You're No Fun Anymore - PyCon AU 2014

Lightning talk presented to PyCon AU 2014 in Brisbane.

Transcript

1. You're no fun anymore PyCon AU 2014 David Beitey (@davidjb or @davidjb_) Slides @ https://speakerdeck.com/davidjb
3. And now for something completely different • Far too serious. • It's Sunday so let's be a little silly. • Ever wondered how much Monty Python influences Python?
4. Let's find out!
5. Reference you likely know: • Python's name is from Flying Circus ◦ Great post from Guido: http://python-history.blogspot.com. au/2009/01/personal-history-part-1-cwi.html • IDLE editor, Cheese Shop, Packages, Wheel • Sketch and comedian references in documentation, tests • Guido "under-analysed" the language's name ◦ "First thing that came to mind"
6. Let's undo all that!
7. Introducing Blancmange • Sweet, creamy dessert. Popular in UK in 70s. • Topic of much of Ep 7 of Flying Circus • Now a Monty Python analysis package for Python easy_install blancmange
8. Introducing Blancmange (console scripts) flying-circus-db Creates an SQLite DB using HTML original sketches and divides into Episodes/Sketches/Words/Actors – who said what. flying-circus Micro-analyse Flying Circus (165,483 spoken words, Michael Palin said 24.3% of them), 500+ sketches, 10100 lines spoken, 249,945 words in scripts) spot-camels keyword [keyword...] Find & display sketches that feature keywords. Open in web browser. completely-different Lookup a random line and output as Python comments, wrapped, and formatted ready for inclusion in your code. blancmange [path] Process a directory of files (eg CPython!) and match with the scripts. Output most common references, those which are missing, and actor stats.
9. Run blancmange over CPython • Naive text analysis of arbitrary source is hard! • Extract episodes from downloaded data wget/PyQuery • Extract words and counts from episodes TextBlob/NLTK • Exclude short words, Python syntax Pygments • Normalise using NLTK. Brown Corpus is from 1979 which is perfect. • Wait about 45 mins. ◦ 76MB of CPython, 250,000 Flying-circus words ◦ Results are pickled for sanity.
10. tl;dr
11. Fast facts • Flying Circus has 76 references to Spam. Cpython has 2225! • Eric Idle gets mentioned 3495 times in Cpython. IDLE, libidle John Cleese mentioned 316 times, Terry Gilliam just 8 times. Aww :( • Almost all references are within the documentation/tests: a. May be good thing or we'd end up with: os. walk → silly.walk • Missing references to: Camel Spotting, Fish Slapping, Hell's Grannies, Blancmange, etc.
12. My Zen of (Monty) Python Monty Python references are one honking great idea -- let's do more of those! • Blancmange = rough edges. < 2 days old. • Want to help or think I'm crazy? Get in touch! • http://github.com/blancmange/blancmange • http://davidjb.com, Twitter: @davidjb_ Slides @ https://speakerdeck.com/davidjb