Added flags to control output structure & another flag for flat dumps

[CollinShoop]

Hey @WoozyMasta 👋🏻 Very cool project, I was looking at making something just like this and figured I should contribute instead.

For my own use, having better control over the output structure would be ideal. I personally do not require resources to be broken up per file, and I also sometimes need all the detailed fields which were being removed via jq for debugging purposes (as opposed to backups, which I presume is the primary use case for this tool), so I've added a few new flags.

[WoozyMasta]

Hey
Looks cool, I've been wanting to make the --output-by-type mode for a long time, special thanks for that.

I'll have to find some time to check it all out. And I think these edits will be included in the new version.

[WoozyMasta]

Do you have time to update the "Commands and Flags" section in the README.md and the .env file with the new options? It would be great.

[WoozyMasta]

@CollinShoop, I researched the solution a little.

Shouldn't the --output-by-type and --flat keys be checked for the presence of a resource. When saving all available namespace resources (a couple of deployments), a large number of empty directories or files are created, especially if there are a large number of CRDs in the cluster.

Do you think it will not be difficult to implement it, and is it worth it?

After all, when I specify specific resources, only they are saved.

For example:

./kube-dump ns -n dev-app -r deployment,svc --output-by-type 

[WoozyMasta]

In any case, I accepted the changes and released the new version 1.1.0.

Thank you so much for the work done.

[CollinShoop]

@WoozyMasta Thanks for the quick review & merge 🥇

@CollinShoop, I researched the solution a little.

Shouldn't the --output-by-type and --flat keys be checked for the presence of a resource. When saving all available namespace resources (a couple of deployments), a large number of empty directories or files are created, especially if there are a large number of CRDs in the cluster.

Do you think it will not be difficult to implement it, and is it worth it?

After all, when I specify specific resources, only they are saved.

For example:

./kube-dump ns -n dev-app -r deployment,svc --output-by-type 

I had the same observation when testing this, but my reasoning for keeping as-is is that it may be better to positively indicate that there are none of a particular resource with an empty file than to omit it which could be ambiguous (ie did the dump even check that resource type?). It's a bit annoying in the case of empty namespaces, though, but I'm not sure of a simple way to avoid that now.

Do you have time to update the "Commands and Flags" section in the README.md and the .env file with the new options? It would be great.

I'll add this to my TODO 👍🏻

[CollinShoop]

Actually, I see you've already updated the README.md and .env 💯