its.backbone.domain
Class DataFactory

java.lang.Object
  its.backbone.domain.DataFactory

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

TmsDataFactory

public abstract class DataFactory

extends java.lang.Object

implements java.io.Serializable

Defines the interface to a domain-specific data extractor. Extractors must extend this abstract class and implement its methods.

Part of the process of creating a transmitter for a new SDD data stream is creating the Extractor for that data stream. The Extractor is sent over the network to SDD Receivers, which can use it to interpret the incoming data frames.

To create a new extractor,

• Extend this abstract DataFactory class, implementing the methods described below. For the SddReceiver to identify the new class when loading it at runtime, the name of the subclass must end with DataFactory.
• Place the subclass and any other classes that are part of its implementation in a temporary directory.
• Strip off any package assignments (the package line at the head of the source files).
• Compile all the source files using javac *.java.
• Bind the resulting class files into a jar file using jar cvfM myjarfile.jar *.class.
• Test the new jar file using java its.backbone.domain.ReadJar myjarfile.jar
• In your transmitter, pass the jar file to SddTransmitter.transmitExtractor.

Author:

Noah S Friedland

See Also:

SddTransmitter.transmitExtractor(java.lang.String), Serialized Form

Constructor Summary

DataFactory()

Method Summary

abstract java.lang.String	getDataOffsetTableName() This method returns the name of the data offset table in the datastream schema.
abstract java.util.Hashtable	getExtractionHashtable(Schema schema, ContentsData dataOffsetTable, ContentsData data) To produce the ExtractedData event, the SddReceiver invokes this method on the extractor it has received.
abstract boolean	hasDataOffsetTable() Indicates whether the Extractor needs to be provided with a data offset table from the data dictionary (schema and contents) in order to interpret the data frame.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DataFactory

public DataFactory()

Method Detail

hasDataOffsetTable

public abstract boolean hasDataOffsetTable()

Indicates whether the Extractor needs to be provided with a data offset table from the data dictionary (schema and contents) in order to interpret the data frame. A data offset table is a relation that specifies the sensor type, its name, and the offset into the data frame where the relevant data is located. If the data frame contains several different types of data, this table also needs to contain the data type (and account for its size). For an example:

Consider a system that produces three types of binary sensor data, all of whom are packaged into a single data frame. The first type is of size 3 bytes, the second type is of 6 bytes and the third type is of 18 bytes. These sensors' readings can appear in arbitrary order within the blob. To gain access to any specific sensor reading a data offset table is required in the SDD data stream's schema. This table should contains columns that describe the following:

• The sensor name.
• The sensor's type (first, second or third).
• The sensor's offset within the data frame.

For an operational example, consider the "LOOPS" table in the SDD data flow at host sdd.its.washington.edu, port 9033.

Returns:

true if the schema and contents for the SDD stream define a data offset table

getDataOffsetTableName

public abstract java.lang.String getDataOffsetTableName()

This method returns the name of the data offset table in the datastream schema. Please see comments above.

Returns:

the given datastream's data offset table name

getExtractionHashtable

public abstract java.util.Hashtable getExtractionHashtable(Schema schema,
                                                           ContentsData dataOffsetTable,
                                                           ContentsData data)
                                                    throws BadOffsetException

To produce the ExtractedData event, the SddReceiver invokes this method on the extractor it has received. The extractor's responsibility is to produce a hashtable of ContentsData objects, one for each sensor type found within the data, keyed by their associated table names from the schema.

Parameters:

schema - the schema from the SDD datastream.

dataOffsetTable - the table named by getDataOffsetTableName, taken from the Contents frame in the SDD datastream, or null if hasDataOffsetTable returns false.

data - the data frame from the SDD datastream, packaged inside a ContentsData object. This object contains a single column entry named "data". The incoming data frame can be retrieved as follows:
byte[] databytes = (byte[])data.getItem("data");

Throws:

BadOffsetException - thrown if the data offset table contains an offset that is not within the data frame