Capturing warnings with the correct source location (hashtag1076)

[vikramsubramanian]

When logging the path from which the warning message occurred, it will show the exact location of where you patched showwarning instead of where the warning message was actually called from.

For example, the first and last warnings are coming from PyTorch Lightning, but here it says that they come from my logging.py module, where I did what's described in the link above.

<img width="1100" alt="image" src="

Is there any viable solution to this? Thanks!

)

[mayil-ai[bot]]

Possible Solution

• The provided code snippets do not contain any direct references to PyTorch Lightning or the showwarning function being patched.
• The issue described is related to incorrect source location in warning messages, which suggests that the showwarning function might not be capturing the correct stack frame when logging the warning.
• To address the issue, locate the custom showwarning function in the codebase (likely in a logging.py module or similar) and ensure it captures the correct stack frame to report the accurate source location of the warning.
• If the custom showwarning function is not found in the provided snippets, it is likely defined elsewhere in the codebase. Search the entire codebase for the definition and patching of showwarning.
• Once located, modify the showwarning function to use the stacklevel parameter or inspect the stack manually to find the original call site of the warning and log it correctly.
• Ensure that the filename and lineno parameters passed to the logging system within showwarning reflect the original call site, not the location of the showwarning function itself.

Code snippets to check

loguru → _logger.py

This snippet includes the logic for capturing the file path and line number when logging, which is directly related to the issue of incorrect source location in warnings. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_logger.py#L1906-L2020 This snippet describes the logger configuration, which could be relevant when investigating how warnings are captured and logged. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_logger.py#L253-L2094

loguru → _better_exceptions.py

This snippet is part of the exception formatting which might be involved in the process of logging warnings and their source locations. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_better_exceptions.py#L331-L380 This snippet deals with extracting frames and source information from tracebacks, which could be related to how the source location of warnings is determined. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_better_exceptions.py#L192-L254