This is just a Sandbox.

Requirements:

• pyenv, for Python versions management
• poetry, for packaging and dependency management

1. Define the python version with pyenv, e.g.:

pyenv local 3.11

so that it also creates a .python-version file in the project root.

2. Create an env variable from the file.

export POETRY_PYTHON_VERSION=$(cat .python-version)

3. Set pyproject_template.toml to use python = "^${POETRY_PYTHON_VERSION".

4. Fill the template with the chosen python versions using the following command,

sed "s/python = \"^\${POETRY_PYTHON_VERSION}\"/python = \"^$POETRY_PYTHON_VERSION\"/" pyproject_template.toml > pyproject.toml

note that it directs the output to create or overwrite the file pyproject.toml.

5. Double check with pyenv versions that the correct version is active. Then force poetry to use this version:

poetry env use python

this already creates the poetry venv, but it doesn't install anything yet. You can debug with poetry debug info and sanity check that the System will show pyenv path and version as desired, and this should match the virtual environment. Poetry Python version does not need to match it.

$ poetry debug info

Poetry
Version: 1.7.1
Python:  3.12.0

Virtualenv
Python:         3.11.6
Implementation: CPython
Path:           ~/.cache/pypoetry/virtualenvs/sandbox-py3.11
Executable:     ~/.cache/pypoetry/virtualenvs/sandbox-py3.11/bin/python
Valid:          True

System
Platform:   linux
OS:         posix
Python:     3.11.6
Path:       ~/.pyenv/versions/3.11.6
Executable: ~/.pyenv/versions/3.11.6/bin/python3.11

6. Finally, you should be able to run poetry install to install all dependencies and poetry shell to manually spawn your virtual environment.