Open vikramsubramanian opened 5 months ago
Based on the provided ticket and code snippets, here is the advice for designing an integration with logot
for loguru
:
Log Capture:
LoguruCapture
backend as it provides a unified approach for capturing logs between stdlib logging and loguru
.capturing
are propagated to the LoguruCapture
constructor.contextvars
-based approach to capture logs in an async context, as loguru
supports asynchronous logging.Log Message Matching:
.format()
-style placeholders in logot
to match loguru
's logging style.loguru
's bind
method for contextual logging.Integration Design:
logot.contrib.loguru
module that contains the LoguruCapture
class and the logged
function with .format()
-style support.stdlib
interface and provides first-class support for loguru
.Code Snippets:
from logot.contrib.loguru import LoguruCapture
with Logot().capturing(LoguruCapture) as logot:
- For log message matching:
```python
from logot.contrib.loguru import logged
logot.wait_for(logged.info("Hello {}"))
logot.wait_for(logged.bind(foo="bar").info("Hello {}"))
Additional Considerations:
loguru
as a hard dependency for logot
.loguru
community's practices and preferences by reviewing existing loguru
extensions or plugins.logot
integration with loguru
.The provided code snippets from the loguru
codebase do not directly impact the design of the logot
integration but provide context on how loguru
is structured and how it handles logging. The integration should be designed to work seamlessly with loguru
's existing API.
I've recently released a 0.1.0 preview of a log testing library called [logot]( It provides a nice fixture for testing [complex log patterns]( which I've found very useful when testing multithreaded and async codebases.
I'd like to provide a
loguru
integration. 🤗Since I'm not hugely familiar withloguru
, I was hoping you'd have some advice on what a good integration would look like.There are a couple of sticking points that I'd like to nail before committing to a v1.0.0 API.
hashtag Log capture
Currently, stdlib logging capture is enabled with a [context manager method](
Since the
loguru
integration would be an optionalextra
, a freestanding context manager could be provided like this:It's a bit awkward having a freestanding context manager for 3rd party integrations, so an alternative would be a capturing backend approach:
Any arguments given to
capturing
would be propagated to theLoguruCapture
constructor. This would then unify capture between stdlib logging andloguru
logging. The default capture backend would be the stdlib capture, and an ` would ensure good typing. 💪The alternative is a whole
Logot
subclass just forloguru
, which seems heavyweight and unpleasant. 🤮hashtaghashtag Log message matching
Rather than awkward regexes,
logot
lets you [match log messages]( with%
placeholders. The hope here is that it's much cleaner and more intuitive to match using the syntax used for log message placeholders. (Similar to howprintf
andscanf
pair nicely).But
loguru
uses lovely.format()
-style placeholders!Maybe it's fine to just use
%
placeholders inloguru
tests? 🤷But it might be nice to support
.format()
-style placeholders like this:Support for matching structured data would also be possible:
hashtaghashtag Advice appreciated! 🙏
Thanks for reading this far!
I'm inclined to go with a
LoguruCapture
backend combined with.format()
-style message matching. That would give first-class support forloguru
without compromising or complicating thestdlib
interface.However, I'm not really familiar with the
loguru
community and ecosystem! Maybe this is unappealing? Maybe it's already been done! 😬Hope you have a nice weekend! ❤️ )