Is there a way to inspect the temporary environment created?

[gitgithan]

Previously I reported in #7730 that uv run test_uv/demo.py works unexpectedly but uv run --python 3.10 test_uv/demo.py fails.

I found a verbose flag in uv run but it only showed the python resolution process, with no information on the temporary environment created or used.

Are details on the temporary environment created available in any part of the cli? If not could anyone point to the source code where envs are get or created? Are they in ~/.cache/uv/environments-v1?

I wish to understand the env creation process triggered by requesting a non-default python like in uv run --python 3.10 test_uv/demo.py to be able to predict what libraries will be available or not.

For issue #7730, uv run test_uv/demo.py worked not because The first case may have succeeded as uv run will use the .venv in the current directory by default as Charlie suggested. I did not have any virtual env in the same directory that contained rich. It turns out it was because uv was using the pip associated with the default python.

Here's how I discovered that:

1. uv python find --> /usr/bin/python3
2. ll /usr/bin/python3 --> /usr/bin/python3 -> python3.8
3. python3.8 -m pip list --> the list contained rich and requests
4. pip3.8 uninstall rich to break uv run test_uv/demo.py, thus proving uv was using environment (is this considered system level?) associated with python3.8 and pip3.8

---

Alternatively to step 3, python3.8 -m pip show pip , then look at Location: /home/hanqi/.local/lib/python3.8/site-packages to find rich inside

[charliermarsh]

In this case, does the script have a # /// script tag?

[gitgithan]

No # /// script, it was the same script as #7730

import requests
from rich.pretty import pprint
import sys

print(sys.version)

import importlib.metadata

rich_version = importlib.metadata.version("rich")
print("Rich Version: ", rich_version)

resp = requests.get("https://peps.python.org/api/peps.json")
data = resp.json()
pprint([(k, v["title"]) for k, v in data.items()][:10])

[charliermarsh]

If you're in a project (i.e., a directory or parent directory with a pyproject.toml), uv run will by default run in a virtual environment located at .venv (relative to the project root).

If you're not, it will run in the discovered interpreter (no environment necessary), unless you pass --with, in which case, it will create an ephemeral environment in the cache: https://github.com/astral-sh/uv/blob/005bb235f03c96d00d0231f1e1a4dd4eca0d4c0b/crates/uv/src/commands/project/environment.rs#L19