jayvdb / piprot

How rotten are your requirements?

NOTE: this package is no longer maintained. You can use pip list --outdated or pipenv update --outdated to show outdated packages (with a little less whimsy).

piprot

How rotten are your requirements?

tl;dr: piprot allows you to check the requirements defined in your requirements files for freshness.

Installation

The latest release of piprot can be installed via PyPI:

pip install -U piprot

Basic Usage

Run piprot and provide a requirements file (if it's not called requirements.txt) and it will tell you the current status of your packages.

> piprot base_requirements.txt
requests (2.3.0) is up to date
six (1.6.1) is 107 days out of date. Latest is 1.7.3
piprot (0.6.0) is up to date
doge (3.4.0) is 129 days out of date. Latest is 3.5.0
Your requirements are 236 days out of date

If your requirements file is named "requirements.txt", you don't need to provide it. piprot will automatically traverse included requirements files.

The --verbatim argument will output your complete requirements file, with some comments about the out of date nature of your packages.

> piprot --verbatim
# Requirements for Piprot
# This actually doubles as a test file

requests==2.3.0
six==1.6.1 # Latest 1.7.3

piprot==0.6.0
# notarequirement==0.1.1

doge==3.4.0 # Latest 3.5.0
# Generated with piprot 0.7.0
# Your requirements are 236 days out of date

Using --outdated will show only the out of date requirements, pretty much the same as running pip list -o, except on a requirements file.

> piprot --outdated
six (1.6.1) is 107 days out of date. Latest is 1.7.3
doge (3.4.0) is 129 days out of date. Latest is 3.5.0
Your requirements are 236 days out of date

The --latest argument will output the requirements lines with the current version replaced with the latest version.

> piprot --latest
ipython (1.1.0) is 331 days out of date. Latest is 2.2.0
ipython==2.2.0 # Updated from 1.1.0
Django (1.5.4) is 241 days out of date. Latest is 1.6.5
Django==1.6.5 # Updated from 1.5.4
requests (1.2.3) is 356 days out of date. Latest is 2.3.0
requests==2.3.0 # Updated from 1.2.3
Your requirements are 928 days out of date

Personally, I like to use --latest and --verbatim together, creating a sort-of ''perfect'' requirements file for me,

> piprot --latest --verbatim
# Development Requirements
ipython==2.2.0 # Updated from 1.1.0

Django==1.6.5 # Updated from 1.5.4
requests==2.3.0 # Updated from 1.2.3
# Generated with piprot 0.8.0
# Your requirements are 928 days out of date

Yep, you can use stdin as well if you really want to, but there are better tools for working with packages installed in your environment.

pip freeze | piprot

(New in 0.9) You can also lookup requirements from a Github repo with the --github, --branch and --path options. Additionally you can use --token to supply a Personal Access Token to remotely test private repositories.

> piprot -g sesh/piprot
requests (2.4.2) is out of date. Latest is 2.4.3
requests-futures (0.9.5) is up to date
six (1.8.0) is up to date
piprot (0.8.2) is up to date
Looks like you've been keeping up to date, time for a delicious beverage!

You can also ignore packages using a norot comment in your requirements file.

# Inside requirements.txt
# Note: two spaces before # norot
Django==1.6.5  # norot

You can set up a delay before piprot throws an error. This is useful when you set up piprot in CI but cannot always upgrade the dependencies.

> piprot --delay 7
ipython (1.1.0) is 5 days out of date. Latest is 1.1.1
All of your dependencies are at most 7 days out of date.
# Displays a warning but does not throw an error

Working with your environment

piprot is designed around working with requirements defined in a requirements file. Check out pip-tools if you're looking for something similar that's designed for use against the packages that you actually have installed.

Tests

To run the test suite, execute python -m unittest discover, within the project directory.

Please ensure that the (limited) tests are all passing before making a pull request. Feel free to add more.