---
title: How to deploy a python / hy package
date: 2022-11-29
---

You can pip install from the local git repo by putting the following in requirements.txt:
```
git+ssh://git@company.ltd/company/repo.git
```
Notice the different format to what git itself uses to point to a remote repository.
It's also possible to install direct from the local source tree, e.g.:
```
python -m pip install /src/program
```
This last way may be the easiest, doing versioning on the git server and installing from /src.

[python-for-hpc](https://betterscientificsoftware.github.io/python-for-hpc/tutorials/python-pypi-packaging/) says:
> A Python project will consist of a root directory with the name of the project. Somewhere inside this will be included a directory which will constitute the main installable package. Most often this has the same name as the project (This is not compulsory but makes things a bit simpler. Inside that package directory, alongside your python files, create a file called `__init__.py`. This file can be empty, and it denotes the directory as a python package. When you pip install, this directory will be installed and become importable.

A simple project may have this structure:

```
graphics
├── LICENSE
├── graphics
│   ├── __init__.py
│   └── screen.hy
├── README.md
├── pyproject.toml
└── setup.cfg
```

```conf
#setup.cfg
[metadata]
name = graphics
author = Name
author_email = name@some.domain
version = 0.1.2
description = A simple ncurses UI class
url = https://some.domain

[options]
python_requires = >= 3.8
include_package_data = True
package_dir = 
packages = find:

# if a command-line utility is required
[options.entry_points]
console_scripts =
    cmd-line-util = graphics.some-module:some-module

[options.packages.find]
where = 

[options.package_data]
* = *.hy
```
```toml
# pyproject.toml
[build-system]
requires = [
  "setuptools >= 40.9.0",
  "wheel",
]
build-backend = "setuptools.build_meta"
```
This means pip will throw away the top-level stuff (README etc.).

The same source goes on to recommend setuptools to make the package installable, provide package metadata, dependencies etc.

The python docs tutorials recommend setup.cfg as a simpler static alternative.
However setup.py could allow using the git tag as a version.


### References

[installing packages](https://packaging.python.org/tutorials/installing-packages/)

[packaging packages](https://packaging.python.org/tutorials/packaging-projects/)