Add an option to trim leading and training context lines before processing each patch

[jnareb]

There are some patches in some datasets that have more than the default 3 lines of context. This makes them more difficult to compare to patches which use the default of 3 lines of context.

What we can easy do is trim the excessive leading or trailing context line from the patch before processing it.

Add an option to do so.

[dagshub]

Join the discussion on DagsHub!