How To Handle Python Dependencies Throughout The Project?

Lets say a developer is working on a project when he realizes he needs to use some package. He uses pip to install it. Now, after installing it, would a the developer write it down

Solution 1:

The command:

pip freeze > requirements.txt 

will copy all of the dependencies currently in your python environment into requirements.txt. http://pip.readthedocs.org/en/latest/reference/pip_freeze.html

Solution 2:

It depends on the project.

If you're working on a library, you'll want to put your dependencies in setup.py so that if you're putting the library on PyPi, people will be able to install it, and its dependencies automatically.

If you're working on an application in Python (possibly web application), a requirements.txt file will be easier for deploying. You can copy all your code to where you need it, set up a virtual environment with virtualenv or pyvenv, and then do pip install -r requirements.txt. (You should be doing this for development as well so that you don't have a mess of libraries globally).

It's certainly easier to write the packages you're installing to your requirements.txt as soon as you've installed them than trying to figure out which ones you need at the end. What I do so that I never forget is I write the packages to the file first and then install with pip install -r.

pip freeze helps if you've forgotten what you've installed, but you should always read the file it created to make sure that you actually need everything that's in there. If you're using virtualenv it'll give better results than if you're installing all packages globally.