
Actinia Quickstart¶
Actinia is an open source REST API for scalable, distributed, high performance processing of geographical data that uses GRASS GIS for computational tasks. Actinia provides a REST API to process satellite images, time series of satellite images, raster and vector data.
Contents
Actinia can be used in different ways:
- curl or similar command line tools
- the Postman extension for browsers
- open a GRASS GIS session and use the ace (actinia command execution) tool
- other interfaces to REST API
In this quickstart, we make use of GRASS GIS to conveniently launch commands from the session to the actinia server (which itself uses GRASS GIS). The idea is to rapidly develop a workflow locally on small data sets to then execute it on the server.
Introduction to ace - actinia command execution¶
The ace
tool (details)
allows the execution of a single GRASS GIS command or a
list of GRASS GIS commands on an actinia REST service
(https://actinia.mundialis.de/). In addition it provides job management,
the ability to list locations, mapsets and map layer the user has access
to as well as the creation and deletion of mapsets.
Th ace
tool must be executed in an active GRASS GIS session and will
use the current location of this session to access the actinia service.
The current location setting can be overridden by the
--location LOCATION_NAME
option. All commands will be executed per
default in an ephemeral database. Hence, generated output must be
exported using augmented GRASS commands.
The option --persistent MAPSET_NAME
allows the execution of commands
in the persistent user database. It can be used with
--location LOCATION_NAME
option.
Setup your environment¶
The user must setup the following environmental variables to specify the actinia server and credentials:
# set credentials and REST server URL
export ACTINIA_USER='demouser'
export ACTINIA_PASSWORD='gu3st!pa55w0rd'
export ACTINIA_URL='https://actinia.mundialis.de/latest'
Access sample data¶
Selected datasets are available to the demo user. To list the locations you have access to, run
ace --list-locations
['latlong', 'nc_spm_08', 'utm_32n', 'latlong']
The following command lists mapsets of current location in the active GRASS GIS session (nc_spm_08):
# running ace in the "nc_spm_08" location:
ace --list-mapsets
['PERMANENT', 'landsat']
Access data from external sources¶
GRASS GIS commands can be augmented with actinia specific extensions.
The +
operator can be specified for an input parameter to import a
web located resource and to specify the export of an output parameter.
Importantly, the name of the local location and mapset must correspond to that on the actinia REST server.
Currently available datasets are (organized by projections):
- North Carolina sample dataset (NC State-Plane metric CRS, EPSG:
3358):
- base cartography (
nc_spm_08/PERMANENT
; source: https://grassbook.org/datasets/datasets-3rd-edition/) - Landsat subscenes (
nc_spm_08/landsat
; source: https://grass.osgeo.org/download/sample-data/)
- base cartography (
- Latitude-Longitude location (LatLong WGS84, EPSG:4326):
- empty (
latlong/PERMANENT/
) - 16-days NDVI, MOD13C1, V006, CMG 0.05 deg res.
(
latlong/modis_ndvi_global/
; source: https://lpdaac.usgs.gov/dataset_discovery/modis/modis_products_table/mod13c1_v006) - LST growing degree days asia 2017 (
latlong/asia_gdd_2017/
; source: https://www.mundialis.de/en/temperature-data/) - LST tropical days asia 2017 (
latlong/asia_tropical_2017/
) - LST temperature daily asia 2017, including min, max and avg
(
latlong/asia_lst_daily_2017/
)
- empty (
- Europe (EU LAEA CRS, EPSG:3035):
- EU DEM 25m V1.1 (
eu_laea/PERMANENT/
; source: https://land.copernicus.eu/imagery-in-situ/eu-dem) - CORINE Landcover 2012, g100_clc12_V18_5 (
eu_laea/corine_2012/
; source: https://land.copernicus.eu/pan-european/corine-land-cover/clc-2012)
- EU DEM 25m V1.1 (
- World Mollweide (EPSG 54009):
- GHS_POP_GPW42015_GLOBE_R2015A_54009_250_v1_0
(
world_mollweide/pop_jrc
; source: https://ghsl.jrc.ec.europa.eu/ghs_pop.php)
- GHS_POP_GPW42015_GLOBE_R2015A_54009_250_v1_0
(
Inspect the REST call prior to submission¶
To generate the actinia process chain JSON request simply add the –dry-run flag:
ace --dry-run r.slope.aspect elevation=elevation slope=myslope
Display a map - map rendering¶
It is very easy (and fast) to render a map:
# check amount of pixels, just FYI
ace --location latlong r.info globcover@globcover
ace --location latlong --render-raster globcover@globcover

ESA Globcover map shown by actinia
Ephemeral processing is the default processing approach of actinia. Commands are executed in an ephemeral mapset which will be removed after processing. You can export the output of GRASS GIS modules to store the computational result for download and further analysis. The following export formats are currently supported:
- raster:
GTiff
- vector:
ESRI_Shapefile
,GeoJSON
,GML
- table:
CSV
,TXT
Script examples¶
Example 1: computing slope and aspect and univariate statistics from an elevation model¶
The following commands (to be stored in a script and executed with
ace
) will import a raster layer from an internet source as raster
map elev
, sets the computational region to the map and computes the
slope. Additional information about the raster layer are requested with
r.info
.
Store the following script as text file ace_dtm_statistics.sh
:
# grass77 ~/grassdata/nc_spm_08/user1/
# Import the web resource and set the region to the imported map
g.region raster=elev+https://storage.googleapis.com/graas-geodata/elev_ned_30m.tif -ap
# Compute univariate statistics
r.univar map=elev
r.info elev
# Compute the slope of the imported map and mark it for export as geotiff file
r.slope.aspect elevation=elev slope=slope_elev+GTiff
r.info slope_elev
Save the script in the text file to /tmp/ace_dtm_statistics.sh
and
run the saved script as
ace --script /tmp/ace_dtm_statistics.sh
The results are provided as REST resources.
To generate the actinia process chain JSON request simply add the –dry-run flag
ace --dry-run --script /tmp/ace_dtm_statistics.sh
The output should look like this:
{
"version": "1",
"list": [
{
"module": "g.region",
"id": "g.region_1804289383",
"flags": "pa",
"inputs": [
{
"import_descr": {
"source": "https://storage.googleapis.com/graas-geodata/elev_ned_30m.tif",
"type": "raster"
},
"param": "raster", "value": "elev"
}
]
},
{
"module": "r.univar",
"id": "r.univar_1804289383",
"inputs": [
{"param": "map", "value": "elev"},
{"param": "percentile", "value": "90"},
{"param": "separator", "value": "pipe"}
]
},
{
"module": "r.info",
"id": "r.info_1804289383",
"inputs": [{"param": "map", "value": "elev"}]
},
{
"module": "r.slope.aspect",
"id": "r.slope.aspect_1804289383",
"inputs": [
{"param": "elevation", "value": "elev"},
{"param": "format", "value": "degrees"},
{"param": "precision", "value": "FCELL"},
{"param": "zscale", "value": "1.0"},
{"param": "min_slope", "value": "0.0"}
],
"outputs": [
{
"export": {"format": "GTiff", "type": "raster"},
"param": "slope", "value": "slope_elev"
}
]
},
{
"module": "r.info",
"id": "r.info_1804289383",
"inputs": [{"param": "map", "value": "slope_elev"}]
}
]
}
Example 2: Orthophoto image segmentation with export¶
Store the following script as text file /tmp/ace_segmentation.sh
:
# grass77 ~/grassdata/nc_spm_08/user1/
# Import the web resource and set the region to the imported map
# we apply a trick for the import of multi-band GeoTIFFs:
# install with: g.extension importer
importer raster=ortho2010+https://apps.mundialis.de/workshops/osgeo_ireland2017/north_carolina/ortho2010_t792_subset_20cm.tif
# The importer has created three new raster maps, one for each band in the geotiff file
# stored them in an image group
r.info map=ortho2010.1
r.info map=ortho2010.2
r.info map=ortho2010.3
# Set the region and resolution
g.region raster=ortho2010.1 res=1 -p
# Note: the RGB bands are organized as a group
i.segment group=ortho2010 threshold=0.25 output=ortho2010_segment_25+GTiff goodness=ortho2010_seg_25_fit+GTiff
# Finally vectorize segments with r.to.vect and export as a GeoJSON file
r.to.vect input=ortho2010_segment_25 type=area output=ortho2010_segment_25+GeoJSON
Run the script saved in a text file as
ace --script /tmp/ace_segmentation.sh
The results are provided as REST resources.
Examples for persistent processing¶
GRASS GIS commands can be executed in a user specific persistent
database. The user must create a mapset in an existing location. This
mapsets can be accessed via ace
. All processing results of commands
run in this mapset, will be stored persistently. Be aware that the
processing will be performed in an ephemeral database that will be moved
to the persistent storage using the correct name after processing.
To create a new mapset in the nc_spm_08 location with the name test_mapset the following command must be executed
ace --location nc_spm_08 --create-mapset test_mapset
Run the commands from the statistic script in the new persistent mapset
ace --location nc_spm_08 --persistent test_mapset --script /path/to/ace_dtm_statistics.sh
Show all raster maps that have been created with the script in test_mapset
ace --location nc_spm_08 --persistent test_mapset g.list type=raster mapset=test_mapset
Show information about raster map elev and slope_elev
ace --location nc_spm_08 --persistent test_mapset r.info elev@test_mapset
ace --location nc_spm_08 --persistent test_mapset r.info slope_elev@test_mapset
Delete the test_mapset
ace --location nc_spm_08 --delete-mapset test_mapset
If the active GRASS GIS session has identical location/mapset settings, then an alias can be used to avoid the persistent option in each single command call:
alias acp="ace --persistent `g.mapset -p`"
We assume that in the active GRASS GIS session the current location is nc_spm_08 and the current mapset is test_mapset. Then the commands from above can be executed in the following way:
ace --create-mapset test_mapset
acp --script /path/to/ace_dtm_statistics.sh
acp g.list type=raster mapset=test_mapset
acp r.info elev@test_mapset
acp r.info slope_elev@test_mapset
Things to try¶
Create new locations¶
# create new location
ace --create-location latlon 4326
# create new mapset within location
ace --location latlon --create-mapset user1
Install GRASS GIS addons (extensions)¶
# list existing addons, see also
# https://grass.osgeo.org/grass7/manuals/addons/
ace --location latlon g.extension -l
# install machine learning addon r.learn.ml
ace --location latlon g.extension r.learn.ml
What next?¶
- Visit the actinia website at https://actinia.mundialis.de
- actinia tutorial: https://neteler.gitlab.io/actinia-introduction
- Further reading: Neteler, M., Gebbert, S., Tawalika, C., Bettge, A., Benelcadi, H., Löw, F., Adams, T, Paulsen, H. (2019). Actinia: cloud based geoprocessing. In Proc. of the 2019 conference on Big Data from Space (BiDS‘2019) (pp. 41–44). EUR 29660 EN, Publications Office of the European Union 5, Luxembourg: P. Soille, S. Loekken, and S. Albani (Eds.). (DOI)