Qiskit/qiskit-community-tutorials#1-hello-quantum-world-… > > "Quantum basics with Q#" > https://docs.microsoft.com/en-us/quantum/quickstart?tabs=tabid-python > > Qutip notebooks: https://github.com/qutip/qutip#run-notebooks-online > > Jupyter Notebooks labeled quantum-computing: > https://github.com/topics/quantum-computing?l=jupyter+notebook > Quantum computing: https://en.wikipedia.org/wiki/Quantum_computing Quantum algorithm: https://en.wikipedia.org/wiki/Quantum_algorithm Quantum Algorithm Zoo: https://quantumalgorithmzoo.org "Quantum Computing for Computer Scientists" https://www.youtube.com/watch?v=F_Riqjdh2oM - TIL about classical and quantum logic gates in terms of linear algebra, matrices, and tensors.

1 0

COVID-19 Resources, Curricula, Prompts
by Wes Turner 28 Mar '20

28 Mar '20

A thread to share [collections of] resources, curriculum ideas, etc. about and for during the COVID-19 epidemic Lots of analyses, some data, some helpful contributions, lots of people learning about exponential growth One video I saw mentioned that a person normal flu infects about 1.3-1.4 other people, but COVID-19 is closer to 3; so what's wrong with this analysis? 1**1.3 1**3 1.3**n 3**n '{:,}'.format(7e9) 1*(3**x) = 7e9 # solve for x with logarithms # Where is the limit with which controls? ## Notebook idea Growth curves: polynomials of degree 0 through 10 ('desic'), exponential, logistic - Exponential growth and epidemics https://youtu.be/Kas0tIxDvrg ## Prompt re: positive, helpful, constructive tone; morale; and amateur data science Here's a prompt for students and teachers alike: Respond to this re: amateur data science, tone, attitude, responsibility: https://www.reddit.com/r/datascience/comments/fm17ja/to_all_data_scientists… ```quote FWIU, there are many unquantified variables: - pre-existing conditions (impossible to factor in without having access to electronic health records; such as those volunteered as part of the Precision Medicine initiative) - policy response - population density - number of hospital beds per capita - number of ventilators per capita - production rate of masks per capita - medical equipment intellectual property right liabilities per territory - treatment protocols - sanitation protocols So, it **is** useful to learn to model exponential growth that's actually logistic due to e.g. herd immunity, hours of sunlight (UVC), effective containment policies. Analyses that compare various qualitative and quantitative aspects of government and community responses and subsequent growth curves should be commended, recognized, and encouraged to continue trying to better predict potential costs. (You can tag epidemiology tools with e.g. "epidemiology" https://github.com/topics/epidemiology ) Are these unqualified resources better spent on other efforts like staying at home and learning data science; rather than asserting superiority over and inadequacy of others? Inclusion criteria for meta-analyses. - "Call to Action to the Tech Community on New Machine Readable COVID-19 Dataset" (March 16, 2020) https://www.whitehouse.gov/briefings-statements/call-action-tech-community-… > “We need to come together as companies, governments, and scientists and work to bring our best technologies to bear across biomedicine, epidemiology, AI, and other sciences. The COVID-19 literature resource and challenge will stimulate efforts that can accelerate the path to solutions on COVID-19.” - https://www.kaggle.com/tags/covid19 - "COVID-19 Open Research Dataset Challenge (CORD-19): An AI challenge with AI2, CZI, MSR, Georgetown, NIH & The White House" https://www.kaggle.com/allen-institute-for-ai/CORD-19-research-challenge - https://en.wikipedia.org/wiki/Precision_medicine#Precision_Medicine_Initiat… ``` ## NIH FigShare instance https://www.niaid.nih.gov/news-events/rapidly-share-discover-and-cite-covid… > NIH is assessing the role of a generalist repository for NIH-funded research and has launched the NIH Figshare instance, a pilot project with the generalist repository Figshare You can archive a tag of a [topic-labeled] GitHub repository [containing notebooks] with FigShare. ## Resource Collections https://github.com/topics/2019-ncov https://github.com/topics/covid-19 https://github.com/topics/epidemiology?l=python https://github.com/topics/epidemiology?l=jupyter+notebook Objectively-scored Kaggle competitions: https://www.kaggle.com/tags/covid19 https://github.com/soroushchehresa/awesome-coronavirus On Tue, Mar 24, 2020, 6:35 AM kirby urner <kirby.urner(a)gmail.com> wrote: > Awesome! > > If you do any kind of Youtube on this specific SIR model I hope you'll > link it from the cell and share it here. > > I see some Youtubes like that already (SIR models, including in Python), > but everyone codes a little differently. > > I'd like to go back to high school and do it all again from a student > perspective, now that the curriculum and tools are so vastly different. > > High school should be for any age, and one keeps going back every 10 years > or so. Learn it the new way. > > Kirby > _______________________________________________ > Edu-sig mailing list -- edu-sig(a)python.org > To unsubscribe send an email to edu-sig-leave(a)python.org > https://mail.python.org/mailman3/lists/edu-sig.python.org/ >

1 4

My SIR Model
by Jorge Garcia 24 Mar '20

24 Mar '20

Here's my mostly pyhthonic model, actually it's done in sagecell.sagemath.org. Problem is, how do you predict the parameters alpha, probability if infection, and beta, probability of recovery. My simulation us interactive in that you can change the values of alpha and beta. A higher alpha ir lower beta with skew the infected curve so it peaks higher and sooner. So how do we reduce alpha? We use vaccines we don't have. How do we increase beta? We use antivirals we don't have. All we can do now us stay home and stay apart to reduce alpha and flatten the infected curve! Stay Healthy, A. Jorge García https://sagecell.sagemath.org/?z=eJytlEtv2zAMgO_-FUSKwXbitE67YVgXr4cdBt8K-9…

3 2

Learning Code IRC for kids?
by Jason Blum 20 Mar '20

20 Mar '20

Hello! I've been toying with the idea of recruiting mentors to commit to one hour a week to moderate a channel on https://gitter.im/ or https://discordapp.com/ or even IRC, specifically geared towards supporting kids working through https://codecademy.com/, https://trinket.io/ or https://twilio.com/quest or any of dozens of other amazing coding tutorials out there. With so many kids home now, seems like the perfect time kick something off. But I wanted to bounce the idea off folks. Does it already exist? Would you be able to participate? I envision a directory listing hour by hour who is moderating the hour and tagging any specializations (#Python, #Repl.it, #Arduino, etc.), but with the understanding that all general tech support is provided, even if only to steer questions towards the right resource on the web. Same IRC etiquette rules would apply. Sample interactions: kid: How do I get python installed on my dad's laptop: moderator: Hi #kid well I assume your dad is ok with this? I find https://installpython3.com/ pretty well maintained. kid: I'm stuck on this line of code [then pastes in 50 lines of code] moderator: Hi @kid ah cool, Ill try to point you in the right direction, but first lemme introduce you to https://dpaste.org/ kid: How do I [whatever]? moderator: Hi @kid! I dunno, but I've been doing this for 20 years and have yet to encounter a problem someone else didn't solve. Lemme help you research it a little. Let's start with [StackOverflow, PyVideo, Google, ReadTheDocs, DjangoPackages, PyPi, etc] Bottom line is just to keep younger coders from giving up, by giving them an introduction to how vast and awesome this community is. This would not be just another coding site and not open-ended tech support, but rather a chat-based gateway to what's already out there. Does such a thing already exist? Would you sign up one hour a week to mentor? How do we eject violations of https://www.python.org/psf/conduct/ Where does this idea fall apart? Thanks! -Jason Blum Father of four at home right now driving me nuts

3 3

REPL.IT works great for sharing Python examples
by kirby urner 10 Mar '20

10 Mar '20

Per usual, we're awash in useful tools, many free, but so little time to learn to use them, it seems. And they all keep evolving. Along those lines: REPL.IT, favored by many a Python teacher, yet so little explored by me. Like with LinkedIn and so many of these Web 2.0 tools, we get so many cool features. Make a Python program public, using 3.8.1 (that's very current now in March 2020), runnable code, blog embeddable. Use a Share feature to create a public Notification. Examples, two of mine from today (tested to work fine on Facebook): https://repl.it/talk/share/Python-Decorators-for-Composing-Functions/30054 https://repl.it/talk/share/Computing-Volumes-in-Tetravolumes-S-and-E-Module… Remember to use Markdown to markup these Notifications (which you can re-edit), same as here on edu-sig where Markdown is the new normal, right? On the first one, I used an embedded asterisk a lot, for Python multiplication, only to discover what I learned about the new edu-sig: Markdown is your friend. Kirby PS: shout out to Jake Vanderplas, the Pycon keynote speaker in Portland that time, who gets my Stupendous Python Teacher Award. @jakevp <https://pycon.blogspot.com/2020/03/march-2-update-on-covid-19.html>

3 2

#1 new book in Children's Computer Programming
by Kent Tong 25 Feb '20

25 Feb '20

Hi, I am glad that my book is currently the #1 new book in the Children's Computer Programming category on Amazon. Please take a look ( https://www.amazon.com/Yes-Kids-can-learn-Python-ebook/dp/B084CY2L43/ref=sr…) to see if you may make use of it at your school. I will be more than happy to provide any assistance that you need :-) [image: image.png] -- Kent Tong IT author and consultant, child education coach

2 2

Python 3.8 breaks backward compatibility with math.factorial(Decimal(int))
by kirby urner 18 Feb '20

18 Feb '20

Now you'll need to use math.factorial(int(Decimal)) or write your own factorial function. The storyline: Apropos of recent threads experimenting with high precision (arbitrary precision) numbers as a way to promote interest in "pure math" applications using Python, I discovered this morning that my Ramanujan Convergence script on repl.it <https://repl.it/@kurner/computepi> was broken all of a sudden (after working before), and set out to discover why. Answer: in Python 3.8, to which repl.it has newly upgraded, math.factorial no longer accepts Decimal type numbers, even if they're integral (integers). You can read the discussion here: https://bugs.python.org/issue33083 Here's the line of code that broke, and was fixed with explicit castings to int. term = (fact(int(c1*i))*(c2 + c3*i))/(pow(fact(int(i)),4)*pow(c4,4*i)) By the time of the final division, the numerator and denominator have been coerced back into Decimals and so will divide with the expected precision (set by the context): Where: ``` c1 = Decimal(4) c2 = Decimal(1103) c3 = Decimal(26390) c4 = Decimal(396) c5 = Decimal(9801) ``` For more context: https://github.com/4dsolutions/Python5/blob/master/Pi%20Day%20Fun.ipynb (I'm updating it now...) Kirby

1 0