Syntax

   @PointCloudInterpretation(CONVERT [,<component>,
      <dest type>,<conversion type>,<round-off direction>]+)

   @PointCloudInterpretation(SCALE [,<component>,<source type>,
      <dest type>,<conversion type>,<round-off direction>]+)

   @PointCloudInterpretation([REJECTABLE], [<mode>], [,<component>,<source type>,
      <dest type>,<conversion type>,<round-off direction>]+)

Arguments

   REJECTABLE

      Specifies if the function will supply a rejection code and message
      to the invalid feature and output it to the rejected port, or will
      fail upon being supplied invalid features.

      A feature can be rejected for the following reasons:
         - The feature has invalid geometry
         - The pointcloud is missing the specific component

   <component> 
   
      The component to scale.

   <source type>   
   
      The source type of the component.

      Range: REAL64, REAL32, UINT64, INT64, UINT32, INT32, UINT16, INT16, UINT8, 
      INT8  
      
   <dest type>  
   
      The destination type of the component.

      Range: REAL64, REAL32, UINT64, INT64, UINT32, INT32, UINT16, INT16, UINT8, 
      INT8, STRING
      
   <conversion type> 
   
      The conversion method to go from the source type to the destination type. 
      
      CAST - uses C-style casts.

      BOUNDEDCAST - Uses C-style casts, but also validates that the source
      values fit into the destination type, effectively preventing underflow
      and overflow. If a source value does not fit, the corresponding
      destination value will either be set to its type's minimum or maximum
      value.

      TYPESCALE - Scales the source values while preserving all proportions
      in regard to the source and destination types' range. This is typically
      used when converting color components, e.g. from UInt8 to UInt16.

      DATASCALE - Finds the minimum and maximum values of the source values
      and uses them to scale the values to the full range of the destination
      type.

      DATASCALEPRESERVE - Similar to DATASCALE, but also sets a scale and
      offset on the destination component so that the original values are
      preserved. This may be useful when you have floating point values but
      need to convert them to an integer type, while still preserving the
      original values. For example, this could be used when reading Real64
      coordinates (e.g. from the POINTCLOUDXYZ reader) and writing to a
      format that only supports integer coordinates (e.g. the LAS writer).

      Range: CAST, BOUNDEDCAST, TYPESCALE, DATASCALE, DATASCALEPRESERVE

   <round-off direction>   
   
      The action to take when converting from floating-point values to integer. 
      ROUND rounds the value to the nearest integer. CEIL gets the next integer 
      which is greater than or equal to the floating-point value. FLOOR gets the 
      next integer which is less than or equal to the floating-point value.

      Range: ROUND, CEIL, FLOOR  

Description

   The @PointCloudInterpretation function is used to alter the data type and 
   component values of point cloud components.

   CONVERT mode

   In CONVERT mode, the data type of the component is changed.

   Multiple components can be converted with one call to this function. The set
   of four parameters (<component>,<dest type>,<conversion type>,
   <round-off direction>) must be specified for each component. The following
   example will scale the red, green, and blue component values to uint16:

   @PointCloudInterpretation (CONVERT, 
      color_red,uint16,TYPESCALE,ROUND,
      color_green,uint16,TYPESCALE,ROUND,
      color_blue,uint16,TYPESCALE,ROUND)

   SCALE mode

   In SCALE mode, the underlying data type of the component is not changed, but 
   the component values are scaled to reflect the intended data type range.

   For example, consider the color component. The color point cloud component is 
   always stored as UInt16, which has a range of 0-65535. However, in practice, 
   some datasets will only have values in the UInt8 range (0-255), despite being 
   stored in a UInt16. When combining point clouds from different sources, if 
   the color values do not use a consistent range, e.g. one point cloud uses 
   UInt16 color values and another uses UInt8 color values, then the points from 
   the UInt8 cloud may appear black when visualized because the color values are 
   so small in comparison. This function can be used to resolve this 
   inconsistency.

   Multiple components can be scaled with one call to this function. The set of 
   five parameters (<component>,<source data type>,<dest data type>,<conversion 
   type>,<round-off direction>) must be specified for each component. The 
   following example will scale the red, green, and blue component values from 
   uint8 to uint16:

      @PointCloudInterpretation (SCALE, 
         color_red,uint8,uint16,TYPESCALE,ROUND,
         color_green,uint8,uint16,TYPESCALE,ROUND,
         color_blue,uint8,uint16,TYPESCALE,ROUND)

   Note that in this mode, STRING is not a valid component type. Additionally,
   CAST, BOUNDEDCAST, and DATASCALEPRESERVE are not valid conversion types.