geoc

geoc is a geospatial command line application that follows the unix philosophy. Each command does one thing well (buffer a layer, crop a raster) by reading a vector layer as a CSV text stream or a raster layer as an ASCII grid, processing the layer or raster, and then writing out the vector layer as a CSV or a raster layer as an ASCII grid. Individual commands can be chained together with unix pipes.

geoc is very much under development (command names may change). Originally is was developed as a complement to geometry commands and to stress test GeoScript Groovy. The commands have not been optimized for large datasets.

geoc is built on the shoulders of giants: GeoTools and the Java Topology Suite. geoc just provides a command line application that wraps the herculean effort that the developers of these two libraries have undertaken.

Key features

1. Git style commands. One command (geoc) and many subcommands (which can be more than one word).
2. By default geoc reads and writes vector layers as CSV and raster layers as ASCII grids.
3. But geoc can read and write to any supported GeoTools DataStore (Shapefiles, PostGIS, H2) or CoverageStore (GeoTIFF, WorldImage, GTOPO).
4. Commands are looked up using Java's Service Provider Interface (SPI) so the framework is extensible.
5. Where appropriate, values are expressions (literals, properties, or CQL with functions)
6. Uses GeoScript Groovy for extremely terse code.

Examples

List commands:

>>> geoc list

Count features in a CSV layer:

>>> cat states.csv | geoc vector count

Buffer feature from a shapefile:

>>> geoc vector buffer -i earthquakes.shp -o earthquake_buffers.shp

Get the envelope of a layer and then calculate the buffer:

>>> cat states.csv | geoc vector envelope | geoc vector buffer -d 0.1

Crop a raster:

>>> geoc raster crop -i raster.tif -b "-120,-40,120,40" -o raster_croped.tif

Here is a longer example that create 100 random points in a GeoPackage database, get's metadata of that layer, and then finally converts the layer to CSV::

>>> geoc vector randompoints -g "0 0 10 10" -n 100 -o test.gpkg -r points100

>>> geoc vector info -i test.gpkg -l points100

>>> geoc vector to -i test.gpkg -f csv

Data Sources

By default, vector commands read and write CSV using WKT for geometry fields and raster commands read and write ASCII grids. But, geoc can read and write any supported GeoTools DataStore or CoverageStore by using Connection Strings. GeoTools uses connection maps to connect to DataStore's. geoc connection strings are these connection maps where the key/value pairs are separated by an '=' sign and multiple key/value pairs are separated by a white space. Values can be single quoted.

Vector

PostGIS

dbtype=postgis database=postgres host=localhost port=5432 user=postgres passwd=postgres

MySQL

dbtype=mysql database=layers host=localhost port=5432 user=me passwd=s$cr$t

H2

test.db

dbtype=h2 database=test.db

dbtype=h2 host=localhost port=5432 schema=public user=me password=s$cr$t

dbtype=h2 jndiReferenceName=layers schema=public

Shapefile

url=data/states.shp

data/states.shp

Memory

memory

Properties

data/states.properties

directory=data/properties

GeoPackage

layers.gpkg

database=layers.gpkg dbtype=geopkg user=me passwd=s$cr$t

Geobuf

layer.pbf

file=layers precision=6 dimension=2

Spatialite

layers.sqlite

dbtype=spatialite database=layers.sqlite

OGR

DatasourceName=states.shp DriverName='ESRI Shapefile' namespace=shp

WFS

http://geoserver.org/wfs?request=getcapabilities

Raster

Raster sources are currently all file based.

data/earth.tif

world.png

Tile

pyramid

Several tile layers can take a pyramid attribute. You can use one of several well known pyramid names:

• globalmercator
• mercator
• globalmercatorbottomleft
• globalgeodetic
• geodetic

or use a file that contains pyramid metadata in csv, xml, or json format.

mbtiles

type=mbtiles file=states.mbtiles

type=mbtiles file=states.mbtiles name=states description='The united states'

states.mbtiles

geopackage

type=geopackage file=states.gpkg name=states pyramid=globalmercator

states.gpkg

tms

type=tms file=/Users/you/tms format=jpeg

type=tms file=/Users/you/tms format=png name=tms pyramid=geodetic

osm

type=osm url=http://a.tile.openstreetmap.org

type=osm urls=http://a.tile.openstreetmap.org,http://b.tile.openstreetmap.org

utfgrid

type=utfgrid file=/Users/me/tiles/states

vectortiles

type=vectortiles name=states file=/Users/me/tiles/states format=mvt pyramid=GlobalMercator

type=vectortiles name=states url=http://vectortiles.org format=pbf pyramid=GlobalGeodetic

Map Layer

Map layer strings contain a layertype, layername, layerprojection, and style properties.

layertype

• layer
• raster
• tile

For layer layertype, you can use the same key value pairs used to specify a Workspace.

For raster layertype, you specify a source=file key value pair.

For tile layertype, you use the same key value pairs used to specify a tile layer.

layername

The name of the layer

style

A SLD or CSS File

Examples

layertype=layer dbtype=geopkg database=/Users/user/Desktop/countries.gpkg layername=countries style=/Users/user/Desktop/countries.sld

layertype=layer file=/Users/user/Desktop/geoc/polygons.csv layername=polygons style=/Users/user/Desktop/geoc/polygons.sld

layertype=layer file=/Users/user/Desktop/geoc/points.properties style=/Users/user/Desktop/geoc/points.sld

layertype=layer file=/Users/user/Projects/geoc/src/test/resources/polygons.shp

layertype=layer directory=/Users/user/Projects/geoc/src/test/resources/points.properties layername=points

layertype=raster source=rasters/earth.tif

layertype=tile file=world.mbtiles

layertype=tile type=geopackage file=states.gpkg

Installation

Just download the latest release and put the geoc/bin directory on your path. geoc also requires Java 8.

If you want to use the OGR Workspace, you need to install the GDAL/OGR native library compiled with JNI support and then set the GEOC_GDAL_HOME variable.

Getting Help

Each command contains a --help option:

>>> geoc vector buffer --help
geoc vector buffer: Buffer the features of the input Layer and save them to the output Layer
--help                      : Print the help message
-c (--capstyle) VAL         : The cap style
-d (--distance) VAL         : The buffer distance
-i (--input-workspace) VAL  : The input workspace
-l (--input-layer) VAL      : The input layer
-o (--output-workspace) VAL : The output workspace
-q (--quadrantsegments) N   : The number of quadrant segments
-r (--output-layer) VAL     : The output layer
-s (--singlesided)          : Whether buffer should be single sided or not

There is also a man page for each subcommand:

>>> man geoc-vector-buffer
geoc-vector-buffer(1)                                    geoc-vector-buffer(1)

NAME
       geoc vector buffer

DESCRIPTION
       Buffer  the  features  of  the  input Layer and save them to the output
       Layer

USAGE
       geoc vector randompoints -n 10 -g "1,1,10,10" | geoc vector  buffer  -d
       10

OPTIONS
       -d --distance: The buffer distance

       -q --quadrantsegments: The number of quadrant segments

       -s --singlesided: Whether buffer should be single sided or not

       -c --capstyle: The cap style

       -o --output-workspace: The output workspace

Finally, there is a bash completion script which makes using geoc with bash much easier.

Install it is your .bash_profile:

source /Users/You/geoc/shell/geoc_bash_comp

Build

Building geoc is very easy but you will need Java 8 and Maven 3.

Check it out:

git checkout https://github.com/jericks/geoc.git

Build it:

cd geoc
mvn clean install

License

geoc is open source and licensed under the MIT License.