Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

mach-nix fails to find scipy dependency

See original GitHub issue

Hello, thank you for your project, it looks super promising.

Just got this:

The Package 'scipy' cannot be found in the dependency DB used by mach-nix.
Please check the following:
  1. Does the package actually exist on pypi? Please check https://pypi.org/project/scipy/
  2. Does the package's initial release date predate the mach-nix release date?
     If so, either upgrade mach-nix itself or manually specify 'pypi_deps_db_commit' and
     'pypi_deps_db_sha256 for a newer commit of https://github.com/DavHau/pypi-deps-db/commits/master
If none of that works, there was probably a problem while extracting dependency information by the crawler maintaining the database.
Please open an issue here: https://github.com/DavHau/pypi-crawlers/issues/new

By trying mach-nix on github.com/mangaki/zero dependencies. I’m pretty sure that scipy exists on PyPI and was released before mach-nix 😛 ; but I don’t know what’s going on.

Issue Analytics

State:
Created 3 years ago
Comments:11 (7 by maintainers)

Top GitHub Comments

1reaction

DavHaucommented, Apr 29, 2020

I added some instructions on how to debug dependency extraction to pypi-crawlers . To test my instructions i debugged scipy’s setup.py already. Actually this setup.py cannot work with distutils / setuptools installers. It definitely imports numpy to early in any case. Our environment doesn’t make a difference there. Executing the setup.py inside the python docker container which is debian based also fails with the same error. I wouldn’t call it broken because of that. Today i learned, the setup.py-style setup is considered legacy and there are other ways to declare dependencies like pyproject.toml (PEP 518) which is also supported by pip. Scipy’s build seems to depend on it. The way i see it, adding support for PEP 518 (pyproject.toml) to pypi-crawlers would be the right way to include scipy and some other libraries. It should be fairly simple since it’s just about parsing the project.toml file. Nothing must be fake-installed.

0reactions

DavHaucommented, May 19, 2020

Mach-nix version 2.0.0 has two new providers nixpkgs and wheel. The problem should be resolved by this

Top Results From Across the Web

mach-nix removes required dependencies from nixpkgs ...

When trying to install cartopy mach-nix removes all python dependencies, which results in a failed build. I assume the dependencies are ...

Mach-nix: Create python environments quick and easy

This lead to failing builds especially for older python versions. Now requires_python is part of the dependency graph and affects resolution.

Serverless python packaging numpy dependency error

I have been running into issues when making function calls from my deployed Python3.7 Lambda function that, from the error message, are related ......

Integration (scipy.integrate) — SciPy v1.9.3 Manual

import scipy.integrate as integrate >>> import scipy.special as special ... of the integral and the second element holding an upper bound on the...

Adding GDAL as dependency to Python package?

Before installing GDAL python binding, you must install GDAL on your system. apt-get install -y libgdal-dev. After that, set variables and ...