[cwhite911]

Adds COPC support to r.in.pdal.

• Detect the CRS of COPC inputs: the point table SRS is not populated before execute(), so fall back to the reader stages. The -p file info output now prints the projection.
• When the computational region is smaller than the point cloud and no reprojection is requested, pass the region bounds to readers.copc so the COPC spatial index skips data outside the region. The GRASS spatial filter still does the exact clip.
• Add a point_table_capacity option (validated >= 1) for tuning the PDAL streaming buffer.
• Initialize all GrassRasterWriter members explicitly. base_segment_ was previously uninitialized and read in processOne(), undefined behavior that could corrupt imported values when no base raster was given.
• Count points on the region edge and report them once instead of one message per point.
• pytest tests cover COPC import, region-bounds pruning, and point_table_capacity.

[petrasovaa]

I have #6465 open, see if there is some overlap.

[cwhite911]

There is some overlap with accessing the CRS information that would make more sense as part of the other PR, but this PR is still addressing other issues. I will add a review with my suggested addition to PR #6465

[cwhite911]

I have #6465 open, see if there is some overlap.

Yes. I'll wait for #6465 to merge and then make any necessary changes here afterwards

[petrasovaa]

Some minor points from AI review:

• The "Filtered spatially N points." statistic silently changes meaning for COPC inputs, since the reader now prunes before GrassLidarFilter sees the points.
• "The compressed LAS (.laz) format can be imported only if PDAL has been compiled with LASzip support" is inaccurate — PDAL reads LAZ via LASzip or lazperf, and lazperf is the default in current builds.