documentation.yaml

"extensionName": |-
  gis
"icon": |2-
  i-lucide-map-pin
"markdownTemplate": |2

  # GIS Extension for NetLogo

  This package contains the NetLogo GIS extension.

  {{> BUILDING.md}}

  {{> USING.md}}

  ## Primitives

  {{#contents}}
  ### {{fullCategoryName}}

  {{#prims}}
  [`{{name}}`](#{{primitive.extensionName}}{{primitive.name}})
  {{/prims}}

  {{/contents}}

  {{#primitives}}
  {{> primTemplate}}
  {{/primitives}}

  {{> LICENSE.md}}
"additionalVariables":
  "netlogoUrl": |-
    http://ccl.northwestern.edu/netlogo/docs/
"primTemplate": |2

  ### `{{name}}`

  ```NetLogo
  {{#examples}}
  {{#isOptional}}({{/isOptional}}{{primitive.fullName}}{{#args}} *{{argumentPlaceholder}}*{{/args}}{{#isOptional}}){{/isOptional}}
  {{/examples}}
  ```

  {{{description}}}
"filesToIncludeInManual":
- |-
  USING.md
- |-
  primitives
"tableOfContents":
  "coord": |-
    Coordinate System Primitives
  "data": |-
    Dataset Primitives
  "vector": |-
    VectorDataset Primitives
  "raster": |-
    RasterDataset Primitives
  "drawing": |-
    Drawing Primitives
"primitives":
- "name": |-
    set-transformation
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      gis-envelope
    "type": |-
      list
  - "name": |-
      netlogo-envelope
    "type": |-
      list
  "description": |2

    Defines a mapping between GIS coordinates and NetLogo coordinates.
    The *gis-envelope* and *netlogo-envelope* parameters must
    each be four-element lists consisting of:

    ```
    [minimum-x maximum-x minimum-y maximum-y]
    ```

    The scale of the transformation will be equal to the minimum of the
    scale necessary to make the mapping between the ranges of x values
    and the scale necessary to make the mapping between the ranges of y
    values. The GIS space will be centered in NetLogo space.

    For example, the following two lists would map all of geographic
    (latitude and longitude) space in degrees to NetLogo world space,
    regardless of the current dimensions of the NetLogo world:

    ```
    (list -180 180 -90 90)
    (list min-pxcor max-pxcor min-pycor max-pycor)
    ```

    However, if you're setting the envelope of the NetLogo world,
    you should probably be using [[gis:set-world-envelope|gis#set-world-envelope]].
- "name": |-
    set-transformation-ds
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      gis-envelope
    "type": |-
      list
  - "name": |-
      netlogo-envelope
    "type": |-
      list
  "description": |2

    Does the same thing as [set-transformation](#set-transformation) above, except that
    it allows the scale for mapping the range of x values to be
    different than the scale for y values. The "-ds" on the
    end stands for "different scales". Using different scales
    will cause distortion of the shape of GIS features, and so it is
    generally not recommended, but it may be useful for some models.

    Here is an example of the difference between [set-transformation](#set-transformation) and [set-transformation-ds](#set-transformation-ds):

    <table width="80%" border="1" rules="cols" style="border-collaps: separate; border-spacing: 4px; text-align: center; margin: 0 auto;">
    <tr>
      <td style="padding: 8px">
        <img alt="" src="images/set-transformation.png" width="200">
      <td style="padding: 8px">
        <img alt="" src="images/set-transformation-ds.png" width="200">
      <tr>
      <td style="padding: 8px; white-space: pre;">

      Using <a href="#set-transformation">set-transformation</a>,
      the scale along the x and y axis is the same, preserving the
      round shape of the Earth in this Orthographic projection.

      <td style="padding: 8px; white-space: pre;">

      Using <a href="#set-transformation-ds">set-transformation-ds</a>, the
      scale along the x axis is stretched so that the earth covers
      the entire NetLogo View, which in this case distorts the shape
      of the Earth.

    </table>
- "name": |-
    set-world-envelope
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      gis-envelope
    "type": |-
      list
  "description": |2

    A shorthand for setting the transformation by mapping the envelope
    of the NetLogo world to the given envelope in GIS space, while
    keeping the scales along the x and y axis the same. It is
    equivalent to:

    ```
    set-transformation gis-envelope (list min-pxcor max-pxcor min-pycor max-pycor)
    ```

    This primitive is supplied because most of the time you'll want
    to set the envelope of the entire NetLogo world, rather than just a
    part of it.
- "name": |-
    set-world-envelope-ds
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      gis-envelope
    "type": |-
      list
  "description": |2

    A shorthand for setting the transformation by mapping the envelope
    of the NetLogo world to the given envelope in GIS space, using
    different scales along the x and y axis if necessary. It is
    equivalent to:

    ```
    set-transformation-ds gis-envelope (list min-pxcor max-pxcor min-pycor max-pycor)
    ```

    See the [pictures](#transformation-example) above for
    the difference between using equal scales for x and y coordinates
    and using different scales.
- "name": |-
    world-envelope
  "tags":
  - |-
    coord
  "type": |-
    reporter
  "returns": |-
    list
  "description": |2+

    Reports the envelope (bounding rectangle) of the NetLogo world,
    transformed into GIS space. An envelope consists of a four-element
    list of the form:

    ```
    [minimum-x maximum-x minimum-y maximum-y]
    ```

- "name": |-
    envelope-of
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      thing
    "type": |-
      anything
  "description": |2+

    Reports the envelope (bounding rectangle) of *thing* in GIS
    coordinates. The *thing* may be an Agent, an AgentSet, a
    RasterDataset, a VectorDataset, or a VectorFeature. An envelope
    consists of a four-element list of the form:

    ```
    [minimum-x maximum-x minimum-y maximum-y]
    ```

- "name": |-
    envelope-union-of
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      envelope1
    "type": |-
      list
  - "name": |-
      envelope2
    "type": |-
      list
  "alternateArguments":
  - "name": |-
      envelope1...
    "type": |-
      repeatable list
  "description": |2

    Reports an envelope (bounding rectangle) that entirely contains the
    given envelopes. An envelope consists of a four-element list of the
    form

    ```
    [minimum-x maximum-x minimum-y maximum-y]
    ```

    No assumption is made about the coordinate system of the arguments,
    though if they are not in the same coordinate system, results will
    be unpredictable.
- "name": |-
    load-coordinate-system
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      file
    "type": |-
      string
  "description": |2

    Loads a new global projection used for projecting or re- projecting
    GIS data as it is loaded from a file. The file must contain a valid
    <a href="https://www.geoapi.org/3.0/javadoc/org/opengis/referencing/doc-files/WKT.html" target="_blank">
    Well-Known Text (WKT)</a> projection description.

    WKT projection files are frequently distributed alongside GIS data
    files, and usually have a ".prj" filename extension.

    Relative paths are resolved relative to the location of the current
    model, or the user's home directory if the current model
    hasn't been saved yet.

    The GIS extension does not support all WKT coordinate systems and
    projections. Only geographic (`"GEOGCS"`) and
    projected (`"PROJCS"`) coordinate systems are
    supported. For projected coordinate systems, only the following
    projections are supported:

    * Albers_Conic_Equal_Area
    * Lambert_Conformal_Conic_2SP
    * Polyconic
    * Lambert_Azimuthal_Equal_Area
    * Mercator_1SP
    * Robinson
    * Azimuthal_Equidistant
    * Miller
    * Stereographic
    * Cylindrical_Equal_Area
    * Oblique_Mercator
    * Transverse_Mercator
    * Equidistant_Conic
    * hotine_oblique_mercator
    * Gnomonic
    * Orthographic

    See <a href="http://remotesensing.org/geotiff/proj_list/" target="_blank">remotesensing.org</a>
    for a complete list of WKT projections and their parameters.
- "name": |-
    set-coordinate-system
  "type": |-
    command
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      system
    "type": |-
      string or list
  "description": |2

    Sets the global projection used for projecting or re- projecting
    GIS data as it is loaded. The *system* must be either a string
    in <a href="http://www.geoapi.org/3.0/javadoc/org/opengis/referencing/doc-files/WKT.html" target="_blank">
    Well-Known Text (WKT) format</a>, or a NetLogo list that consists
    of WKT converted to a list by moving each keyword inside its
    associated brackets and putting quotes around it. The latter is
    preferred because it makes the code much more readable.

    The same limitations on WKT support apply as described above in the
    documentation for [load-coordinate-system](#load-coordinate-system)
- "name": |-
    project-lat-lon
  "type": |-
    reporter
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      latitude
    "type": |-
      number
  - "name": |-
      longitude
    "type": |-
      number
  "returns": |-
    list
  "description": |2

    Report the position, in NetLogo space, of the given latitude
    and longitude pair according to the current map projection and
    transformation.

    Like the [[gis:location-of|gis#location-of]] primitive, the reported xcor and ycor
    values are reported in a two-item list of `[xcor ycor]`
    and an empty list if the specified point is outside of
    the bounds of the netlogo world.
    For instance:
    ```
    let location-of-abbey-road-studios gis:project-lat-lon 51.5320787 -0.1802646
    let abbey-road-xcor item 0 location-of-abbey-road-studios
    let abbey-road-ycor item 1 location-of-abbey-road-studios
    ```

    Note that this primitive assumes that the given lat/lon pair
    are relative to the WGS84 datum/ellipsoid. If your
    data is based on GPS observations or GeoJson files, then your
    data is already relative to WGS84. If you are unsure about
    what datum your data is, then you should probably just assume
    it is WGS84 and use this primitive. However, if you do know
    that your data is relative to a different datum and that
    extra degree of precision is important to you (if you are,
    say, comparing values from location-of and project-lat-lon)
    then you should use [[gis:project-lat-lon-from-ellipsoid|gis#project-lat-lon-from-ellipsoid]] and
    specify the desired datum's ellipsoid.
- "name": |-
    project-lat-lon-from-ellipsoid
  "type": |-
    reporter
  "tags":
  - |-
    coord
  "arguments":
  - "name": |-
      latitude
    "type": |-
      number
  - "name": |-
      longitude
    "type": |-
      number
  - "name": |-
      ellipsoid-radius
    "type": |-
      number
  - "name": |-
      ellipsoid-inverse-flattening
    "type": |-
      number
  "returns": |-
    list
  "description": |2

    Report the position, in NetLogo space, of the given latitude
    and longitude pair according to the current map projection and
    transformation and the given ellipsoid parameters.

    Like the [[gis:location-of|gis#location-of]] primitive, the reported xcor and ycor
    values are reported in a two-item list of `[xcor ycor]`
    and an empty list if the specified point is outside of
    the bounds of the netlogo world.

    The two defining parameters of a  ellipsoid for
    the purposes of this primitive are the radius and the
    inverse flattening metric. These parameters can be
    easily found by examining either the WKT definition
    of a given projection/datum pair or the .prj file for
    the desired datum. For example, if you open the .prj file
    for a shapefile exported with the WGS66 datum in a text editor,
    you will see, somewhere in the file, this bit of text:
    `DATUM["D_WGS_1966",SPHEROID["NWL_9D",6378145,298.25]]`.
    If you look at the `SPHEROID` section of that text, the
    first number is the radius of that ellipoid and the
    second is the inverse flattening.

    Once we have these numbers, we can project data that is
    relative to WGS66 like so:
    ```
    let location gis:project-lat-lon my-lat my-lon 6378145 298.25
    ```

    For more on earth ellipoids, see: https://en.wikipedia.org/wiki/Earth_ellipsoid
- "name": |-
    load-dataset
  "type": |-
    reporter
  "reports": |-
    Raster dataset
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      file
    "type": |-
      string
  "description": |2-
    Loads the given data file, re-projecting the data as necessary.

    Relative paths are resolved relative to the location of the current
    model, or the user's home directory if the current model
    hasn't been saved yet.

    For ESRI shapefiles and ESRI grid files, if there is a ".prj" file
    associated with the file, then [[gis:load-dataset|gis#load-dataset]] will consult that file
    and re-project to the current global projection if needed. If no ".prj"
    file is found, then the data is assumed to use the same projection as
    the current global coordinate system.

    For GeoJSON files, as per the most-recent specification (RFC 7946),
    the coordinate system for GeoJSON files is always WGS84 and will be
    imported accordingly.

    Currently, three types of data file are supported:

    * "**.shp**" (ESRI shapefile): contains vector data,
      consisting of points, lines, or polygons. When the target file is a
      shapefile, [[gis:load-dataset|gis#load-dataset]] reports a VectorDataset.
    * "**.asc**" or "**.grd**" (ESRI ASCII grid):
      contains raster data, consisting of a grid of values. When
      the target file is an ASCII grid file, [[gis:load-dataset|gis#load-dataset]]
      reports a RasterDataset.
    * "**.geojson**" or "**.json**" (GeoJSON): contains vector data
      similar to shapefiles and similarly reports a VectorDataset.

    Note that not all aspects of the GeoJSON standard are supported.
    In particular, to be properly imported, a GeoJSON file must
    satisfy the following:

    * It only contain numeric or string data within the properties.
      all other json data will be stringified.
    * All "Features" within a "FeatureCollection" must be of the same
      shape type ("Point", "LineString", etc.) Additionally, if not all
      the "Features" in the "FeatureCollection" have the same set of
      property names, default values will be supplied where there
      are missing entries (0 for numbers and "" for strings.)
    * It must not use "GeometryCollection", which is not supported
    Elevation/Z data is partially supported. For both .geojson and
    .shp files, single points with Z data will have their Z
    coordinate moved to a new "_Z" property that can be accessed
    with [[gis:property-value|gis#property-value]] like any other property. Any Z
    information for other shape types will be discarded upon import.
- "name": |-
    store-dataset
  "type": |-
    command
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      dataset
    "type": |-
      Dataset
  - "name": |-
      file
    "type": |-
      string
  "description": |2-
    Exports the given dataset to the given file.

    For VectorDatasets, two file formats are supported, ESRI shapefiles
    and GeoJSON. If the given file name ends in ".geojson" or ".json",
    then the file will be exported as a GeoJSON file. If the file name
    ends in any other extension or no extension at all, the dataset
    will be exported as a shapefile and the associated file extensions
    will be supplied (".shp", ".prj", etc.)

    For RasterDatasets, only ESRI ASCII grid files are supported and
    the associated file extensions will be automatically supplied.

    Relative paths are resolved relative to the location of the current
    model, or the user's home directory if the current model hasn't
    been saved yet.
- "name": |-
    type-of
  "type": |-
    reporter
  "returns": |-
    string
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      dataset
    "type": |-
      Dataset
  "description": |-
    Reports the type of the given GIS dataset: either "VECTOR" or "RASTER"
- "name": |-
    patch-dataset
  "type": |-
    reporter
  "reports": |-
    raster
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      patch-variable
    "type": |-
      reference
  "description": |2

    Reports a new raster whose cells correspond directly to NetLogo
    patches, and whose cell values consist of the values of the given
    patch variable. This primitive is basically the inverse of [[gis:apply-raster|gis#apply-raster]].
    [[gis:apply-raster|gis#apply-raster]] copies values from a raster dataset to a patch variable, while this
    primitive copies values from a patch variable to a raster dataset.
- "name": |-
    turtle-dataset
  "type": |-
    reporter
  "returns": |-
    dataset
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      turtle-set
    "type": |-
      turtleset
  "description": |2

    Reports a new, point VectorDataset built from the turtles in the
    given agentset. The points are located at locations of the turtles,
    translated from NetLogo space into GIS space using the current
    coordinate transformation. And the dataset's properties consist
    of all of the turtle variables common to every turtle in the
    agentset.
- "name": |-
    link-dataset
  "type": |-
    reporter
  "returns": |-
    dataset
  "tags":
  - |-
    data
  "arguments":
  - "name": |-
      link-set
    "type": |-
      linkset
  "description": |2

    Reports a new, line VectorDataset built from the links in the given
    agentset. The endpoints of each line are at the location of the
    turtles connected by each link, translated from NetLogo space into
    GIS space using the current coordinate transformation. And the
    dataset's properties consist of all of the link variables
    common to every link in the agentset.
- "name": |-
    shape-type-of
  "type": |-
    reporter
  "returns": |-
    string
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  "description": |2-

    Reports the shape type of the given dataset. The possible output
    values are "POINT", "LINE", and "POLYGON".
- "name": |-
    property-names
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  "description": |2

    Reports a list of strings where each string is the name of a
    property possessed by each VectorFeature in the given
    VectorDataset, suitable for use in [gis:property-value](#property-value).
- "name": |-
    feature-list-of
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  "description": |-
    Reports a list of all VectorFeatures in the given dataset.
- "name": |-
    vertex-lists-of
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  "description": |2

    Reports a list of lists of Vertex values. For point datasets, each
    vertex list will contain exactly one vertex: the location of a
    point. For line datasets, each vertex list will contain at least
    two points, and will represent a "polyline", connecting
    each adjacent pair of vertices in the list. For polygon datasets,
    each vertex list will contain at least three points, representing a
    polygon connecting each vertex, and the first and last vertices in
    the list will be the same.
- "name": |-
    centroid-of
  "type": |-
    reporter
  "returns": |-
    vertex
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  "description": |2

    Reports a single Vertex representing the centroid (center of
    gravity) of the given feature. For point datasets, the centroid is
    defined as the average location of all points in the feature. For
    line datasets, the centroid is defined as the average of the
    locations of the midpoints of all line segments in the feature,
    weighted by segment length. For polygon datasets, the centroid is
    defined as the weighted sum of the centroids of a decomposition of
    the area into (possibly overlapping) triangles. See <a href="http://www.faqs.org/faqs/graphics/algorithms-faq/" target="_blank">this FAQ</a>
    for more details on the polygon centroid algorithm.
- "name": |-
    random-point-inside
  "type": |-
    reporter
  "returs": |-
    vertex
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  "description": |2-
    Reports a single randomly-generated Vertex that lies within the
    given feature polygon. Generated points are uniformly
    distributed within the polygon and both polygon holes and
    multi-polygon features are supported.
    ```
    ; create 100 turtles randomly distributed throughout a VectorFeature [[vf]]
    crt 100 [
      let loc gis:location-of gis:random-point-inside vf
      set xcor item 0 loc
      set ycor item 1 loc
    ]
    ```
- "name": |-
    location-of
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      Vertex
    "type": |-
      vertex
  "description": |2

    Reports a two-element list containing the x and y values (in that
    order) of the given vertex translated into NetLogo world space
    using the current transformation, or an empty list if the given
    vertex lies outside the NetLogo world.
- "name": |-
    set-property-value
  "tags":
  - |-
    vector
  "type": |-
    command
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      value
    "type": |-
      string or number
  "description": |2-
    Sets the value of the given property of the given VectorFeature. The
    type of the given value (string or number) must match the property
    type of the VectorFeature. This command may be used in conjunction
    with store-dataset to make changes to VectorFeatures and export
    them back as GIS datasets.
- "name": |-
    property-value
  "tags":
  - |-
    vector
  "type": |-
    reporter
  "returns": |-
    anything
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  - "name": |-
      property-name
    "type": |-
      string
  "description": |2

    Reports the value of the property with the given name for the given
    VectorDataset. The reported value may be a number, a string, or a
    boolean value, depending on the type of the field in the underlying
    data file.

    For shapefiles, values from dBase `CHARACTER` and
    `DATE` fields are returned as strings, values from
    `NUMBER` and `FLOAT` fields are returned as numbers,
    and values from `LOGICAL` fields are returned as boolean
    values. `MEMO` fields are not supported. `DATE`
    values are converted to strings using ISO 8601 format
    (`YYYY-MM-DD`).
- "name": |-
    find-features
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      specified-value
    "type": |-
      string or number
  "description": |2

    Reports a list of all VectorFeatures in the given dataset
    whose value for the property *property-name* matches
    *specified-value* (a string or number).

    For strings, value comparison is not case sensitive, and the
    wildcard character "*" will match any number of occurrences
    (including zero) of any character.

    For numbers, beware that there are some numbers that are too
    big to be represented as an integer in NetLogo, so integer
    data imported into NetLogo may start to lose precision
    if they are larger than around 10^14.
- "name": |-
    find-one-feature
  "type": |-
    reporter
  "returns": |-
    feature
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      specified-value
    "type": |-
      string or number
  "description": |2

    Reports the first VectorFeature in the dataset whose value
    for the property *property-name* matches the given string or
    number.  Reports `nobody` if no matching VectorFeature is
    found.

    For strings, Value comparison is not case sensitive, and the
    wildcard character "*" will match any number of occurrences
    (including zero) of any character.  Features are searched in
    the order that they appear in the data file that was the
    source of the dataset, and searching stops as soon as a
    match is found.

    For numbers, beware that there are some numbers that are too
    big to be represented as an integer in NetLogo, so if you
    want to be able to identify features based on a unique ID,
    keep the IDs you use on the relatively small side to play it
    safe. Any number that can fit into a 32 bit unsigned integer
    will work, but any larger than 10^14 and you could run into
    issues. (Helpfully, the NetLogo app will warn you if you try
    to type one of these too-large numbers into the editor if
    you want to check.)
- "name": |-
    find-less-than
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      value
    "type": |-
      string or number
  "description": |2

    Reports a list of all VectorFeatures in the given dataset whose
    value for the property *property-name* is less than the given
    *value*. String values are compared using case-sensitive
    lexicographic order as defined in the <a href="http://docs.oracle.com/javase/1.5.0/docs/api/java/lang/String.html#compareTo(java.lang.String)" target="_blank">
    Java Documentation</a>. Using a string value for a numeric property
    or a numeric value for a string property will cause an error.
- "name": |-
    find-greater-than
  "tags":
  - |-
    vector
  "type": |-
    reporter
  "returns": |-
    list
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      value
    "type": |-
      string or number
  "description": |2

    Reports a list of all VectorFeatures in the given dataset whose
    value for the property *property-name* is greater than the
    given *value*. String values are compared using case-sensitive
    lexicographic order as defined in the <a href="http://docs.oracle.com/javase/1.5.0/docs/api/java/lang/String.html#compareTo(java.lang.String)" target="_blank">
    Java Documentation</a>. Using a string value for a numeric property
    or a numeric value for a string property will cause an error.
- "name": |-
    find-range
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      minimum-value
    "type": |-
      number or string
  - "name": |-
      maximum-value
    "type": |-
      number or string
  "description": |2

    Reports a list of all VectorFeatures in the given dataset whose
    value for the property *property-name* is strictly greater
    than *minimum-value* and strictly less than
    *maximum-value*. String values are compared using
    case-sensitive lexicographic order as defined in the <a href="http://docs.oracle.com/javase/1.5.0/docs/api/java/lang/String.html#compareTo(java.lang.String)" target="_blank">
    Java Documentation</a>. Using a string value for a numeric property
    or a numeric value for a string property will cause an error.
- "name": |-
    property-minimum
  "type": |-
    reporter
  "returns": |-
    number or string
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  "description": |2

    Reports the smallest value for the given property over all of the
    VectorFeatures in the given dataset. String values are compared
    using case-sensitive lexicographic order as defined in the <a href="http://docs.oracle.com/javase/1.5.0/docs/api/java/lang/String.html#compareTo(java.lang.String)" target="_blank">
    Java Documentation</a>.
- "name": |-
    property-maximum
  "tags":
  - |-
    vector
  "type": |-
    reporter
  "returns": |-
    number or string
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  "description": |2

    Reports the largest value for the given property over all of the
    VectorFeatures in the given dataset. String values are compared
    using case-sensitive lexicographic order as defined in the <a href="http://docs.oracle.com/javase/1.5.0/docs/api/java/lang/String.html#compareTo(java.lang.String)" target="_blank">
    Java Documentation</a>.
- "name": |-
    apply-coverage
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      property-name
    "type": |-
      string
  - "name": |-
      patch-variable
    "type": |-
      reference
  "description": |2

    Copies values from the given property of the VectorDataset's
    features to the given patch variable. The dataset must be a
    `polygon` dataset; points and lines are not supported.

    For each patch, it finds all VectorFeatures that intersect that
    patch. Then, if the property is a string property, it computes the
    majority value by computing the total area of the patch covered by
    VectorFeatures having each possible value of the property, then
    returning the value which represents the largest proportion of the
    patch area. If the property is a numeric property, it computes a
    weighted average of property values from all VectorFeatures which
    intersect the patch, weighted by the proportion of the patch area
    they cover.

    There are two exceptions to this default behavior:

    * If a percentage of a patches' area greater than the
      coverage-maximum-threshold is covered by a single VectorFeature,
      then the property value from that VectorFeature is copied directly.
      If more than one VectorFeature covers a percentage of area greater
      than the threshold, only the first will be used.

    * If the total percentage of a patches' area covered by
      VectorFeatures is less than the coverage-minimum-threshold, the
      target patch variable is set to Not A Number.

    By default, the minimum threshold is 10% and the maximum threshold
    is 33%. These values may be modified using the four primitives that
    follow.
- "name": |-
    create-turtles-from-points
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      breed
    "type": |-
      breed
  - "name": |-
      commands
    "type": |-
      commandBlock
  "description": |-
    For each point in a VectorDataset of points, create a turtle of
    the specified breed at the point's location. For each agent variable
    (as defined in `<breeds>-own`), if there is a property with the same
    name in the dataset, set that variable's value to be the value of
    that property. Finally, execute any commands in the optional
    command block. To use generic turtles as the chosen breed, simply
    supply `turtles` as the breed argument.

    Property names and variable names are compared case-insensitively.
    Keep in mind that when importing shapefiles, property names may be
    modified for backwards compatibility reasons. The names given
    by [[gis:property-names|gis#property-names]] can always be trusted as authoritative. For
    manually specifying a mapping between property names and variable
    names, see the [[create-turtles-from-points-manual|gis#create-turtles-from-points-manual]] primitive.

    For multi-point datasets, a turtle is created at each point of
    multi-point feature, each with the same set of variable values.

    Built-in variables such as "label" and "heading" are supported.
    NetLogo color numeric representations are supported for setting
    "color" and "label-color", as well as the 15 default color string
    representations ("red", "blue", "black", etc.).

    As an example: say you wanted to create a turtle of breed
      "cities/city" for each city in a dataset of cities
    like the one included in the "GIS General Examples" model from
    the models library. The cities dataset has four properties,
    "NAME", "COUNTRY", "POPULATION", and "CAPITAL". To map them all
    to NetLogo turtle variables and set their shapes to circles, you
    could do this:

    ```
    extensions [gis]
    breed [cities city]
    cities-own [name country population capital]
    globals [cities-dataset]

    to setup
      set cities-dataset gis:load-dataset "cities.shp"
      gis:create-turtles-from-points cities-dataset cities [
        set shape "circle"
      ]
    end
    ```
- "name": |-
    create-turtles-from-points-manual
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorDataset
    "type": |-
      dataset
  - "name": |-
      breed
    "type": |-
      breed
  - "name": |-
      property-mapping
    "type": |-
      list
  - "name": |-
      commands
    "type": |-
      commandBlock
  "description": |-
    Like [[gis:create-turtles-from-points|gis#create-turtles-from-points]], creates a turtle for each point in a
    VectorDataset of points and populates their variables with the values of
    corresponding gis properties.

    This primitive can be used to specify
    additional mappings between gis property names and NetLogo variable names.
    These mappings are specified as a list of lists of strings like so:
    `[["property-name" "turtle-variable-name"] ["property-name" "turtle-variable-name"] (etc.)]`

    These manual mappings modify the automatic mapping process that takes
    place in the [[gis:create-turtles-from-points|gis#create-turtles-from-points]] primitive, so you only need
    to specify the changes you want to make to the default mappings, and
    the rest of the mappings will be untouched.

    To return to the cities example from the
    create-turtles-from-points entry, the variable name "capital" is not very
    descriptive. something like "is-capital?" fits the NetLogo style much better.
    To make that change, you would modify the example like so.

    ```
    extensions [gis]
    breed [cities city]
    cities-own [name country population is-capital?]
    globals [cities-dataset]

    to setup
      set cities-dataset gis:load-dataset "cities.shp"
      ;; Since we only want to change how the "CAPITAL" property is mapped, we only need to specify that one change.
      gis:create-turtles-from-points-manual cities-dataset cities [["CAPITAL" "is-capital?"]] [
        set shape "circle"
      ]
      ;; Each city turtle still has a name, country, and population set just like the non-manual version.
    end
    ```

- "name": |-
    create-turtles-inside-polygon
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  - "name": |-
      breed
    "type": |-
      breed
  - "name": |-
      n
    "type": |-
      number
  - "name": |-
      commands
    "type": |-
      commandBlock
  "description": |-
    Randomly create "n" turtles of the given breed within
    the given VectorFeature and for each agent variable
    (as defined in `<breeds>-own`), if there is a property with the same
    name in the dataset, set that variable's value to be the value of
    that property. Finally, execute any commands in the optional
    command block. To use generic turtles as the chosen breed, simply
    supply [[turtles]] as the breed argument.

    Property names and variable names are compared case-insensitively.
    Keep in mind that when importing shapefiles, property names may be
    modified for backwards compatibility reasons. The names given by
    [[gis:property-names|gis#property-names]] can always be trusted as authoritative. For
    manually specifying a mapping between property names and variable
    names, see the [[gis:create-turtles-inside-polygon-manual|gis#create-turtles-inside-polygon-manual]] primitive.

    Built-in variables such as "label" and "heading" are supported.
    NetLogo color numeric representations are supported for setting
    "color" and "label-color", as well as the 15 default color string
    representations ("red", "blue", "black", etc.).

    As an example: say you had a VectorDataset of polygons representing
    different zip codes within a state and you want to create 100
    different turtles within each zip code and have each turtle
    know which zip code it originated in. The VectorDataset
    has a field named "zip", so you should add a variable
    named "zip" to the turtles with [[turtles-own]]. Then,
    loop through each VectorFeature in the VectorDataset and use
    the [[gis:create-turtles-inside-polygon|gis#create-turtles-inside-polygon]] primitive to create 100
    new turtles.

    ```
    extensions [gis]
    globals [dataset]
    turtles-own [zip]

    to setup
      set dataset gis:load-dataset "dataset.shp"
      gis:set-world-envelope envelope-of dataset
      gis:set-drawing-color red
      gis:draw dataset 1

      foreach gis:feature-list-of dataset [ this-vector-feature ->
        gis:create-turtles-inside this-vector-feature turtles 100 [
          set shape "person"
        ]
      ]
    end
    ```
- "name": |-
    create-turtles-inside-polygon-manual
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      VectorFeature
    "type": |-
      feature
  - "name": |-
      breed
    "type": |-
      breed
  - "name": |-
      n
    "type": |-
      number
  - "name": |-
      property-mapping
    "type": |-
      list
  - "name": |-
      commands
    "type": |-
      commandBlock
  "description": |2-
    Like [[gis:create-turtles-inside-polygon|gis#create-turtles-inside-polygon]], creates "n" different turtles within the given VectorFeature and populates their
    agent variables with values corresponding to the property values of the VectorFeature.

    This primitive can be used to specify
    additional mappings between gis property names and NetLogo variable names.
    These mappings are specified as a list of lists of strings like so:
    `[["property-name" "turtle-variable-name"] ["property-name" "turtle-variable-name"] (etc.)]`

    These manual mappings modify the automatic mapping process that takes
    place in the [[gis:create-turtles-inside-polygon|gis#create-turtles-inside-polygon]] primitive, so you only need
    to specify the changes you want to make to the default mappings, and
    the rest of the mappings will be untouched.

    See the [[gis:create-turtles-from-points-manual|gis#create-turtles-from-points-manual]] entry for an example of
    how to override default mappings with manual ones.
- "name": |-
    coverage-minimum-threshold
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    vector
  "description": |-
    Reports the current coverage minimum threshold used by [[gis:apply-coverage|gis#apply-coverage]].
- "name": |-
    set-coverage-minimum-threshold
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      new-threshold
    "type": |-
      number
  "description": |-
    Sets the current coverage minimum threshold to be used by [[gis:apply-coverage|gis#apply-coverage]].
- "name": |-
    coverage-maximum-threshold
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    vector
  "description": |-
    Reports the current coverage maximum threshold used by [[gis:apply-coverage|gis#apply-coverage]].
- "name": |-
    set-coverage-maximum-threshold
  "type": |-
    command
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      new-threshold
    "type": |-
      number
  "description": |-
    Sets the current coverage maximum threshold to be used by [[gis:apply-coverage|gis#apply-coverage]].
- "name": |-
    intersects?
  "type": |-
    reporter
  "returns": |-
    boolean
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      x
    "type": |-
      anything
  - "name": |-
      y
    "type": |-
      anything
  "description": |2

    Reports true if the given objects' spatial representations
    share at least one point in common, and false otherwise. The
    objects x and y may be any one of:

    * a VectorDataset, in which case the object's spatial
        representation is the union of all the points, lines, or polygons
        the dataset contains.
    * a VectorFeature, in which case the object's spatial
        representation is defined by the point, line, or polygon the
        feature contains.
    * A turtle, in which case the spatial representation is a point.
    * A link, whose spatial representation is a line segment
      connecting the two points represented by the turtles the link is
      connecting.
    * A patch, whose spatial representation is a rectangular polygon.
    * An agentset, whose spatial representation is the union of the
      representations of all of the agents it contains.
    * A list containing of any of the items listed here, including
      another list. The spatial representation of such a list is the
      union of the spatial representations of its contents.
- "name": |-
    contains?
  "type": |-
    reporter
  "returns": |-
    boolean
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      x
    "type": |-
      anything
  - "name": |-
      y
    "type": |-
      anything
  "description": |2

    Reports true if every point of *y*'s spatial
    representation is also a part of *x*'s spatial
    representation. Note that this means that polygons do contain their
    boundaries. The objects x and y may be any one of

    * a VectorDataset, in which case the object's spatial
      representation is the union of all the points, lines, or polygons
      the dataset contains.
    * a VectorFeature, in which case the object's spatial
      representation is defined by the point, line, or polygon the
      feature contains.
    * A turtle, in which case the spatial representation is a point.
    * A link, whose spatial representation is a line segment
      connecting the two points represented by the turtles the link is
      connecting.
    * A patch, whose spatial representation is a rectangular polygon.
    * An agentset, whose spatial representation is the union of the
      representations of all of the agents it contains.
    * A list containing of any of the items listed here, including
      another list. The spatial representation of such a list is the
      union of the spatial representations of its contents.
- "name": |-
    contained-by?
  "type": |-
    reporter
  "returns": |-
    boolean
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      x
    "type": |-
      anything
  - "name": |-
      y
    "type": |-
      anything
  "description": |2

    Reports true if every point of *x*'s spatial
    representation is also a part of *y*'s spatial
    representation. The objects x and y may be any one of:

    * a VectorDataset, in which case the object's spatial
      representation is the union of all the points, lines, or polygons
      the dataset contains.
    * a VectorFeature, in which case the object's spatial
      representation is defined by the point, line, or polygon the
      feature contains.
    * A turtle, in which case the spatial representation is a point.
    * A link, whose spatial representation is a line segment
      connecting the two points represented by the turtles the link is
      connecting.
    * A patch, whose spatial representation is a rectangular polygon.
    * An agentset, whose spatial representation is the union of the
      representations of all of the agents it contains.
    * A list containing of any of the items listed here, including
      another list. The spatial representation of such a list is the
      union of the spatial representations of its contents.
- "name": |-
    have-relationship?
  "type": |-
    reporter
  "returns": |-
    boolean
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      x
    "type": |-
      anything
  - "name": |-
      y
    "type": |-
      anything
  "description": |2

    Reports true if the spatial representations of the two objects have
    the given spatial relationship, and false otherwise. The spatial
    relationship is specified using a **Dimensionally Extended Nine-
    Intersection Model (DE-9IM)** matrix. The matrix consists of 9
    elements, each of which specifies the required relationship between
    the two objects' interior space, boundary space, or exterior
    space. The elements must have one of six possible values:

    * "T", meaning the spaces must intersect in some way
    * "F", meaning the spaces must not intersect in any way
    * "0", meaning the dimension of the spaces'
      intersection must be zero (i.e., it must be a point or non-empty
      set of points).
    * "1", meaning the dimension of the spaces'
      intersection must be one (i.e., it must be a line or non-empty set
      of line segments).
    * "2", meaning the dimension of the spaces'
      intersection must be two (i.e., it must be a polygon or set of
      polygons whose area is greater than zero).
    * "*", meaning that the two spaces may have any
      relationship.

    For example, this matrix:
          <table width="50%" border="1" style="text-align: center; margin: 0 auto;">
    <tr>
      <td rowspan="2" colspan="2">
      <td colspan="3">
        x
      <tr>
      <td> Interior <td> Boundary <td> Exterior <tr>
      <td rowspan="3">
        y
      <td> Interior <td> T <td> * <td> * <tr>
      <td> Boundary <td> * <td> * <td> * <tr>
      <td> Exterior <td> F <td> F <td> * </table>

    would return true if and only if some part of object *x*'s
    interior lies inside object *y*'s interior, and no part of
    object *x*'s interior or boundary intersects object
    *y*'s exterior. This is essentially a more restrictive
    form of the `contains?` primitive; one in which polygons are
    not considered to contain their boundaries.

    The matrix is given to the `have-relationship?` primitive as
    a string, whose elements are given in the following order:
          <table width="25%" border="1" style="text-align: center; margin: 0 auto;" align="center">
    <tr>
      <td> 1 <td> 2 <td> 3 <tr>
      <td> 4 <td> 5 <td> 6 <tr>
      <td> 7 <td> 8 <td> 9 </table>

    So to use the example matrix above, you would write:

    ```
    gis:have-relationship? x y "T*****FF*"
    ```

    A much more detailed and formal description of the DE-9IM matrix
    and the associated point-set theory can be found in the <a href="http://www.opengeospatial.org/standards/sfs" target="_blank">OpenGIS Simple
    Features Specification for SQL</a>.

    The objects x and y may be any one of:

    * a VectorDataset, in which case the object's spatial
      representation is the union of all the points, lines, or polygons
      the dataset contains.
    * a VectorFeature, in which case the object's spatial
      representation is defined by the point, line, or polygon the
      feature contains.
    * A turtle, in which case the spatial representation is a point.
    * A link, whose spatial representation is a line segment
      connecting the two points represented by the turtles the link is
      connecting.
    * A patch, whose spatial representation is a rectangular polygon.
    * An agentset, whose spatial representation is the union of the
      representations of all of the agents it contains.
    * A list containing of any of the items listed here, including
      another list. The spatial representation of such a list is the
      union of the spatial representations of its contents.
- "name": |-
    relationship-of
  "type": |-
    reporter
  "returns": |-
    string
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      x
    "type": |-
      anything
  - "name": |-
      y
    "type": |-
      anything
  "description": |2

    Reports the **Dimensionally Extended Nine-Intersection Model
    (DE-9IM)** matrix that describes the spatial relationship of the
    two objects. The matrix consists of 9 elements, each of which
    describes the relationship between the two objects' interior
    space, boundary space, or exterior space. Each element will
    describe the dimension of the intersection of two spaces, meaning
    that it may have one of four possible values:

    * "-1", meaning the spaces do not intersect
    * "0", meaning the dimension of the spaces'
      intersection is zero (i.e., they intersect at a point or set of
      points).
    * "1", meaning the dimension of the spaces'
      intersection is one (i.e., they intersect along one or more lines).
    * "2", meaning the dimension of the spaces'
      intersection is two (i.e., their intersection is a non-empty
      polygon).


    For example, the two polygons x and y shown here:
    <img alt="" src="images/intersecting-polygons.png">

    have the following DE-9IM matrix:
    <table width="50%" border="1" style="text-align: center; margin: 0 auto;" align="center">
      <tr>
      <td rowspan="2" colspan="2">
      <td colspan="3">
        x
      <tr>
      <td> Interior <td> Boundary <td> Exterior <tr>
      <td rowspan="3">
        y
      <td> Interior <td> 2 <td> 1 <td> 2 <tr>
      <td> Boundary <td> 1 <td> 0 <td> 1 <tr>
      <td> Exterior <td> 2 <td> 1 <td> 2 </table>

    Which would be reported by the `relationship-of` primitive
    as the string "212101212".

    A much more detailed and formal description of the DE-9IM matrix
    and the associated point-set theory can be found in the <a href="http://www.opengeospatial.org/standards/sfs" target="_blank">OpenGIS Simple
    Features Specification for SQL</a>.

    The objects x and y may be any one of:

    * a VectorDataset, in which case the object's spatial
      representation is the union of all the points, lines, or polygons
      the dataset contains.
    * a VectorFeature, in which case the object's spatial
      representation is defined by the point, line, or polygon the
      feature contains.
    * A turtle, in which case the spatial representation is a point.
    * A link, whose spatial representation is a line segment
      connecting the two points represented by the turtles the link is
      connecting.
    * A patch, whose spatial representation is a rectangular polygon.
    * An agentset, whose spatial representation is the union of the
      representations of all of the agents it contains.
    * A list containing of any of the items listed here, including
      another list. The spatial representation of such a list is the
      union of the spatial representations of its contents.
- "name": |-
    intersecting
  "type": |-
    reporter
  "returns": |-
    agentset
  "infix": !!bool |-
    true
  "tags":
  - |-
    vector
  "arguments":
  - "name": |-
      patch-set
    "type": |-
      patchset
  - "name": |-
      data
    "type": |-
      anything
  "description": |2

    Reports a new agent set containing only those members of the given
    agent set which intersect given GIS *data*, which may be any
    one of: a VectorDataset, a VectorFeature, an Agent, an Agent Set,
    or a list containing any of the above.
- "name": |-
    width-of
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  "description": |2

    Reports the number of columns in the dataset. Note that this is the
    number of cells from left to right, not the width of the dataset in
    GIS space.
- "name": |-
    height-of
  "tags":
  - |-
    raster
  "type": |-
    reporter
  "returns": |-
    number
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  "description": |2

    Reports the number of rows in the dataset. Note that this is the
    number of cells from top to bottom, not the height of the dataset
    in GIS space.
- "name": |-
    raster-value
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      x
    "type": |-
      number
  - "name": |-
      y
    "type": |-
      number
  "description": |2

    Reports the value of the given raster dataset in the given cell.
    Cell coordinates are numbered from left to right, and from top to
    bottom, beginning with zero. So the upper left cell is (0, 0), and
    the bottom right cell is (`gis:width-of dataset` - 1,
    `gis:height-of dataset` - 1).
- "name": |-
    set-raster-value
  "type": |-
    command
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      x
    "type": |-
      number
  - "name": |-
      y
    "type": |-
      number
  - "name": |-
      value
    "type": |-
      number
  "description": |2

    Sets the value of the given raster dataset at the given cell to a
    new value. Cell coordinates are numbered from left to right, and
    from top to bottom, beginning with zero. So the upper left cell is
    (0, 0), and the bottom right cell is (`gis:width-of dataset`
    - 1, `gis:height-of dataset` - 1).
- "name": |-
    minimum-of
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  "description": |-
    Reports the highest value in the given raster dataset.
- "name": |-
    maximum-of
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  "description": |-
    Reports the lowest value in the given raster dataset.
- "name": |-
    sampling-method-of
  "type": |-
    reporter
  "returns": |-
    string
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  "description": |2

    Reports the sampling method used to compute the value of the given
    raster dataset at a single point, or over an area smaller than a
    single raster cell. Sampling is performed by the GIS extension
    primitives [[gis:raster-sample|gis#raster-sample]], [[gis:resample|gis#resample]], [[gis:convolve|gis#convolve]],
    and [[gis:apply-raster|gis#apply-raster]]. The sampling
    method will be one of the following:

    * `"NEAREST_NEIGHBOR"`: the value of the cell
      nearest the sampling location is used.
    * `"BILINEAR"`: the value of the four nearest
      cells are sampled by linear weighting, according to their
      proximity to the sampling site.
    * `"BICUBIC"`: the value of the sixteen nearest
      cells are sampled, and their values are combined by weight
      according to a piecewise cubic polynomial recommended by Rifman
      (see *Digital Image Warping*, George Wolberg, 1990, pp
      129-131, IEEE Computer Society Press).
    * `"BICUBIC_2"`: the value is sampled using the
      same procedure and the same polynomial as with `BICUBIC`
      above, but using a different coefficient. This method may produce
      somewhat sharper results than `BICUBIC`, but that result
      is data dependent.

    For more information on these sampling methods and on raster
    sampling in general, see <a href="https://en.wikipedia.org/wiki/Image_scaling" target="_blank">this wikipedia
    article</a>.
- "name": |-
    set-sampling-method
  "type": |-
    command
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      sampling-method
    "type": |-
      string
  "description": |2

    Sets the sampling method used by the given raster dataset at a
    single point, or over an area smaller than a single raster cell.
    Sampling is performed by the GIS extension primitives [[gis:raster-sample|gis#raster-sample]], [[gis:resample|gis#resample]], [[gis:convolve|gis#convolve]],
    and [[gis:apply-raster|gis#apply-raster]]. The sampling
    method must be one of the following:

    *  `"NEAREST_NEIGHBOR"`
    *  `"BILINEAR"`
    *  `"BICUBIC"`
    *  `"BICUBIC_2"`

    See [[gis:sampling-method-of|gis#sampling-method-of]] above
    for a more specific description of each sampling method.
- "name": |-
    raster-sample
  "type": |-
    reporter
  "returns": |-
    number
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      sample-location
    "type": |-
      anything
  "description": |2

    Reports the value of the given raster over the given location. The
    location may be any of the following:

    * A list of length 2, which is taken to represent a point in
      netlogo space (`[xcor ycor]`) of the sort reported by
      [[gis:location-of|gis#location-of]] Vertex. The raster
      dataset is sampled at the point of that location.
    * A list of length 4, which is taken to represent an envelope in
      GIS space, of the sort reported by [[envelope-of|gis#envelope-of]]. The raster dataset is sampled
      over the area of that envelope.
    * A patch, in which case the raster dataset is sampled over the
      area of the patch.
    * A turtle, in which case the raster dataset is sampled at the
      location of that turtle.
    * A Vertex, in which case the raster dataset is sampled at the
      location of that Vertex.

    If the requested location is outside the area covered by the raster
    dataset, this primitive reports the special value representing
    "not a number", which is printed by NetLogo as
    "NaN". Using the special "not a number" value
    as an argument to primitives that expect a number may cause an
    error, but you can test the value reported by this primitive to
    filter out "not a number" values. A value that is not a
    number will be neither less than nor greater than a number value,
    so you can detect "not a number" values using the
    following:

    ```
    let value gis:raster-sample dataset turtle 0
    ; set color to blue if value is a number, red if value is "not a number"
    ifelse (value <= 0) or (value >= 0)
    [ set color blue ]
    [ set color red ]
    ```

    If the requested location is a point, the sample is always computed
    using the method set by [[gis:set-sampling-method|gis#set-sampling-method]]. If the
    requested location is an area (i.e., an envelope or patch), the
    sample is computed by taking the average of all raster cells
    covered by the requested area.
- "name": |-
    raster-world-envelope
  "type": |-
    reporter
  "returns": |-
    list
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      x
    "type": |-
      number
  - "name": |-
      y
    "type": |-
      number
  "description": |2

    Reports the GIS envelope needed to match the boundaries of NetLogo
    patches with the boundaries of cells in the given raster dataset.
    This envelope could then be used as an argument to [[set-transformation-ds|gis#set-transformation-ds]].

    There may be more cells in the dataset than there are patches in
    the NetLogo world. In that case, you will need to select a subset
    of cells in the dataset by specifying which cell in the dataset you
    want to match with the upper-left corner of the NetLogo world.
    Cells are numbered from left to right, and from top to bottom,
    beginning with zero. So the upper left cell is (0, 0), and the
    bottom right cell is (`gis:width-of dataset` - 1, `gis:height-of dataset` - 1).
- "name": |-
    create-raster
  "tags":
  - |-
    raster
  "type": |-
    reporter
  "returns": |-
    dataset
  "arguments":
  - "name": |-
      width
    "type": |-
      number
  - "name": |-
      height
    "type": |-
      number
  - "name": |-
      envelope
    "type": |-
      list
  "description": |2

    Creates and reports a new, empty raster dataset with the given
    number of columns and rows, covering the given envelope.
- "name": |-
    resample
  "type": |-
    reporter
  "returns": |-
    dataset
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      envelope
    "type": |-
      list
  - "name": |-
      width
    "type": |-
      number
  - "name": |-
      height
    "type": |-
      number
  "description": |2

    Reports a new dataset that consists of the given RasterDataset
    resampled to cover the given envelope and to contain the given
    number of columns and rows. If the new raster's cells are
    smaller than the existing raster's cells, they will be
    resampled using the method set by [[gis:set-sampling-method|gis#set-sampling-method]]. If the new
    cells are larger than the original cells, they will be sampled
    using the `"NEAREST_NEIGHBOR"` method.
- "name": |-
    convolve
  "type": |-
    reporter
  "returns": |-
    dataset
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      kernel-rows
    "type": |-
      number
  - "name": |-
      kernel-columns
    "type": |-
      number
  - "name": |-
      kernel
    "type": |-
      list
  - "name": |-
      key-column
    "type": |-
      number
  - "name": |-
      key-row
    "type": |-
      number
  "description": |2

    Reports a new raster whose data consists of the given raster
    convolved with the given kernel.

    A convolution is a mathematical operation that computes each output
    cell by multiplying elements of a kernel with the cell values
    surrounding a particular source cell. A kernel is a matrix of
    values, with one particular value defined as the "key
    element", the value that is centered over the source cell
    corresponding to the destination cell whose value is being
    computed.

    The values of the kernel matrix are given as a list, which
    enumerates the elements of the matrix from left to right, top to
    bottom. So the elements of a 3-by-3 matrix would be listed in the
    following order:
          <table width="25%" border="1" style="text-align: center; margin: 0 auto;" align="center">
    <tr>
      <td> 1 <td> 2 <td> 3 <tr>
      <td> 4 <td> 5 <td> 6 <tr>
      <td> 7 <td> 8 <td> 9 </table>

    The key element is specified by column and row within the matrix.
    Columns are numbered from left to right, beginning with zero. Rows
    are numbered from top to bottom, also beginning with zero. So, for
    example, the kernel for the horizontal <a href="https://en.wikipedia.org/wiki/Sobel_operator" target="_blank">Sobel operator</a>,
    which looks like this:
          <table width="25%" border="1" style="text-align: center; margin: 0 auto;" align="center">
    <tr>
      <td> 1 <td> 0 <td> -1 <tr>
      <td> 2 <td> 0 <br> <small>(key)</small> <td> -2 <tr>
      <td> 1 <td> 0 <td> -1 </table>

    would be specified as follows:

    ```
    let horizontal-gradient gis:convolve dataset 3 3 [1 0 -1 2 0 -2 1 0 -1] 1 1
    ```
- "name": |-
    apply-raster
  "type": |-
    command
  "tags":
  - |-
    raster
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      patch-variable
    "type": |-
      reference
  "description": |2


    Copies values from the given raster dataset to the given patch
    variable, resampling the raster as necessary so that its cell
    boundaries match up with NetLogo patch boundaries. This resampling
    is done as if using [[gis:resample|gis#resample]] rather
    than [[gis:raster-sample|gis#raster-sample]], for the sake
    of efficiency. However, patches not covered by the raster are
    assigned values of "not a number" in the same way that
    [[gis:raster-sample|gis#raster-sample]] reports values for
    locations outside the raster.
- "name": |-
    drawing-color
  "tags":
  - |-
    drawing
  "type": |-
    reporter
  "returns": |-
    list or number
  "description": |2

    Reports the color used by the GIS extension to draw vector features
    into the NetLogo drawing layer. Color can be represented either as
    a NetLogo color (a single number between zero and 140) or an RGB
    color (a list of 3 numbers). See details in the [Colors](/programming.html#colors) section of the
    Programming Guide.
- "name": |-
    set-drawing-color
  "type": |-
    command
  "tags":
  - |-
    drawing
  "arguments":
  - "name": |-
      color
    "type": |-
      list or number
  "description": |2

    Sets the color used by the GIS extension to draw vector features
    into the NetLogo drawing layer. *Color* can be represented
    either as a NetLogo color (a single number between zero and 140) or
    an RGB color (a list of 3 numbers). See details in the [Colors](/programming.html#colors) section of the Programming Guide.
- "name": |-
    draw
  "type": |-
    command
  "tags":
  - |-
    drawing
  "arguments":
  - "name": |-
      vector-data
    "type": |-
      anything
  - "name": |-
      line-thickness
    "type": |-
      number
  "description": |2

    Draws the given vector data to the NetLogo drawing layer, using the
    current GIS drawing color, with the given line thickness. The data
    may consist either of an entire VectorDataset, or a single
    VectorFeature. This primitive draws only the boundary of polygon
    data, and for point data, it fills a circle with a radius equal to
    the line thickness.
- "name": |-
    fill
  "type": |-
    command
  "tags":
  - |-
    drawing
  "arguments":
  - "name": |-
      vector-data
    "type": |-
      anything
  - "name": |-
      line-thickness
    "type": |-
      number
  "description": |2

    Fills the given vector data in the NetLogo drawing layer using the
    current GIS drawing color, using the given line thickness around
    the edges. The data may consist either of an entire VectorDataset,
    or a single VectorFeature. For point data, it fills a circle with a
    radius equal to the line thickness.
- "name": |-
    paint
  "type": |-
    command
  "tags":
  - |-
    drawing
  "arguments":
  - "name": |-
      RasterDataset
    "type": |-
      dataset
  - "name": |-
      transparency
    "type": |-
      number
  "description": |2

    Paints the given raster data to the NetLogo drawing layer. The
    highest value in the dataset is painted white, the lowest is
    painted in black, and the other values are painted in shades of
    gray scaled linearly between white and black.

    The *transparency* input determines how transparent the new
    image in the drawing will be. Valid inputs range from 0 (completely
    opaque) to 255 (completely transparent).
- "name": |-
    import-wms-drawing
  "type": |-
    command
  "tags":
  - |-
    drawing
  "arguments":
  - "name": |-
      server-url
    "type": |-
      string
  - "name": |-
      spatial-reference
    "type": |-
      string
  - "name": |-
      layers
    "type": |-
      string
  - "name": |-
      transparency
    "type": |-
      string
  "description": |2

    Imports an image into the NetLogo drawing layer using the
    <a href="http://www.opengeospatial.org/standards/wms" target="_blank">Web Mapping Service</a>
     protocol, as defined by the <a href="http://www.opengeospatial.org/" target="_blank">
    Open Geospatial Consortium</a>.

    The *spatial reference* and *layers* inputs should be
    given as strings. The *spatial reference* input corresponds to
    the **SRS** parameter to the **GetMap** request as defined in
    section 7.2.3.5 of version 1.1.1 of the WMS standard. The
    *layers* input corresponds to the **LAYERS** parameter to
    the as defined in 7.2.3.3 of version 1.1.1 of the WMS standard.

    You can find the list of valid spatial reference codes and layer
    names by examining the response to a **GetCapabilities** request
    to the WMS server. Consult the relevant standard for instructions
    on how to issue a **GetCapabilities** request to the server and
    how to interpret the results.

    The *transparency* input determines how transparent the new
    image in the drawing will be. Valid inputs range from 0 (completely
    opaque) to 255 (completely transparent).