pipx fails on some packages where pipsi works

pipx doesn’t find the binary for this package but pipsi does.

$ pipx install py-spy
No binaries associated with package py-spy. 
$ pipx --version
0.12.3.3
$ pipsi --version
pipsi, version 0.9, python /usr/bin/python3

$ pipsi install py-spy
Already using interpreter /usr/bin/python3
Using base prefix '/usr'
New python executable in /home/user/.local/venvs/py-spy/bin/python3
Also creating executable in /home/user/.local/venvs/py-spy/bin/python
Installing setuptools, pkg_resources, pip, wheel...done.
Collecting py-spy
  Using cached https://files.pythonhosted.org/packages/26/c2/bacd6ff83b43e1eff51d0d93a1fc930daad02fa93c0499d6facece57f786/py_spy-0.1.10-py2.py3-none-manylinux1_x86_64.whl
Installing collected packages: py-spy
Successfully installed py-spy-0.1.10
  Linked script /home/user/.local/bin/py-spy
Done.

[ghost]

@cs01 Thanks for checking. It works for me too on the latest pipx.

[cs01]

@amcgregor nice! I am a big fan of venv and I use a similar shell function I called venvhere which I think does the same thing as your venv function :

venvhere () 
{ 
    python3 -m venv venv --prompt $(basename $PWD) && source venv/bin/activate && pip install -U pip --quiet
}

Why does this seem to be a good idea vs. “proper” use of virtual environments, directly?

It is only a good idea for certain workflows. I agree that creating and manually managing venv’s for development of a project is the best way to do it. pipx is definitely not meant to be a replacement for all venv interactions.

Its goal is to make CLI binaries accessible globally, that’s it. Using venv is the best way for pipx to do that. pipx saves you a few steps, and makes a single UI to list, upgrade, uninstall, these binaries/packages. I actually don’t know of an easy way to know which entry points a package offers other than using pipx or inspecting the package’s source code.

If you don’t find value in it, that is totally fine! No need to use it. I made it to solve a problem: simplify and improve the installation process of another Python program I made.