An improved TileClient that avoids the REST interface where possible (#…

…156)

* Improve local use

* Docs updates

* BaseTileClient -> BaseTileClientInterface

* Re-enable rest-based local client

* Bump version to 1.0.0

* Fix data paths

MANIFEST.in

@@ -1,5 +1,5 @@
-graft localtileserver/tileserver/static
-graft localtileserver/tileserver/templates
+graft localtileserver/web/static
+graft localtileserver/web/templates
 global-exclude *.pyc
-graft localtileserver/tileserver/data
-exclude localtileserver/tileserver/data/*.aux.xml
+graft localtileserver/tiler/data
+exclude localtileserver/tiler/data/*.aux.xml

doc/source/api/index.rst

@@ -8,7 +8,7 @@ Python Client
 .. autofunction:: localtileserver.get_or_create_tile_client
 
 
-.. autoclass:: localtileserver.client.BaseTileClient
+.. autoclass:: localtileserver.client.BaseTileClientInterface
    :members:
    :undoc-members:
 

doc/source/user-guide/hillshade.rst

@@ -9,12 +9,9 @@ see in a terrain model. Hillshades are often used as an underlay in a map, to
 make the data appear more 3-Dimensional.
 
 
-This example was adopted from `EarthPy <https://earthpy.readthedocs.io/en/latest/gallery_vignettes/plot_dem_hillshade.html>`_
-
-
 .. note::
 
-  This example requires ``rasterio`` to be installed.
+  This example was adopted from `EarthPy <https://earthpy.readthedocs.io/en/latest/gallery_vignettes/plot_dem_hillshade.html>`_
 
 
 .. code:: python

doc/source/user-guide/index.rst

@@ -40,6 +40,20 @@ Here is the "one-liner" to visualize a large geospatial image with
 The :class:`localtileserver.TileClient` class utilizes the ``_ipython_display_``
 method to automatically display the tiles with ``ipyleaflet`` in a Notebook.
 
+You can also get a single tile by:
+
+.. jupyter-execute::
+
+  # z, x, y
+  client.get_tile(10, 163, 395)
+
+
+And get a thumbnail preview by:
+
+.. jupyter-execute::
+
+  client.thumbnail()
+
 
 🍃 ``ipyleaflet`` Tile Layers
 -----------------------------

doc/source/user-guide/rasterio.rst

@@ -21,3 +21,17 @@ This will only work when opening a raster in read-mode.
     m = Map(center=client.center(), zoom=client.default_zoom)
     m.add_layer(t)
     m
+
+
+``localtileserver`` actually uses ``rasterio`` under the hood for everything
+and keeps a reference to a ``rasterio.DatasetReader`` for all clients.
+
+
+.. code-block:: python
+
+    from localtileserver import examples
+
+    # Load example tile layer from publicly available DEM source
+    client = examples.get_elevation()
+
+    client.rasterio

doc/source/user-guide/rgb.rst

@@ -16,6 +16,19 @@ viewing a different set of bands:
   # First, create TileClient using example file
   client = examples.get_landsat()
 
+
+.. jupyter-execute::
+
+  client.thumbnail(band=[7, 5, 4])
+
+
+.. jupyter-execute::
+
+  client.thumbnail(band=[5, 3, 2])
+
+
+.. jupyter-execute::
+
   # Create 2 tile layers from same raster viewing different bands
   l = get_leaflet_tile_layer(client, band=[7, 5, 4])
   r = get_leaflet_tile_layer(client, band=[5, 3, 2])
@@ -49,6 +62,11 @@ See https://girder.github.io/large_image/tilesource_options.html#style
     ]
   }
 
+  client.thumbnail(style=style)
+
+
+.. jupyter-execute::
+
   l = get_leaflet_tile_layer(client, style=style)
 
   m = Map(center=client.center(), zoom=client.default_zoom)

flask.env

@@ -1,2 +1,2 @@
-export FLASK_APP=localtileserver/tileserver/__init__.py
+export FLASK_APP=localtileserver/web/__init__.py
 export FLASK_ENV=development

localtileserver/__init__.py

@@ -2,7 +2,7 @@
 from localtileserver._version import __version__
 from localtileserver.client import RemoteTileClient, TileClient, get_or_create_tile_client
 from localtileserver.report import Report
-from localtileserver.tiler.utilities import get_cache_dir, make_vsi, purge_cache
+from localtileserver.tiler import get_cache_dir, make_vsi, purge_cache
 from localtileserver.validate import validate_cog
 from localtileserver.widgets import (
     LocalTileServerLayerMixin,