Skip to content

Latest commit

 

History

History
54 lines (38 loc) · 1.52 KB

polystore.rst

File metadata and controls

54 lines (38 loc) · 1.52 KB
Raw

TDM Polystore

General idea:

  • all objects are timeseries with annotations:
  • a timeseries is a collection of snapshots, * each snapshot has a timestamp and a geometrical footprint, * plus annotations and provenance information; * a geometrical footprint can go from a point to a volume;
  • use timescaledb + postgis to hold ghosts of the datasets
  • single shot events are

Datacube query interface

latitude = ( 4.5217, 4.5925) longitude = (-71.7926, -71.6944) product_class = "l7" product_instance = "ledaps_meta_river" ingestion_timestamp = xxxx

data = dc.load(latitude=latitude, longitude=longitude, product_class=,
product_instance=, measurements = ['red', 'nir', 'pixel_qa'])

this will return a list of xarray(s)

use case sensor trace

dimensions (latitude:

with dims print( gpm_data )

<xarray.Dataset> Dimensions: (latitude: 3, longitude: 3, time: 366) Coordinates:

  • time (time) datetime64[ns] 2015-01-01T11:59:59.500000 ...
  • latitude (latitude) float64 12.95 12.85 12.75
  • longitude (longitude) float64 14.25 14.35 14.45
Data variables:
total_precipitation (time, latitude, longitude) int32 0 0 0 0 0 0 0 0 ... liquid_precipitation (time, latitude, longitude) int32 0 0 0 0 0 0 0 0 ... ice_precipitation (time, latitude, longitude) int32 0 0 0 0 0 0 0 0 ... percent_liquid (time, latitude, longitude) uint8 15 15 15 15 15 ...
Attributes:
crs: EPSG:4326