Add README for scripts folder #76
Conversation
|
|
||
| To install Poetry, follow the Poetry [installation guide](https://python-poetry.org/docs/#installation). | ||
|
|
||
| To install from the lockfile, run `poetry update`. |
There was a problem hiding this comment.
Is this poetry update step supposed to create a new environment?
Because that doesn't seem to be what it is doing, in the way I tried it. We have a conda environment.yml file in the examples/ directory to create the example.parquet file. And so that environment was activated when, after updating the example, I ran those instructions. And it seems to have updated my conda installation instead of creating a separate virtual environment.
There was a problem hiding this comment.
🤦♂️ Sorry, I wrote this doc too quickly. To install from a lockfile you use poetry install, not poetry update. The latter updates the lockfile with new dependencies (but still constrained to the semver dependencies in pyproject.toml).
I don't use Conda so I'm not familiar of any ways Poetry could interact with it. One way to check how Poetry is behaving is by setting the config settings to create a virtualenv in the project directory:
poetry config virtualenvs.create true
poetry config virtualenvs.in-project true
Then when I run poetry install within scripts/ I see this:
> poetry install
Creating virtualenv geoparquet-scripts in /Users/kyle/github/mapping/geoparquet/scripts/.venv
Installing dependencies from lock file
Package operations: 2 installs, 0 updates, 0 removals
• Installing numpy (1.22.3)
• Installing pyarrow (7.0.0)
And then you should see a .venv directory in the local directory, and be able to open that environment (even outside of poetry run) with a normal source ./.venv/bin/activate.
There was a problem hiding this comment.
That said, I suggested this setup because it's what I'm most familiar with, but it's ok if the project standardizes on another system. Mostly I like having the reproducibility of lockfiles while only needing to strictly define top-level dependencies (see: ref1, ref2). Poetry, Pipenv, and pip-compile+requirements.txt+virtualenv all have similar goals.
We could explore conda-lock.
For #68 (comment)