Worst TMS EVER

This python flask server, serv TMS data from copernicus raster, its the worst TMS cause the computing time long and requests going in time out.

For now the server take the last image found and create a tile from it.

Caution

Unzip ./generator/sentinel2/utils/data/S2A_OPER.kml.zip before use

Once launch

Server will be launch on port 5000

You can access to data with url like this:

http://localhost:5000/sentinel2/4108/3017/13

http://localhost:5000/sentinel2/{x}/{y}/{z}

The request can take a long time, if the data was not already computed the wtmse will download the data and process the tile.

The processing can be configure by the request:

http://localhost:5000/sentinel2/4108/3017/13?first_clip=500,8500&third_clip=500,8500&second_clip=500,8500&bands=2,2,2

http://localhost:5000/sentinel2/{x}/{y}/{z}?first_clip=500,8500&third_clip=500,8500&second_clip=500,8500

http://localhost:5000/sentinel2/{x}/{y}/{z}?zone=31TCJ&date=20180620

Docker-compose

Use docker-compose for testing:

docker-compose up

Access to the map with:

http://localhost:85

The data are pre-downloaded and work only on few tiles

TODO

• Define the generator requested in the request
• Pass argument to improve the wtmse (date of image, color clip, band used)
• Handle error properly in generators
• Add the possibility to configure generator with config files or env var
• Write readme for real
• Use TCI instead of compose image in no band ask case
• Change sentinel downloader in abstract class and implement local_sentinel_downloader
• Implement amazon_bucket_sentinel_downloader
• Implement google_bucket_sentinel_downloader
• Change thread poll to process poll, because of GIL ()
• Define min and max Z by ENV var instead of hard wrinting in code.
• Looking if decorator is better than sentinel2/init.py way to register in generator factory