Ingest and visualize landsat data

tms.py

import rasterio
import geopyspark as gps
import numpy as np
import matplotlib
import matplotlib.cm as cm
from pyspark import SparkContext


conf = gps.geopyspark_conf(master="local[*]", appName="ingest-example", )
conf.set(key='spark.kryoserializer.buffer.max', value='256m')
conf.set(key='spark.driver.maxResultSize', value="8g")
conf.set(key='spark.driver.memory', value="8g")
pysc = SparkContext(conf=conf)

raster_layer_b3 = gps.geotiff.get(layer_type=gps.LayerType.SPATIAL, uri="file:///landsat_data/LC81070352015218LGN00_B3.TIF")
raster_layer_b5 = gps.geotiff.get(layer_type=gps.LayerType.SPATIAL, uri="file:///landsat_data/LC81070352015218LGN00_B5.TIF")
raster_layer_bqa = gps.geotiff.get(layer_type=gps.LayerType.SPATIAL, uri="file://landsat_data/LC81070352015218LGN00_BQA.TIF")

combined_layer = gps.combine_bands(
    [raster_layer_b3, raster_layer_b5, raster_layer_bqa]
).convert_data_type(new_type=gps.CellType.INT32)
combined_layer.collect_metadata()

tiled_raster_layer = combined_layer.tile_to_layout(gps.GlobalLayout(), target_crs=3857)

for layer  in tiled_raster_layer.pyramid().levels.values():
    gps.write(uri="file:///tmp/bbb", layer_name="japan", tiled_raster_layer=layer)

cm = gps.ColorMap.nlcd_colormap()
layers = []

for zoom in range(10, 13):
   layers.append(gps.query(uri="file:///tmp/bbb",
                           layer_name="japan",
                           layer_zoom=zoom))

    
def render_image(tiles):
    print('here')
    arr = tiles[0].cells[0]
    arr1 = tiles[1].cells[0] 
    res = arr/arr1
    
    
    norm = matplotlib.colors.Normalize(vmin=min(res), vmax=max(res), clip=True)
    mapper = cm.ScalarMappable(norm=norm, cmap=cm.Greys_r)

    return Image.fromarray(mapper.to_rgba(res), mode='RGBA')


landsat_pyramid = gps.Pyramid(layers)

tms = gps.TMS.build(source=landsat_pyramid, display=render_image)
tms.bind(host='0.0.0.0', requested_port=8889)
tms.url_pattern