tl;dr: Please use
-undefined dynamic_lookup instead of
-framework Python to build Python extension modules on OS X, no matter what
python-config says. Only use
-framework Python if you intend to embed an interpreter.
Python 3.5 is scheduled to be released on September 13, 2015. Since Homebrew carries the newest stable version of its packages, the Homebrew
python3 formula will adopt Python 3.5 very quickly upon its release. This post explains how migrating to Python 3.5 will affect Homebrew users who have set up a Python 3 development environment using Homebrew packages.
What would you do with 2 petaflops of processing power? It’s not an idle question: NSF’s XSEDE infrastructure offers researchers access to large-scale computing resources. I spent the last week at the San Diego Supercomputer Center, which hosts an annual Summer Institute to introduce scientists from many disciplines to high-performance computing technologies. Tissue engineering doesn’t have a lot of petascale problems but I did learn a lot about tools that can help me write more performant code on single and multi-processor systems—even in Python, my favorite programming language. ~And~ I got to take a selfie with a supercomputer.
I needed an accurate and inexpensive temperature logger to calibrate our stage incubator for live cell microscopy and make sure that it was stable over time, so I built one. This was a very quick project, using a $30 Dataq DI-145 data logger and a $4 epoxy-coated thermistor from Adafruit. (To power the thermistor from USB, which is safe and easy, add a micro-USB breakout board, a matching USB cable, and a small solderless breadboard for the complete package!)
Here are the incantations you need to perform to create a Cairo surface from a grayscale image in a Numpy array, doodle on it, and convert the doodled-upon image back into a Pillow image so you can save it as e.g. a JPEG.