Streetcover bug

[barneydobson]

Description

Found some oddly formatted lanes entries. This fixes it.

TODO:

• test
• update from Separate network grpahfcnns #247

[cheginit]

I am not sure about this ad-hoc solution. I'd suggest solving the issue at its source, i.e., add a new post-processing step for the street graph to ensure that lane's edge attribute is an int.

[barneydobson]

I am not sure about this ad-hoc solution. I'd suggest solving the issue at its source, i.e., add a new post-processing step for the street graph to ensure that lane's edge attribute is an int.

Yep good suggestion - I'll look at next week

[dalonsoa]

This needs appropriate typing to include the case of the input not being a string, and update the docstring, accordingly. If it is not a string, what would it be? Output typing also needs updating.

[barneydobson]

I am not sure about this ad-hoc solution. I'd suggest solving the issue at its source, i.e., add a new post-processing step for the street graph to ensure that lane's edge attribute is an int.

this will require a bit more than I want to do in this PR so I will make a new issue