Samyak Sarnayak

I would suggest flake8 as a linting tool. Since you're using it with black you'll have to create a compatible configuration. ([ref](https://black.readthedocs.io/en/stable/compatible_configs.html#flake8))

Because you're planning to have type annotations everywhere, I would suggest using [mypy](https://github.com/python/mypy) to perform static type checking. This could also be added as a CI step (#9).

Would the scraper need to run all analyzers on all files in the cloned repo? Or would it only download the files?

I understand the flow now. I will try to understand the Pastebin scraper and then start work on this.

Ah I see. Well, I think I will stick with long versions for flags that can have negative values. Thank you making kong, I really like how simple it is...

Though, if I may ask > `=` is not valid with short flags Is that a convention? or is it something that kong does not support?

I tried this with `head` and it correctly parses arguments of the form `-a -10`. Both of these work as expected: ```bash echo "hello world" | head -c 5 ```...

I tried fixing it with this: https://github.com/Samyak2/kong/commit/86d1f36058eb5c30172560250196a28dd9d6e065 It did not work. I don't think that's the right approach as it should only be considered if it's followed by a short...

I noticed this warning when building metacall today. I see that it's used for these features `cxx_alignas cxx_alignof cxx_constexpr cxx_final cxx_noexcept cxx_nullptr cxx_sizeof_member cxx_thread_local` in the `write_compiler_detection_header` call present in...

Porting this to PyTorch lightning would be great. It has support for multiple optimizers ([reference](https://github.com/PyTorchLightning/pytorch-lightning-bolts/blob/master/pl_bolts/models/gans/basic/basic_gan_module.py)) and training on TPU is very simple ([reference](https://pytorch-lightning.readthedocs.io/en/latest/tpu.html)).