Refactor SDF handling logic for non-existing string paths

[flferretti]

The non-existing paths will be handled by pathlib.Path.read_text(), eventually raising a FileNotFoundError.

[flferretti]

The pre-commit.ci is failing as #48 needs to be merged before this

[diegoferigo]

I'm not 100% sure that the new logic covers entirely the old logic. Also, the previous refactoring might have introduced some unchecked cases.

It's probably worth to have a unit test for such a core component of the library like loading the supported types of resources.

[flferretti]

I'm not 100% sure that the new logic covers entirely the old logic. Also, the previous refactoring might have introduced some unchecked cases.

It's probably worth to have a unit test for such a core component of the library like loading the supported types of resources.

I added additional tests and refactored forcing is_urdf to be more robust

[diegoferigo]

I think the three original cases were correct, and now something is missing:

1. If the resource is a path, expect that is exists and read its content.
2. If the resource is a string:

• a) if its length is shorter than the max path length, check if it's a file. If it is, read its content, otherwise treat it as a string containing the URDF/SDF.
• b) if its length is longer than the max path length, treat is as a string containing the URDF/SDF.

How is 2a handled now? The logic seems changed.

[flferretti]

I think the three original cases were correct, and now something is missing:

1. If the resource is a path, expect that is exists and read its content.

2. If the resource is a string:


* a) if its length is shorter than the max path length, check if it's a file. If it is, read its content, otherwise treat it as a string containing the URDF/SDF.

* b) if its length is longer than the max path length, treat is as a string containing the URDF/SDF.

How is 2a handled now? The logic seems changed.

Solved by rebase. I had to make the assumption that the passed string path has an extension, I couldn't find any other way to distinguish between a URDF shorter than MAX_PATH and a string path.

[diegoferigo]

Great! Now it looks much better, thanks @flferretti.