Improve river estimation stage - address downriver wall artefacts

[rosepearson]

The river estimation approach appears to introduce an 'wall artefact' down river amplified from any high elevation values detected along the measured cross section data.

As an example:

"See below the strange artifact of a ridgeline in the river that seems like an exaggerated version of a small positive feature in the surveyed XS (0.5 m), increased to 3 m in the interpolated data and extending to the next XS upstream and downstream."

Starting point:
Could it be due to the taking the nearest point along the measured sections instead of interpolating the value nearest to the sample locations.. https://github.com/rosepearson/GeoFabrics/blob/main/src/geofabrics/bathymetry_estimation.py#L772

[rosepearson]

Also support providing just a thalweg and wetted area instead of river banks. If providing the wetter area the thalweg is not optional it is required.

[rosepearson]

Update documentation to make it clearer what formats are supported.

• "Left" / "Right" river banks - extend to any two columns.
• What the deal with how the centreline is defined? From the sea? or a check for which is closest to the ocean contours?
• Supported wetted area in place of left/right banks.