Command-line interface

A command-line interface is provided for extracting information from data files or authority codes, and for performing coordinate operations. That command-line tool is provided for allowing users to experiment some Apache SIS functionalities without writing Java code. For each example, key Java APIs are listed for allowing users to reproduce the functionality in their own application.

Installation

Select "Apache SIS 0.7 binary" from the downloads page an unzip in any directory. The directory structure will be as below:

apache-sis-0.7
├─ bin
├─ conf
├─ data
├─ lib
└─ log

The bin sub-directory contains a sis command for Unix systems (Linux or MacOS — we have not yet done a sis.bat file for Windows). The bin sub-directory can be added to the PATH environment variable for convenience, but this is not mandatory. Example:

export PATH=$PWD/apache-sis-0.7/bin:$PATH

The remaining of this page assumes that the bin directory is on the search path. If not, the same examples can still be executed by replacing the sis command by ./apache-sis-0.7/bin/sis.

Optional dependencies

Every JAR files present in the lib sub-directory will be included on the classpath during sis command execution. By default, the lib directory contains only the sis.jar file. However users can add other JAR files in that directory for the following optional dependencies:

  • Derby driver — needed only if Apache SIS can not locate a JavaDB driver in the JDK installation directory. This is normally not needed on Oracle JDK6, JDK7 or JDK8 distribution.

  • UCAR NetCDF library — by default, SIS uses its own embedded NetCDF reader which supports only the classical NetCDF format, as standardized by OGC. If there is a need to read files encoded in GRID or HDF formats, then one can copy the UCAR NetCDF library in the lib sub-directory. If presents, the UCAR library should be detected and used automatically.

Usage

Invoking sis without argument shows a summary of available commands and all options. For executing a command, the syntax is:

sis <command> [options] [files]

Available commands are:

help Show a help overview.
about Show information about Apache SIS and system configuration.
mime-type Show MIME type for the given file.
metadata Show metadata information for the given file.
crs Show Coordinate Reference System (CRS) information for the given file or code.
identifier Show identifiers for metadata and referencing systems in the given file.
transform Convert or transform coordinates from given source CRS to target CRS.

The set of legal options and the expected number of file arguments depend on the command being executed. However all commands support the --help option, which lists the options available for that command. Available options will be from the following list:

--sourceCRS The Coordinate Reference System of input data.
--targetCRS The Coordinate Reference System of output data.
--format The output format: xml, wkt, wkt1 or text.
--locale The locale to use for the command output.
--timezone The timezone for the dates to be formatted.
--encoding The encoding to use for the command output.
--colors Whether colorized output shall be enabled.
--brief Whether the output should contains only brief information.
--verbose Whether the output should contains more detailed information.
--debug Prints full stack trace in case of failure.
--help Lists the options available for a specific command.

The --locale, --timezone and --encoding options apply to the command output sent to the standard output stream, but usually do not apply to the error messages sent to the standard error stream. The reason is that command output may be targeted to a client, while the error messages are usually for the operator.

Examples

The following examples first show how to get a Coordinate Reference System (CRS) from different kinds of definitions. CRS contain necessary information for locating points on Earth. Those information include the geodetic datum, the map projection (if any), the axes and their units of measurement. When two such CRS are known (the source and the target), it is possible to convert or transform points between those CRS. However the coordinate operation depends not only on the source and target CRS, but also on the area of interest as shown by the example transforming city coordinates in USA and in Canada. Then another example show how to get a wider set of metadata, with the CRS as only one specific element of those metadata.

Getting ISO 19111 Coordinate Reference Systems (CRS)

Apache SIS can read Coordinate Reference System (CRS) definitions from various sources:

  • Codes from the EPSG geodetic dataset
  • Texts in Well Known Text (WKT) format version 1 and 2
  • XML files conform to Geographic Markup Language (GML) schema version 3.2

The easiest way to see a CRS definition is to use an EPSG code. The EPSG geodetic dataset is a widely-used source of definitions for thousands of Coordinate Reference Systems. Appache SIS provides a list of supported codes, which can be queried from the command line. For example the following command prints the definition of the "JGD2011 / Japan Plane Rectangular CS VIII" Coordinate Reference System:

sis crs EPSG:6676

The first time that a command requires the EPSG dataset, Apache SIS will propose to download and install a local copy. This automatic download happen only when using the command-line tools described in this page. Developers who wish to use the EPSG dataset in their applications can use the Maven artifact documented in download page.

The above command should produce the following output (click to expand):
ProjectedCRS["JGD2011 / Japan Plane Rectangular CS VIII",
  BaseGeodCRS["JGD2011",
    Datum["Japanese Geodetic Datum 2011",
      Ellipsoid["GRS 1980", 6378137.0, 298.257222101]],
    Unit["degree", 0.017453292519943295]],
  Conversion["Japan Plane Rectangular CS zone VIII",
    Method["Transverse Mercator"],
    Parameter["Latitude of natural origin", 36.0],
    Parameter["Longitude of natural origin", 138.5],
    Parameter["Scale factor at natural origin", 0.9999],
    Parameter["False easting", 0.0],
    Parameter["False northing", 0.0]],
  CS[Cartesian, 2],
    Axis["Northing (X)", north],
    Axis["Easting (Y)", east],
    Unit["metre", 1],
  Scope["Large and medium scale topographic mapping, cadastral and engineering survey."],
  Area["Japan - onshore - Honshu between approximately 137°45'E and 139°E - Niigata-ken; Nagano-ken; Yamanashi-ken; Shizuoka-ken."],
  BBox[34.54, 137.32, 38.58, 139.91],
  Id["EPSG", 6676, "8.9", URI["urn:ogc:def:crs:EPSG:8.9:6676"]],
  Remark["Replaces JGD2000 / Japan Plane Rectangular CS VIII (CRS code 2450) with effect from 21st October 2011."]]

The above output is compatible with version 2 of Well Known Text (WKT) format. However some applications are restricted to WKT version 1. For an output using the legacy WKT 1 format, one can add the --format wkt1 option to the above command-line.

The “WKT 2” specification allows some flexibility in keyword names and in the way to specify units of measurement. By default, the command-line uses this flexibility for producing less verbose but still legal WKT 2 output. If an output closer to WKT 2 recommendations is desired, one can add the --format wkt2 option to the above command. The result will be slightly more verbose.

Apache SIS can also read and write CRS definitions in the Geographic Markup Language (GML) format. For GML output, one can add the --format xml option to the above command. The result will be much more verbose than WKT outputs.

Java API for accessing functionalities shown in above examples are:

  • Convenience static methods in org.apache.sis.referencing package:
    • CRS.forCode(String)
    • CRS.fromWKT(String)
    • CRS.fromXML(String)
  • Classes in other packages (less convenient but give more control):
    • org.apache.sis.io.wkt.WKTFormat — control WKT version, syntax highlighting, etc.
    • org.apache.sis.xml.MarshallerPool — control GML version, link resolutions, etc.

Getting a verified EPSG identifier from a Coordinate Reference System

Because the EPSG geodetic dataset is so commonly used, a very common practice is to specify a CRS using only its EPSG code instead than its full definition. Such codes can be written in different ways. For example all the following strings are for the same code:

  • "EPSG:4326"
  • "urn:ogc:def:crs:EPSG::4326"
  • "http://www​.opengis​.net/def/crs/epsg/0/4326" (not yet supported on the command-line)
  • "http://www​.opengis​.net/gml/srs/epsg​.xml#4326" (not yet supported on the command-line)

In a Well Known Text (WKT) string, the code can appear at the bottom of the definition in an optional element like ID["EPSG", 4326] or AUTHORITY["EPSG", "4326"]. However in many cases the code is missing. Sometime Apache SIS can find a missing code by comparing a given CRS against the definitions in the EPSG database.

The following example reads a WKT for the “NTF (Paris) / Lambert zone II” Coordinate Reference System, but in which the ID["EPSG", 27572] element has been intentionally omitted. Furthermore, the “NTF (Paris) / Lambert zone II” name has been replaced by “NTF (Paris) / zone to be discovered by the demo” name. Executing the following command:

sis identifier http://sis.apache.org/examples/crs/MissingIdentifier.wkt

produces an output like below:

    urn:ogc:def:crs:EPSG:8.9:27572    | NTF (Paris) / Lambert zone II

As we can see, Apache SIS has been able to find back the identifier code and the actual CRS name.

Sometime a WKT declares wrongly an EPSG code. The most frequent case is a WKT that defines a Coordinate Reference System with (longitude, latitude) axes, but declare an EPSG code for a CRS with (latitude, longitude) axes. Apache SIS can detect such mismatches. For example executing the following command:

sis identifier http://sis.apache.org/examples/crs/WrongAxisOrder.wkt

produces an output like below:

!   urn:ogc:def:crs:EPSG:8.9:4979    | WGS 84

Legend:
!   Identified object matches definition provided by authority except for coordinate system axes.

Apache SIS can perform such analysis because it “understands” the CRS definition. This analysis capability can be tested by altering the CRS definition. The next example asks the identifier of a CRS which is normally defined as below:

ProjectedCRS["WGS 84 / Mercator 41",
  (... definition omitted for brevity ...)
    Method["Mercator (variant B)"],
    Parameter["Latitude of 1st standard parallel", -41.0],
  (... definition omitted for brevity ...)

However in this example, we will provide a CRS defined as below:

ProjectedCRS["Scaled Mercator",
  (... definition omitted for brevity ...)
    Method["Mercator (variant A)"],
    Parameter["Scale factor at natural origin", 0.7557992272019596"],
    Parameter["Latitude of natural origin", -0.0],
  (... definition omitted for brevity ...)

Executing the following command:

sis identifier http://sis.apache.org/examples/crs/EquivalentDefinition.wkt

produces an output like below:

    urn:ogc:def:crs:EPSG:8.9:3994    | WGS 84 / Mercator 41

In above example, Apache SIS used the fact that a “Mercator (variant A)” projection with a “Scale factor at natural origin” parameter value of 0.755799… on the WGS84 datum is numerically equivalent to a “Mercator (variant B)” projection with a “Latitude of 1st standard parallel” parameter value of 41° on the same datum. This recognition allowed SIS to return the EPSG:3994 code even if it stands for a CRS defined as a “Mercator (variant B)” projection rather than variant A.

Java API for accessing functionalities shown in above example are:

  • Convenience static method in org.apache.sis.referencing package:
    • IdentifiedObjects.lookupURN(IdentifiedObject, Citation)
  • Class in other package (less convenient but give more control):
    • org.apache.sis.referencing.factory.IdentifiedObjectFinder

Performing coordinate conversions or transformations

Coordinates represented in a given CRS can be transformed into coordinates represented in another CRS. The coordinate transformations depend mostly on the source and target CRS, but the area of interest can also have an influence; while optional, that area should be specified when it is known.

The following example transform coordinates from the North American Datum 1927 (EPSG:4267) to WGS84 (EPSG:4326). The example is run twice: once for cities in USA, then once for cities in Canada: (Note: the application may log warnings to the console. Those warnings can be ignored)

wget http://sis.apache.org/examples/coordinates/AmericanCities.csv
wget http://sis.apache.org/examples/coordinates/CanadianCities.csv
sis transform --sourceCRS EPSG:4267 --targetCRS EPSG:4326 AmericanCities.csv
sis transform --sourceCRS EPSG:4267 --targetCRS EPSG:4326 CanadianCities.csv

The first execution should print the following header, followed by transformed coordinate values. Note the operation code (EPSG:1173), domain of validity (United State) and accuracy.

# Source:      NAD27 (EPSG:4267)
# Destination: WGS 84 (EPSG:4326)
# Operations:  NAD27 to WGS 84 (4) (EPSG:1173)
# Domain:      United States (USA) - onshore
# Accuracy:    10.0 metres

The second execution should print the following header, followed by transformed coordinate values. Note that the operation code (EPSG:1172), domain of validity (Canada) and accuracy are not the same than in previous example.

# Source:      NAD27 (EPSG:4267)
# Destination: WGS 84 (EPSG:4326)
# Operations:  NAD27 to WGS 84 (3) (EPSG:1172)
# Domain:      Canada - onshore and offshore
# Accuracy:    20.0 metres

The difference between those two operations become more visible by adding the --verbose option to the above sis transform commands. This option shows the coordinate operation in Well Known Text (WKT) or pseudo-WKT format. When transforming coordinates in USA, the operation contains the following parameter values:

Method["Geocentric translations (geog2D domain)"],
  Parameter["X-axis translation", -8.0, Unit["metre", 1]],
  Parameter["Y-axis translation", 160.0, Unit["metre", 1]],
  Parameter["Z-axis translation", 176.0, Unit["metre", 1]]

But when transforming coordinates in Canada, the operation rather contains the following parameter values:

Method["Geocentric translations (geog2D domain)"],
  Parameter["X-axis translation", -10.0, Unit["metre", 1]],
  Parameter["Y-axis translation", 158.0, Unit["metre", 1]],
  Parameter["Z-axis translation", 187.0, Unit["metre", 1]],

As seen in the above examples, the parameter values differ slightly with the geographic area of the coordinates to transform. Those parameters could also be different if datum shift grids are available. For example in USA:

Method["NADCON"],
  Parameter["Latitude difference file", "conus.las"],
  Parameter["Longitude difference file", "conus.los"],

Java API for accessing functionalities shown in above examples are:

  • Convenience static method in org.apache.sis.referencing package:
    • CRS.findOperation(CoordinateReferenceSystem, CoordinateReferenceSystem, GeographicBoundingBox)
  • Class in other package (less convenient but give more control):
    • org.apache.sis.referencing.operation.DefaultCoordinateOperationFactory

Extracting ISO 19115 Metadata

SIS can read the header of a data file and print the metadata in an ISO 19115 structure. The data file given in argument can be a local file or a URL. By default, the metadata are shown in a relatively compact tree-table format. But the metadata can optionally be exported as an ISO 19139 compliant XML document.

Note: As of SIS 0.7, only CF-compliant NetCDF files or ISO 19139 XML documents can be given in argument to this command.

The following example show the metadata of a NetCDF file accessible from the web:

sis metadata https://github.com/opengeospatial/geoapi/raw/master/geoapi-netcdf/src/test/resources/org/opengis/wrapper/netcdf/NCEP-SST.nc

Below is an extract of the above command output:
Metadata
  ├─Identification info
  │   ├─Citation
  │   │   ├─Title……………………………………………………………………………… Sea Surface Temperature Analysis Model
  │   │   ├─Date
  │   │   │   ├─Date……………………………………………………………………… Sep 22, 2005 2:00:00 AM
  │   │   │   └─Date type………………………………………………………… Creation
  │   │   └─Identifier
  │   │       ├─Code……………………………………………………………………… NCEP/SST/Global_5x2p5deg/SST_Global_5x2p5deg_20050922_0000.nc
  │   │       └─Authority
  │   │           └─Title………………………………………………………… edu.ucar.unidata
  │   ├─Abstract………………………………………………………………………………… NCEP SST Global 5.0 x 2.5 degree model data
  │   ├─Descriptive keywords
  │   │   ├─Keyword………………………………………………………………………… EARTH SCIENCE > Oceans > Ocean Temperature > Sea Surface Temperature
  │   │   ├─Type………………………………………………………………………………… Theme
  │   │   └─Thesaurus name
  │   │       └─Title…………………………………………………………………… GCMD Science Keywords
  │   ├─Resource constraints
  │   │   └─Use limitation……………………………………………………… Freely available
  │   ├─Spatial representation type……………………………… Grid
  │   └─Extent
  │       ├─Geographic element
  │       │   ├─West bound longitude…………………………… 180°W
  │       │   ├─East bound longitude…………………………… 180°E
  │       │   ├─South bound latitude…………………………… 90°S
  │       │   ├─North bound latitude…………………………… 90°N
  │       │   └─Extent type code……………………………………… true
  │       └─Vertical element
  │           ├─Minimum value……………………………………………… 0
  │           └─Maximum value……………………………………………… 0
  └─Content info
      └─Dimension
          ├─Descriptor………………………………………………………………… Sea temperature
          └─Sequence identifier………………………………………… SST

Adding the --format xml option to the above command will format the same metadata in a XML document. The output is not shown in this page because of its verbosity.

Java API for accessing functionalities shown in above examples are:

  • Convenience static methods in org.apache.sis.xml package:
    • XML.marshal(…)
    • XML.unmarshal(…)
  • Class in other package (less convenient but give more control):
    • org.apache.sis.xml.MarshallerPool

Performance consideration

If there is a large amount of files to process, invoking the above command many time may be inefficient since it would restart a new JVM on every invocation. If the operation requires the EPSG dataset, booting the Derby database also has a significant cost. For such cases, it is more efficient to loop inside a small Java program using the SIS API.