The usual Mapnik workflow is to load the map data into a SQL-based database and then define map layers with SQL statements. To render a tile, Mapnik needs to execute a couple of SQL queries. The benefit of this approach is that you don't need to reload your database when you change your map style. You can just change your SQL query and Mapnik picks it up. In our model, we decided to precompute the features we need for each tile. We need to know the map style before we create the database. However, when rendering the map tile, we only fetch the features that we need to render.
-We haven't open sourced the RocksDB Mapnik plugin or the database loading pipeline. However, the spatial indexing is available in RocksDB under a name [SpatialDB](https://github.com/facebook/rocksdb/blob/master/include/rocksdb/utilities/spatial_db.h). The API is focused on map rendering use-case, but we hope that it can also be used for other spatial-based applications.
+We haven't open sourced the RocksDB Mapnik plugin or the database loading pipeline. However, the spatial indexing is available in RocksDB under a name [SpatialDB](https://github.com/facebook/rocksdb/blob/main/include/rocksdb/utilities/spatial_db.h). The API is focused on map rendering use-case, but we hope that it can also be used for other spatial-based applications.
Let's take a tour of the API. When you create a spatial database, you specify the spatial indexes that need to be built. Each spatial index is defined by a bounding box and granularity. For map rendering, we create a spatial index for each zoom levels. Higher zoom levels have more granularity.
projects / ceph.git / blobdiff