Implementation:Sktime Pytorch forecasting Show Versions

Knowledge Sources	Sktime_Pytorch_forecasting
Domains	Time_Series, Forecasting, Deep_Learning
Last Updated	2026-02-08 08:00 GMT

Overview

show_versions, _get_deps_info, and _get_sys_info are utility functions that collect and display system information and dependency versions for debugging and issue reporting.

Description

show_versions prints a formatted summary of the Python version, executable path, OS platform, and version numbers for key dependencies (pytorch-forecasting, torch, lightning, numpy, pandas, and others). It delegates to _get_sys_info for system-level details and _get_deps_info for package version lookups. _get_deps_info supports two version resolution strategies: "distributions" (via importlib metadata / PEP 440 package names) and "import" (via module __version__ attributes). This module is adapted from scikit-learn's and sktime's utilities of the same name.

Usage

Use show_versions to quickly print environment details when debugging issues or filing bug reports. It provides a standardized summary of all relevant package versions and system configuration, making it easy to reproduce and diagnose environment-specific problems.

Code Reference

Source Location

• Repository: Sktime_Pytorch_forecasting
• File: pytorch_forecasting/utils/_maint/_show_versions.py
• Lines: 1-142

Signature

def show_versions():

def _get_sys_info():

def _get_deps_info(deps=None, source="distributions"):

Import

from pytorch_forecasting.utils._maint._show_versions import show_versions
from pytorch_forecasting.utils._maint._show_versions import _get_deps_info
from pytorch_forecasting.utils._maint._show_versions import _get_sys_info

I/O Contract

Inputs (show_versions)

Name	Type	Required	Description
(none)	--	--	Takes no parameters; uses the DEFAULT_DEPS_TO_SHOW list internally

Inputs (_get_sys_info)

Name	Type	Required	Description
(none)	--	--	Takes no parameters

Inputs (_get_deps_info)

Name	Type	Required	Description
deps	list[str] or None	No	List of package names to query; defaults to ["pytorch-forecasting"] if None
source	str	No	Version lookup strategy: "distributions" uses importlib metadata with PEP 440 names, "import" uses __version__ attribute with import names; defaults to "distributions"

Outputs

Name	Type	Description
show_versions	None	Prints system info and dependency versions to stdout; no return value
_get_sys_info	dict	Dictionary with keys "python", "executable", and "machine" containing system details
_get_deps_info	dict	Dictionary mapping package names to their version strings (or None if not installed)

Usage Examples

from pytorch_forecasting.utils._maint._show_versions import show_versions

# Print all system and dependency information
show_versions()
# Output:
# System:
#     python: 3.10.12 (main, ...)
# executable: /usr/bin/python3
#    machine: Linux-5.15.0-x86_64
#
# Python dependencies:
#            pip: 23.2.1
# pytorch-forecasting: 1.1.1
#          torch: 2.1.0
#      lightning: 2.1.0
#          numpy: 1.26.0
#          ...

from pytorch_forecasting.utils._maint._show_versions import _get_deps_info, _get_sys_info

# Get system info as a dictionary
sys_info = _get_sys_info()
print(sys_info["python"])   # e.g., "3.10.12 (main, ...)"
print(sys_info["machine"])  # e.g., "Linux-5.15.0-x86_64"

# Get specific dependency versions
deps = _get_deps_info(deps=["torch", "numpy", "pandas"])
print(deps["torch"])   # e.g., "2.1.0"
print(deps["numpy"])   # e.g., "1.26.0"

Related Pages

• Principle:Sktime_Pytorch_forecasting_Dependency_Diagnostics