logger fixes: close streams and persist during warm start

[BrotherHa]

Hallo,

I have added an abstract 'dummy' method to close the logger in instance to AbstractLoggerSpec and a method with the same function for the TensorboardLoggerSpec. The method is called in sr.py after the logging file is completed. This enables the use of the parameter overwrite = True for the TensorBoardLoggerSpec and solves issue (Closes #761) .

Best regards

[MilesCranmer]

I'm confused because according to this: JuliaLogging/TensorBoardLogger.jl#109, this file closing should already happen automatically? So I am not sure why it does not in our case.

Also it should call Base.close rather than looping through the files

[BrotherHa]

Effectively I reimplemented the Base.close method as close_logger to call it from logger_spec.py. If it is possible to call the Base.close method directly from logger_spec.py that would definitely be the better option.

[BrotherHa]

Concerning JuliaLogging/TensorBoardLogger.jl/109 I am not sure, why it does not close the logger, but if possible it would be great to already solve this issue on the level of the underlying Julia-code I think.

[MilesCranmer]

Effectively I reimplemented the Base.close method as close_logger to call it from logger_spec.py. If it is possible to call the Base.close method directly from logger_spec.py that would definitely be the better option.

I think you can do jl.close?

[coveralls]

Pull Request Test Coverage Report for Build 12206177822

Details

• 12 of 12 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.05%) to 93.568%

---

Totals
Change from base Build 12203983899:	0.05%
Covered Lines:	1382
Relevant Lines:	1477

---

💛 - Coveralls

[MilesCranmer]

Thanks @BrotherHa!