Trace level logging for reads

[jisantuc]

With read errors, it's difficult to tell what's actually going on in any client application using geotrellis-contrib.

With trace level logging of:

• uri
• warp options
• extent

that would be easier

[metasim]

+1

[metasim]

Ideally would be great to know all the scenarios where GDAL is making IO calls.

[metasim]

geotrellis/gdal-warp-bindings#34

[echeipesh]

Added this gem when debugging new OGC services:

geotrellis-contrib/vlm/src/main/scala/geotrellis/contrib/vlm/avro/GeotrellisReprojectRasterSource.scala

Lines 66 to 77 in e0fc7d6

	lazy val tileBounds = sourceLayer.metadata.mapTransform.extentToBounds(sourceExtent)
	lazy val pixelsRead = (tileBounds.size * sourceLayer.metadata.layout.tileCols * sourceLayer.metadata.layout.tileRows).toDouble
	lazy val pixelsQueried = (targetRasterExtent.cols.toDouble * targetRasterExtent.rows.toDouble)
	def msg = s"""${GREEN}ead($extent)${RESET} =
	|\t${BOLD}FROM${RESET} $uri ${sourceLayer.id}
	|\t${BOLD}TARGET${RESET} ${targetRasterExtent.extent} ${targetRasterExtent.cellSize} @ ${crs}
	|\t${BOLD}SOURCE${RESET} $sourceExtent ${sourceLayer.metadata.cellSize} @ ${sourceLayer.metadata.crs}
	|\t${BOLD}READ${RESET} ${pixelsRead/pixelsQueried} read/query ratio for ${tileBounds.size} tiles""".stripMargin
	if (tileBounds.size < 1024) // Assuming 256x256 tiles this would be a very large request
	logger.debug(msg)
	else
	logger.warn(msg + " (large read)")

it has been invaluable and ANSI colors are Purdy