#58 Better cache decorators and another take on type hints

Python Bytes - Podcast tekijän mukaan Michael Kennedy and Brian Okken - Maanantaisin

Kategoriat:

Sponsored by DigitalOcean: http://do.co/python

Brian #1: Instagram open sources MonkeyType

  • Carl Meyer, an engineer on Instagram’s infrastructure team.
  • (Note: we talked about Dropbox’s pyannotate in episode 54. pyannotate is not on Python3 yet and generates comment style annotations that are Py2 compatible)
  • MonkeyType is Instagram’s tool for automatically adding type annotations to your Python 3 code via runtime tracing of types seen.
  • Requires Python 3.6+
  • Generates only Python 3 style type annotations (no type comments)

Michael #2: cachetools

  • Extensible memoizing collections and decorators
  • Think variants of Python 3 Standard Library @lru_cache function decorator
  • Caching types:
    • cachetools.Cache Mutable mapping to serve as a simple cache or cache base class.
    • cachetools.LFUCache Least Frequently Used (LFU) cache implementation
    • cachetools.LRUCache Least Recently Used (LRU) cache implementation
    • cachetools.TTLCache LRU Cache implementation with per-item time-to-live (TTL) value.
    • And more
  • Memoizing decorators
    • cachetools.cached Decorator to wrap a function with a memoizing callable that saves results in a cache.
    • Note that cache need not be an instance of the cache implementations provided by the cachetools module. cached() will work with any mutable mapping type, including plain dict and weakref.WeakValueDictionary.
    • Can pass key function for hash insertions and lock object for thread safety.

Brian #3: Going Fast with SQLite and Python

  • Charles Leifer
  • Many projects start with SQLite, as it’s distributed with Python as sqlite3.
  • This article discusses some ways to achieve better performance from SQLite and shares some tricks.
    • transactions, concurrency, and autocommit
    • user-defined functions
    • using pragmas
    • compilation flags

Michael #4: The graphing calculator that makes learning math easier.

  • A full graphing calculator
  • Programmable in Python
  • Exam approved: Take the SAT and the ACT.
  • Free browser emulator

Brian #5: Installing Python Packages from a Jupyter Notebook

  • Jake VanderPlas
  • using conda import sys !conda install --yes --prefix {sys.prefix} numpy
  • using pip import sys {sys.executable} -m pip install numpy
  • plus a discussion of why this is weird in Jupyter

Michael #6: Videos from PyConDE 2017 are online

  • via Miroslav Šedivý @eumiro
  • Lots of interesting talk titles
  • Almost all in English

Visit the podcast's native language site