optimize file based sources

[smnorris]

For files manually compiled, consider downloading from source where available (vs manually compiling and storing on s3).

Where manual compilation is still required:

• include just the layer of interest in the file
• standardize format

[smnorris]

Looking at the first candidate (migratory bird sancuaries / national wildlife area), it doesn't seem like ogr can reach the .gdb via vsicurl - curl needs the -L flag to follow redirects and download successfully:

curl -v -L -O "https://data-donnees.az.ec.gc.ca/api/file?path=%2Fspecies%2Fprotectrestore%2Fcanadian-protected-conserved-areas-database%2FDatabases%2FProtectedConservedArea_2022.gdb.zip"

Edit - nope.

vsicurl follows the redirects - but gdal requests the file size and the server responds with a 404. Use this config option as a workaround CPL_VSIL_CURL_USE_HEAD=NO

This source should work:
/vsicurl?use_head=no&url=https://data-donnees.az.ec.gc.ca/api/file?path=%2Fspecies%2Fprotectrestore%2Fcanadian-protected-conserved-areas-database%2FDatabases%2FProtectedConservedArea_2022.gdb.zip