-
Notifications
You must be signed in to change notification settings - Fork 16
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
get sdist requires from egg-info #88
get sdist requires from egg-info #88
Conversation
Tests found at least one case where this errors out. I'll look into it. |
Should I delete the files afterward or does it not matter as code meant for the CI? (which I assume just cleans itself up) |
…/Avasam/stub_uploader into get-sdist-requires-from-egg-info
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks, remarks below.
stub_uploader/metadata.py
Outdated
for info in file_in: | ||
folder_name = info.name | ||
# Assume a single folder | ||
break |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would add a check here, whether folder_name
was actually set, or maybe even check that there really is exactly one folder to account for broken eggs. But for the former case, the following should be simpler and have that required error check:
for info in file_in: | |
folder_name = info.name | |
# Assume a single folder | |
break | |
# Assume a single folder | |
folder_name = next(iter(file_in)).name |
You can then also drop the initialization of folder_name
.
Or just use *
instead of folder_name
in the glob below.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Or just use * instead of folder_name in the glob below.
New code looks much cleaner like this, and makes no assumption. Especially since it's extracting to a temp folder now.
stub_uploader/metadata.py
Outdated
) | ||
|
||
|
||
def get_sdist_requires( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nit: "get" makes this function sound simpler than what it actually does. Maybe "extract"?
Also I think it would be safer to extract the sdist into an empty temporary directory than to where ever we currently are.
use glob to not have to assume match count
At least some of the test failures will be fixed by #95 |
I was thinking integration tests might be too fragile, but I can just pin the version I'm testing against. I'm sure you'd like me to add tests for this? Quick testing script I've been using: from __future__ import annotations
import requests
from packaging.requirements import Requirement
from stub_uploader.metadata import extract_sdist_requires, validate_response
for req_name in ("pycocotools", "click-default-group", "numpy", "mypy", "pywin32"): # Pin these
req = Requirement(req_name)
resp = requests.get(f"https://pypi.org/pypi/{req.name}/json")
validate_response(resp, req)
data = resp.json()
sdist_data = next(
(
url_data
for url_data in reversed(data["urls"])
if url_data["packagetype"] == "sdist"
),
None,
)
if sdist_data is None:
print(f"{req_name}: sdist_data is None!")
continue
requirements = extract_sdist_requires(sdist_data, req)
print(f"{req_name}: {list(requirements)}") Results: pycocotools: [<Requirement('matplotlib>=2.1.0')>, <Requirement('numpy')>]
click-default-group: [<Requirement('click')>]
numpy: []
mypy: [<Requirement('typing_extensions>=3.10')>, <Requirement('mypy_extensions>=1.0.0')>, <Requirement('tomli>=1.1.0')>, <Requirement('typed_ast<2,>=1.4.0')>, <Requirement('psutil>=4.0')>, <Requirement('pip')>, <Requirement('typed_ast<2,>=1.4.0')>, <Requirement('lxml')>]
pywin32: sdist_data is None! |
CC @AlexWaygood This was the idea I mentioned.
When the requirements can't be found in
requires_dist
, fallback to searching in *.egg-info. All without executing any arbitrary code.I move the
f"Expected dependency {req} to be present in the allowlist"
error higher so it can fail earlier and leave the network requests last.It's my first time playing around with downloading and manipulating archives in python. Let me know what I can do better here.
I tested the idea with
pycocotools
(best case scenario)click-default-group
(on my windows machine with 7zip there's an extra folder between the gz and tar, still worked),numpy
(no *.egg-info folder),mypy
(has empty lines and [extras]) andpywin32
(no Source Distribution)