`requirements.txt` dependencies, getting only high level dependencies

Question

I have a three python projects A, B and C. Each depending on each other.

How can I now "clean up" my requirements.txt that only the high level dependencies that are required at this "level" are there.

Example

boto3==1.2.4

A==0.0.1
boto3==1.2.4
cookiecutter==1.4.0

B==0.0.1
slacker==0.9.9
boto3==1.2.4
cookiecutter==1.4.0

The bold packages are the ones which should be filtered out.

Why do you want to do this? If a package is already installed it will be skipped by pip anyway. — Selcuk, Jun 22 '16 at 11:40
If I refactor stuff I lose track what is in which package. I want to automate cleaning up my requirements.txt and don'T see duplicates as a good code quality. — lony, Jun 22 '16 at 11:43

score 8 · Accepted Answer · answered Jun 22 '16 at 11:53

8

pipdeptree can display the dependencies tree, and for a given package tell you which other package(s) depends on it.

answered Jun 22 '16 at 11:53

1

Specifically: https://github.com/naiquevin/pipdeptree#using-pipdeptree-to-write-requirementstxt-file – jonrsharpe Jun 22 '16 at 12:16
2

Then, you can use grep to get only top level dependencies: `pipdeptree | grep -e '^\S'` – moppag Jan 08 '19 at 15:13