Skip to content

Latest commit

 

History

History
110 lines (87 loc) · 3.38 KB

solution-python-geopyspark.md

File metadata and controls

110 lines (87 loc) · 3.38 KB
Raw

openEO Hackathon - Solution Python client / GeoPyspark back-end

Task 1

  • Prerequisites: Python 3.5, pip
  • Clone/Download: openeo-python-client
  • Hint: Use anaconda for python versioning
  • Recommendation: Use Linux based operating system.

Installing and loading the required packages, in your console:

# inside the project directory
pip install -r requirements.txt
pip install -r requirements-dev.txt
pip install --user -e .

Connecting to the openEO GeoPyspark back-end:

import openeo
endpoint = "http://openeo.vgt.vito.be/openeo"
session = openeo.session("me", endpoint=endpoint)

Requesting the capabilities that are provided by the back-end:

session.list_capabilities()

Requesting the supported file formats to see whether GeoTiff (GTiff) or not:

session.get_outputformats()

Task 2

Requesting the processes offered by the back-end:

session.get_all_processes()

Requesting the arguments required for the ndvi process is currently not supported by the GeoPyspark back-end, but the call would be:

# session.get_process('NDVI')

Requesting the products offered by the back-end:

session.list_collections()

Requesting information about the Sentinel-2 dataset, including the temporal and spatial extent, is currently not supported by the GeoPyspark back-end, but the calls would be:

# collection = session.get_collection("COPERNICUS/S2")
# collection["extent"]
# collection["time"]

Task 4

If you haven't done so yet, download the GeoJSON file containing the polygon into the working directory of the Python client.

Construct and execute the process graph. The execute call is synchronous, so it computes the time series on the fly:

from shapely.geometry import asShape
import json

with open("polygon.json","r") as f:
    polygon = asShape(json.load(f))

image_collection = session \
    .imagecollection('PROBAV_L3_S10_TOC_NDVI_333M') \
    .date_range_filter(start_date="2017-11-01", end_date="2017-11-30")

timeseries = image_collection.zonal_statistics(polygon,'mean').execute()

Task 5

If you haven't done so yet, download the UDF file containing Python code to calculate the NDVI to the working directory of the Python client.

Connect to the back-end as described in task 1. Request information about available products and processes as shown in task 2.

After connecting to the back-end, we are building the process graph and downloading the file to disk:

import openeo
import os

bbox = {
    "left": 6.8371137,
    "top": 50.5647147,
    "right": 6.8566699,
    "bottom": 50.560007,
    "srs": "EPSG:4326"
}
time = {
    "start": "2017-10-10",
    "end": "2017-10-30"
}

session = openeo.session("me", endpoint="http://openeo.vgt.vito.be/openeo")
with open("raster_collections_ndvi.py", "r")  as f:
    udf_code = f.read()
    image_collection = session.image("CGS_SENTINEL2_RADIOMETRY_V101") \
                .date_range_filter(start_date=time["start"], end_date=time["end"]) \
                .bbox_filter(left=bbox["left"],right=bbox["right"],bottom=bbox["bottom"],top=bbox["top"],srs=bbox["srs"]) \
                .apply_tiles(udf_code) \
                .max_time() \
                .download("task_5_out.geotiff", "GTIFF")