RStudio equivalent for Python

August 22, 2012 7:44 PM

What is the closest equivalent to RStudio for Python?

I love RStudio - the integrated help files, integrated package searching and downloading, the extra little features like Knitr and RMarkdown, and Git integration. I've pretty much given up on Python and have been using R for years now, because I could never find an IDE for Python that did all the things I wanted.

So what's the closest thing to RStudio for Python development? In particular, I want something with integrated help, auto-complete for functions and arguments, the ability to browse, download and install Python libraries from a central repository, and the ability to manage "projects", with a collection of associated files in a directory

posted by Jimbob to Computers & Internet (11 answers total)

Guido likes PyCharm (and so do I - I spend 8 hours a day in it, at least).
posted by dmd at 7:46 PM on August 22, 2012

(and it's come a LONG way since that 2-year-old post was written)
posted by dmd at 7:47 PM on August 22, 2012

I also love PyCharm. Also PHPStorm by the same company. I spend a ridiculous amount of time in those two apps. My only significant complaint is the gig of memory it can consume, but thats a drawback of any full-featured IDE. I don't think it can install libraries though (although I've never actually checked); for that i use pip.
posted by cgg at 8:02 PM on August 22, 2012

Response by poster: Just checking out PyCharm at the moment. It does look like it can install packages!However, it's also on a 30-day limited trial. I should have said, I prefer free/open source.
posted by Jimbob at 8:06 PM on August 22, 2012

Like you, I've tried to find a Python counterpart of RStudio and I finally settled on Python (x,y), which contains the Spyder IDE. I'm not sure about installing Python libraries with it, but it's free.
posted by Bokmakierie at 8:17 PM on August 22, 2012

Response by poster: Okay, PyCharm review: Tried to set up a new project, selected a freshly installed Python 2.7.3 as the interpreter. Tried to get PyCharm to install some packages - I started with pandas, but it complained that numpy wasn't installed. So, the packaging system doesn't take into account dependencies. Okay. So then I tried to install numpy, which is pretty standard - PyCharm threw up a window filled with errors as to why numpy couldn't be installed (missing libraries, missing batch files, repositories not found). Then PyCharm froze and had to be killed.

Starting to remember the perverted handing of packages is why I quit Python and switched to R...
posted by Jimbob at 8:22 PM on August 22, 2012

The general goodness in Python package handling is to use pip on the command line--that's where "it just works" happens. I've never seen an IDE that handles this smoothly.

That said, I've worked in a variety of IDEs with Python, and the two big ones are PyCharm and PyDev in Eclipse (or in Aptana Studio). Without expecting them to do package handling, I've been quite satisfied with them--but part of that satisfaction derives from me doing webapps in Django, so the IDE handling the full stack of HTML, Django templates, Python, and git is what made me happy.
posted by fatbird at 9:48 PM on August 22, 2012

I'm assuming you're on a Windows machine? Either way, for package management and Python, you really need to get familiar with the big guys, all command-line based: pip or easy_install. I perfer pip because it's easier to uninstall packages.

For a free IDE, I'd try Aptana. It was decent given the price when i tried it a couple years ago.
posted by cgg at 9:51 PM on August 22, 2012 [1 favorite]

Response by poster: Should say that I'm using both Windows and OSX, but would prefer OSX. Once again, since RStudio is so completely platform-agnostic and works the same on Windows, OSX and on it's Linux server version, pointing out which OS I'm on slipped my mind...
posted by Jimbob at 11:53 PM on August 22, 2012

Just started using Python. Pip for package management and installing, Spyder for Rstudio-like IDE, pandas and numpy for R-like data handling, matplotlib for graphics. The Enthought Python distribution is free for academics and has a lot of useful packages integrated.

Packages are more complex in python. I agree it's harder than RStudio. I figure this is because python has to be a lot more flexible and can do a lot more things, being a general-purpose language. Pip is quick once you get rolling with it.
posted by PercussivePaul at 12:08 AM on August 23, 2012 [1 favorite]

Nthing PyCharm and use pip or easy_install. Nothing else I've tried (Aptana, PyDev, Spyder) has come close to the level of refactoring, auto-completion, and code analysis that it supports. It's well worth the money and seems to go on 50% sales pretty often.
posted by ReadEvalPost at 7:35 AM on August 23, 2012