Python 3.5¶

This section introduces the Python 3.5 packaging workflow built around declarative setup.cfg, a small setup.py compatibility shim, and split runtime and development requirements.

Applied Project¶

Project Setup¶

The applied project is Historic Calculator, release 5.0.0. This snapshot is set in 2015, the release year of Python 3.5, and contains the historic_calculator package plus the hist_calc command-line script.

Run the Project¶

Application commands are documented in the section README.

Background¶

This project belongs to the Python 3.5 and 2016 packaging era, when declarative configuration became normal for setuptools projects. Setuptools 30.3.0 added [metadata] and [options] sections in setup.cfg, giving projects a declarative place for:

Recording package metadata
Finding packages in a src layout
Declaring runtime dependencies
Registering console-script entry points
Configuring wheel and test tooling

setup.cfgsetup.pyrequirements.txtrequirements-dev.txt

The setup.cfg file records the package metadata, dependencies, entry point, and tool settings:

[metadata]
name = historic_calculator
version = 5.0.0
description = A tiny vector calculator using NumPy and Click.

[options]
package_dir =
    =src
packages = find:
install_requires =
    numpy
    click

[options.packages.find]
where = src
exclude =
    tests
    tests.*

[options.entry_points]
console_scripts =
    hist_calc = historic_calculator.main:cli

[bdist_wheel]
universal = 0

[tool:pytest]
testpaths = tests
addopts = -ra

[metadata] stores the distribution name, version, and human-readable description.
[options] maps packages to the src directory, discovers packages, and declares runtime dependencies.
[options.packages.find] tells setuptools where to find packages and which test packages to exclude.
[options.entry_points] registers the hist_calc command through console_scripts.
[bdist_wheel] configures the wheel build.
[tool:pytest] records pytest discovery and reporting settings.

The compatibility setup.py stays small because the metadata moved into setup.cfg:

from setuptools import setup

setup()

setup() lets older tools continue to invoke the setuptools build while reading configuration from setup.cfg.

The runtime dependency pins live in requirements.txt:

numpy==1.11.3
click==6.6

numpy==1.11.3 pins the NumPy release used for array-style calculations.
click==6.6 pins the Click release used for the command-line interface.

The development dependency pins live in requirements-dev.txt:

-r requirements.txt
pytest==3.0.7

-r requirements.txt reuses the runtime dependency set.
pytest==3.0.7 pins the test runner used by this project snapshot.

Dependency Management¶

Overview¶

Declarative configuration and split requirements made project setup clearer, but the model still had important limits:

⚠️ requirements.txt and requirements-dev.txt pinned direct install inputs, not a complete lockfile with hashes.
⚠️ Runtime dependencies were duplicated conceptually between package metadata and requirements files.
⚠️ setup.py was still required as a compatibility shim for many tools.

Runtime and build dependencies¶

Use Dockerfile.devEnv for the Python 3.5 development environment. It keeps the historical interpreter, pip, setuptools, wheel support, and build tooling isolated from the host machine.

NumPy provides array-style calculations, and Click provides the structured hist_calc command-line interface. requirements.txt pins the runtime dependency set, while requirements-dev.txt extends it with pytest for local development and tests.

Install the runtime dependencies:

pip install -r requirements.txt

Install the development dependencies when you need the test runner and local development tools:

pip install -r requirements-dev.txt

Build the package¶

Build a pure-Python wheel:

python setup.py bdist_wheel

Build a source distribution alongside it:

python setup.py sdist

PEP 427 wheels are fully usable in this era. setup.cfg already declares [bdist_wheel] universal = 0.

Install the package¶

Install the project:

pip install .