Favorite Development Environment Setup

[shaunwbell]

Whats your favorite working environment?

I tend to use Sublime Text 3 as my text editor. Its cross platform and extensible so I can have the same environment on any machine I use. It does code highlighting and my favorite feature...multi-selection editing. It also has a builtin python console but I've not used it.

For interactive development, I tend to use Spyder3 (packaged with the conda distribution) which sits on top of Ipython and provides a matlab like experience with a text editor (with code analysis) and array/variable viewing.

[MCasari-PMEL]

I also like Spyder3 as an IDE. The debugging capability is good, and like Shaun mentioned it can be set up to look like Matlab, if you like that sort of thing.

I have been using Atom a lot lately. It is Github's cross-platfrom text editor with a ton of plug-ins, including Python console and auto-complete. It also has direct plugins to git/github and will do Markdown Preview. Similar to Sublime, but is open-source so no need to pay for continued use.

At home I've also been playing around with Microsoft Visual Studio Code, a source code editor which is also open source and cross-platform (yes, it will work on macOS & Linux). It has debugging capability, including C and Python (with a plug-in). I've actually been impressed with it, but haven't started the move into professional dev with it yet.

[crdietrich-noaa]

I use a combo of Spyder, PyCharm and Jupyter: Spyder for data driven dev, support for 2d dataframes in the variable explorer. It's also nice like Jupyter to execute code snippets in an existing iPython session. It's faster to kill the interpreter and start a new session than Jupyter. PyCharm - any package or threaded program development. I try to keep my print() debugging to a minimum here and use the debug tools. Code inspections and formatting are far superior and definitely make the code better (PEP 8/Docstrings etc). Guido uses it so it's got that going for it :) Jupyter for per dataset development - Import local custom packages and edit them in PyCharm as I work with the data. Jupyter has more granular plotting control (%inline) so it's easier to replot a lot at once and ensure a sequence. Formatting and code completion is ok but printing DataFrame slices or multiindexes are easier to read in Jupyter imo. Also Jupyter uses a lot of Javascript, I try not to run other javascript heavy applications in the same browser, if the browser crashes you'll lose edits. Sidenote: I've found Atom to be very slow (Javascript again), however it's markdown rendering is great and I use it for documentation.

On Thu, May 18, 2017 at 9:09 AM, Matt Casari notifications@github.com wrote:

I also like Spyder3 as an IDE. The debugging capability is good, and like Shaun mentioned it can be set up to look like Matlab, if you like that sort of thing.

I have been using Atom a lot lately. It is Github's cross-platfrom text editor with a ton of plug-ins, including Python console and auto-complete. It also has direct plugins to git/github and will do Markdown Preview. Similar to Sublime, but is open-source so no need to pay for continued use.

At home I've also been playing around with Microsoft Visual Studio Code, a source code editor which is also open source and cross-platform (yes, it will work on macOS & Linux). It has debugging capability, including C and Python (with a plug-in). I've actually been impressed with it, but haven't started the move into professional dev with it yet.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/NOAA-PMEL/Python_Users_Group/issues/3#issuecomment-302455615, or mute the thread https://github.com/notifications/unsubscribe-auth/APuxXy8c-NShChimWZdhFX9V4cocfwqBks5r7G1JgaJpZM4Nfb8I .

-- Colin Dietrich Research Scientist Carbon Program | Pacific Marine Environmental Laboratory NOAA-UW Joint Institute for the Study of the Atmosphere and Ocean 206-526-6195 | Alt: colinrd@uw.edu