Move from Excel to Python with Pandas Transcripts
Chapter: Appendix: Python language concepts
Lecture: Concept: Virtual environments
0:01 One of the challenges of installing packages globally has to do with the versioning.
0:05 The other really has to do with managing deployments and dependencies.
0:09 Let's talk about the versioning part first.
0:11 Suppose my web application I am working on right now
0:14 requires version 2.9 of requests.
0:18 But somebody else's project required an older version with older behavior,
0:22 version 2.6 let's say. I don't think those are actually incompatible,
0:25 but let's just imagine that they were.
0:28 How would I install via pip version 2.6 and version 2.9 and keep juggling those,
0:33 how would I run those two applications on my machine without continually reconfiguring it-
0:37 the answer is virtual environments.
0:39 And, virtual environments are built into Python 3
0:43 and are also available through a virtual env package
0:46 that you can install for Python 2
0:48 and the idea is this- we can crate basically a copy,
0:52 change our paths and things like that around so that when,
0:55 you ask for Python or this looks for Python packages,
0:58 it looks in this little local environment, we create one of these small environments
1:02 just for a given application, so we would create one for our web app
1:06 that uses request 2.9 and another one for the one that uses request 2.6
1:10 and we would just activate those two depending on which project we are trying to run,
1:14 and they would live happily side by side.
1:17 The other challenge you can run into is if you look
1:19 at what you have installed on your machine,
1:21 and you run some Python application and it works,
1:24 how do you know what listed in your environment is actually required to run your app,
1:28 if you need to deploy it or you need to give it to someone else,
1:31 that could be very challenging. So with virtual environments
1:34 we can install just the things a given application requires to run
1:37 and be deployed so when we do something like "pip list",
1:41 it will actually show us exactly what we need to set up
1:44 and use for our app to run in production.
1:47 Typically we tie virtual environments one to one to a given application.
1:51 So how do we create one?
1:53 This example uses virtual env which we would have to install via pip,
1:57 you could also use venv, just change virtual env to venv in Python 3
2:01 and it will have the same effect, but this works,
2:03 like I said in Python 2 and 3, so here you go.
2:06 So we are going to run Python 3 and we are going to say run the module, virtual env,
2:09 and create a new environment into ./localenv.
2:14 Here you can see it creates a copy from Python 3.5.
2:17 Then we go into that environment, there is a bin directory
2:20 and there is an activate program that we can run and notice,
2:23 we'll use the . (dot) to apply that to this shell
2:26 and not to create a new separate shell environment for that when it runs
2:30 because we wanted to modify our shell environment, not a temporary one.
2:34 So we say . activate and that will actually change our environment,
2:38 you can see the prompt change, if we say "pip", we get the local pip,
2:41 if we ask "which python", you'll see it's this one that is in my user profile
2:45 not the one in the system.
2:47 Now, few changes for Windows, if I did exactly the same thing in Windows,
2:51 I would have .\localenv of course, I might not use Python 3,
2:56 I just say Python and make sure I have the right path to Python 3
2:59 because that is not a feature in the Python 3 that comes on Windows,
3:03 and I wouldn't use the source activate you don't need to do that in Windows,
3:06 but you would call activate.bat, otherwise,
3:10 it's pretty much the same. Also, the "which" command doesn't exist on Windows,
3:13 use "where" and it gives you the same functionality.
3:16 So we can create one of these virtual environments in the terminal,
3:18 but you might suspect that PyCharm has something for us as well,
3:21 and PyCharm actually has the ability to create
3:24 and manage virtual environments for us,
3:26 basically it does what you just saw on the screen there.
3:30 So here we give it a name, we give it a location here,
3:32 we say blue_yellow_python, this is going to be for a Blue / Yellow band web application,
3:37 we are going to base this on Python 3.5.1
3:39 and we are going to put into my Python environments and under this name.
3:43 Then I just say OK and boom, off it goes, set it as the active interpreter
3:46 and manage it just the same as before in PyCharm
3:49 using its ability to install packages and see what is listed and so on.