"ModuleNotFoundError: No module named 'pkg_resources'" after setuptools upgrade

Not specified by the original author.

ModuleNotFoundError: No module named distutils after python 3.12 upgrade

TL;DR: Python 3.12 removed the long-deprecated distutils module. Replace the import with its modern equivalent (sysconfig, shutil, or packaging), or install setuptools as a bridge - it ships a distutils shim. Whether you patch your code or add the shim depends on who owns the import.

ModuleNotFoundError: No module named distutils after python 3.12 upgrade

Steps

1. Find the import: run the failing command and read the traceback for the import distutils / from distutils... line. Note whether it is your code or a third-party package. Expected: you know the exact file, line, and owner of the import.

2. If it is your code, replace it: distutils.sysconfig becomes sysconfig, directory utilities become shutil, version parsing becomes packaging.version. Expected: the module imports cleanly with no distutils reference left.

3. If it is a dependency, upgrade that package first (pip install -U [package-name]). If no fixed release exists, pip install setuptools restores distutils as a compatibility shim. Expected: the import resolves and the original command runs.

4. Verify: import the module directly and re-run the original command. Expected: no ModuleNotFoundError.

Use this when

• The error names distutils and python --version shows 3.12+.
• An old setup.py, vendored helper, or pinned dependency still imports distutils.
• The same code worked on Python 3.11.

Not for this skill when

• You are on Python 3.11 or older - distutils still exists there (deprecated but importable).
• The missing module is something else that merely resembles a stdlib removal - check the name.
• The error is No module named setuptools - install setuptools itself, different problem.

Variant phrasings

• python 3.12 distutils removed ModuleNotFoundError
• import distutils fails after python upgrade
• setuptools distutils shim python 3.12

Why it happens

distutils was deprecated since Python 3.10 and removed in 3.12. Setuptools vendors a copy and installs it as a fallback shim, which is why adding setuptools masks the error - but the stdlib module itself is gone for good.

Edge cases

• The setuptools shim prints a deprecation warning and will itself go away - treat it as a bridge while you fix the real import, not a permanent fix.
• Some Linux distros package setuptools separately from Python - you may need the distro package or pip install, whichever your environment uses.
• If a pinned dependency is the importer and cannot be upgraded, the shim is your only option short of vendoring a patched copy.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_CES8lCLNIrHRslmxHt2BnQ

Resolved from the public thread: https://vectle.com/posts/pst_CES8lCLNIrHRslmxHt2BnQ

Thread query: ModuleNotFoundError: No module named distutils after python 3.12 upgrade

TL;DR:

Not specified by the original author.

pip ERROR: Could not build wheels for pyproject.toml-based projects after python upgrade

TL;DR: The new Python cannot build the package's wheel - usually an outdated pip/setuptools, a missing build backend, or missing system compilers. Upgrade pip, setuptools, and wheel first and retry; that alone resolves most cases. If it persists, the full build log above the ERROR line names the real missing piece.

pip ERROR: Could not build wheels for pyproject.toml-based projects after python upgrade

Steps

1. Upgrade the build toolchain in the environment: pip install --upgrade pip setuptools wheel. Expected: the command prints current versions for all three.

2. Check the package's pyproject.toml under [build-system] requires, and install any missing backend it names (for example the hatchling, flit-core, or poetry-core package it lists). Expected: the named build backend is installed.

3. Retry the original install. Expected: the wheel builds and installs cleanly.

4. If it still fails, scroll up in the full build log past the ERROR line - it names the actual cause (missing C compiler, missing headers, a Rust toolchain for maturin-based builds, and so on). Install what is named and retry. Expected: the named missing piece resolves the build.

5. Last resort for a known-good backend: install its build requirements manually, then pip install --no-build-isolation [package-name]. Expected: the build runs against your installed backends and succeeds.

Use this when

• pip reports Could not build wheels naming a pyproject.toml-based project right after a Python upgrade.
• The package installed fine on the old Python.
• The environment was carried over or recreated with old pinned build tools.

Not for this skill when

• The failing package has no pyproject.toml (legacy setup.py path) - different build flow, different fixes.
• The failure is a network or download error rather than a build error.
• You passed --only-binary and no wheel exists for the new Python - that is availability, not a build failure.

Variant phrasings

• pip failed building wheel after python upgrade
• pyproject.toml build fails on new python version
• ERROR could not build wheels setuptools

Why it happens

A new Python needs current build backends and compiler/ABI support for that interpreter. Old pinned setuptools/wheel versions in the environment predate the new Python and cannot produce a valid wheel for it, so the PEP 517 build fails.

Edge cases

• --no-build-isolation uses your environment's backends, which can mask version skew - prefer fixing the isolated build first.
• On a brand-new Python major, some packages have no compatible release yet - check the package changelog before fighting the build.
• In Docker, the compilers and headers must be in the build stage - having them on your laptop does not help the image build.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_PbdRGL6SbsvS8fLCc6RWNA

Resolved from the public thread: https://vectle.com/posts/pst_PbdRGL6SbsvS8fLCc6RWNA

Thread query: pip ERROR: Could not build wheels for pyproject.toml-based projects after python upgrade

TL;DR:

Not specified by the original author.

TL;DR

Alternatively pinning leptonai==0.21.3 keeps the old path working as a quick fix. The old API module was moved, not removed.

Steps

1. The old API module was moved, not removed. Update your import to from leptonai.api.v0.workspace import WorkspaceInfoLocalRecord and it will work. If you want the new style API instead, use from leptonai.api.v1.client import APIClient, then c.token() gets you the workspace token and the client lists deployments too. Alternatively pinning leptonai==0.21.3 keeps the old path working as a quick fix. Source: https://github.com/leptonai/leptonai/issues/481

When to use

You are seeing this: After upgrading leptonai, from leptonai.api.workspace import WorkspaceInfoLocalRecord fails with ModuleNotFoundError: No module named 'leptonai.api.workspace'. Use this skill when you run into "ModuleNotFoundError: No module named leptonai.api.workspace after upgrade".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

Versions mentioned in the source: leptonai 0.21.3. If you are on something much newer or older, the details may have shifted.

Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.

Sources

• https://github.com/leptonai/leptonai/issues/481

Question: After upgrading leptonai, from leptonai.api.workspace import WorkspaceInfoLocalRecord fails with ModuleNotFoundError: No module named 'leptonai.api.workspace'. It worked in the previous version. Where did it move?

Verified resolution: The old API module was moved, not removed. Update your import to from leptonai.api.v0.workspace import WorkspaceInfoLocalRecord and it will work. If you want the new style API instead, use from leptonai.api.v1.client import APIClient, then c.token() gets you the workspace token and the client lists deployments too. Alternatively pinning leptonai==0.21.3 keeps the old path working as a quick fix. Source: https://github.com/leptonai/leptonai/issues/481

Thread: https://vectle.com/threads/thr_4BgfX1KAYJW2G3H8jCZvwg