[bug] Incompatibility with numba 0.57.0 version information

[meyer89]

Issue Report Checklist

• [x] Searched the issues page for similar reports
• [x] Read the relevant sections of the documentation
• [x] Browse the tutorials and tests for usefull code snippets and examples of use
• [x] Reproduced the issue after updating with pip install --upgrade pandapower (or git pull)
• [x] Tried basic troubleshooting (if a bug/error)
  • [x] Restarted interpreter
  • [x] Checked the pythonpath
• [x] Completed the Problem Description, What steps reproduce the problem and Versions sections below

Problem Description

Pandapower is not compatible with numba 0.57.0. Installing the older numba version 0.56.4 solves the problem.

The interface changed in numba and the attribute version_version is no longer available.

What steps reproduce the problem?

Use fresh virtual environment or docker image.

1. Execute pip install numba==0.57.0
2. Execute pip install pandapower
3. Execute python -c "from pandapower.auxiliary import _check_if_numba_is_installed; print(_check_if_numba_is_installed(True))"

What is the expected output? What do you see instead?

Expected output:

True

Actually observed output:

numba cannot be imported and numba functions are disabled. Probably the execution is slow. Please install numba to gain a massive speedup. (or if you prefer slow execution, set the flag numba=False to avoid this warning!)

False

Paste Traceback/Error Below (if applicable)

>>> from numba._version import version_version as numba_version
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: cannot import name 'version_version' from 'numba._version' (/usr/local/lib/python3.10/site-packages/numba/_version.py)

Old content of numba._versions.py

version_version = '0.56.4'
version_full = '288a38bbd5a15418a211bf067878dfdf3c139509'
def get_versions(default={}, verbose=False):
    return {'version': version_version, 'full': version_full}

New content

import json

version_json = '''
{
 "date": "2023-05-01T10:59:47-0500",
 "dirty": false,
 "error": null,
 "full-revisionid": "4fd4e39c672d119b54a2276d170f270764d2bce7",
 "version": "0.57.0"
}
'''  # END VERSION_JSON

def get_versions():
    return json.loads(version_json)

Importing the function get_versions() and evaluation the dictionary key version would work in both numba versions.

Versions

• pandas version: 2.0.1
• networkx version: 3.1
• scipy version: 1.10.1
• numpy version: 1.24.3
• packaging version: 23.1
• tqdm version: 4.65.0
• deepdiff version: 6.3.0
• Operating System name/version: Docker-Image python:3.10 @ 2b8b079d7548

Labels (if access available)

• [ ] Choose a labels at the right side (e.g. bug, controller, OPF...)

[SteffenMeinecke]

With this set of packages I get the a NumbaDeprecationWarning:

C:\Users\smeinecke\Documents\GIT\pandapower\pandapower\pypower\makeLODF.py:25: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.
  def update_LODF_diag(LODF): # pragma: no cover

Maybe @rbolgaryn, you can have a look since it it about LODF?

deepdiff 5.8.1 matplotlib 3.7.1 networkx 2.6.3 numba 0.57.0 numpy 1.23.5 packaging 23.0 pandas 2.0.1 scipy 1.10.1

[AnkurArohi]

I have encountered the same problem, there is a way out of it although.

The auxiliary script in pandapower is ending up in exception because the way the script is trying to extract the numba version, is not compatible with the numba _version script. Hence the output of the

from numba._version import version_version as numba_version

leads already to exception.

I will try to create a PR.

[CarliJoy]

Is this bug fixed?

[CarliJoy]

Btw. instead of using private imports of modules that are likely to change to determine the version of modules, I would recommend to use importlib.metadata.version("numba").

This works from Python 3.8 onwards.

See https://stackoverflow.com/questions/20180543/how-do-i-check-the-versions-of-python-modules

[CarliJoy]

I added an MR doing exactly this.

[RedBatProject]

it still didnt work for me, i have python 3.10 and numba 0.57

[SteffenMeinecke]

sorry to say @RedBatProject but reckon it is something with your python installation, since it is working with python 3.10 and numba 0.57 on all laptops I know.

[brauliobarahona]

hallo everybody, I observed the same issue as @RedBatProject :

 python3.10 -c "from pandapower.auxiliary import _check_if_numba_is_installed; print(_check_if_numba_is_installed(True))"
numba cannot be imported and numba functions are disabled.
Probably the execution is slow.
Please install numba to gain a massive speedup.
(or if you prefer slow execution, set the flag numba=False to avoid this warning!)

False

I am running on macOS Sonoma, python 3.10, pandapower 2.13.1, numba 0.57

Package         Version
--------------- ------------
deepdiff        6.6.1
llvmlite        0.40.1
networkx        3.2.1
numba           0.57.0
numpy           1.24.4
ordered-set     4.1.0
packaging       23.2
pandapower      2.13.1
pandas          2.1.2
pip             23.3.1
python-dateutil 2.8.2
pytz            2023.3.post1
scipy           1.11.3
setuptools      65.5.0
six             1.16.0
tqdm            4.66.1
tzdata          2023.3

Note that with numba 0.56 the import works