Light-weight project manager for Python projects.
This is a tiny tool built to automate repetitive manual tasks, when creating a new Python development project and managing project dependencies. What you will end up with, when using jetzt?
- A new project directory under your current directory.
- A new virtualenv named
venv
under the project directory with the System-level installed Python 3.6+ version. - The virtualenv will have
pip
andsetuptools
installed and updated to latest available versions. - Additional packages will be installed based on the selected project type.
- Depending on the project type, other resources may be created to initialize a new project (all from within this repo).
This has been tested on the following setups:
- MacOS Sierra 10.12.6 and newer
- Python 3.6+ (installed via Homebrew)
- Zsh shell (with oh-my-zsh)
- Fedora 29
- Python 3.6+
- Bash and Bourne shells
Things should probably work nicely on older releases of macOS / OS X and on various Linux distros too. Python 3.6+ is expected.
A changelog is maintained.
Run jetzt --scaffold
in the directory, where you want to place your project directory, and follow the prompt.
You have couple of options for a new project.
Each project will have the following packages installed as development dependencies:
- flake8
- pytest
- jetzt
These will be listed in requirements-dev.txt
.
A metadata file jetzt_metadata.json
is maintained, and this will be used to manage the project metadata, including installed dependencies. This should be committed to git along with the rest of the project files.
This scaffolding option is meant to be a good starting point for a generic Python project. This option will not install any additional packages.
This scaffolding option is meant to be a good starting point for a new Flask-based webapp / API project. This option will install the following packages:
- Flask
These will be listed in requirements.txt
.
This scaffolding option is meant to be a good starting point for a new data science / analytics project. This option will install the following packages:
- jupyter + ipykernel
- pandas
- matplotlib
- seaborn
These will be listed in requirements.txt
.
Also a data
-directory will be created, and this notebook will be copied to the project root (the notebook is included in the installation package).
After scaffolding, start a jupyter server by running jupyter-notebook
in the project root directory.
All dependency management takes place within a virtualenv (created with jetzt --scaffold
), so make sure you have activated the project's environment before running these commands with source venv/bin/activate
. There is naturally the standard way of using pip to install dependencies, and manually add them to requirements.txt
, etc.
Jetzt includes an option to install and manage the dependencies for you. Please continue reading.
To install a package requests
, run jetzt --install
and follow the prompt. You have an option to install the package as a production dependency or as a development dependency. What does all this mean, you might ask? Jetzt will:
- Install the package requests (latest available version).
- Add the package
requests
tojetzt_metadata.json
with a version requirement set to minimum of the currently installed version. The packages, which requests depends on, are not added. You can install a dependency as aDEV
dependency too.
At the moment, you can only install one package at a time.
Example of a defined dependency with a version (as a installation requirement, or as seen in jetzt_metadata.json
):
requests>=2.21.0
Version pinning: To install a specific version of a package, add the version, just as you would with pip. For example: requests==2.20.1
. This will pin the version like so:
requests==2.20.1
To list installed dependencies, run jetzt --list
.
To list outdated dependencies, run jetzt --outdated
.
To update an outdated dependency, run jetzt --update
. This will allow you to choose an outdated dependency to update. After the update, the command will update the outdated dependency list automatically.
To remove an installed dependency, run jetzt --remove
. This will list all installed dependencies, and let's you choose which one you want to remove. After the update, the command will update the outdated dependency list automatically.
To reinstall all dependencies (excluding editable), run jetzt --reinstall
. The dependency versions are pinned temporarily for installation, so yo will end up with the same version as before the reinstallation. Make sure, you are in active virtualenv.
Run jetzt --create-requirements
to generate standard requirements.txt
and requirements-dev.txt
based on jetzt_metadata.json
.
MIT.