model.xml

    <project name='project_01' pubsub='auto' threads='1' use-tagged-token='true'> 
    <description> 
    	This example is a basic demonstration of using the moving relative range (MRR)
	algorithm. The moving relative range provides a measure of volatility for a 
	nonstationary time series, where the mean and variance of the series changes
	over time.  

	This project contains a single continuous query consisting of the following: 
	1.	A source window that receives the data to analyze
	2.	A calculate window that performs the moving relative range calculation
    </description>
      <contqueries>
        <contquery name='cq_01' trace='w_calculate'>
          <windows>
            <window-source name='w_data' insert-only='true' index='pi_EMPTY'> 
            <description> 
		This source window receives the included example input data file via
		a file socket connector. The input data stream is placed into two fields
		for each observation: an ID that acts as the data stream's key, named id;
		and a variable x1 that represents an element of the time series.
            </description>
              <schema>
                <fields>
                  <field name='id'  type='int64' key='true'/>
                  <field name='x1' type='double'/>
                </fields>
              </schema>
              <connectors>
                <connector class='fs' name='publisher'>
                  <properties>
                    <property name='type'>pub</property>
                    <property name='fstype'>csv</property>
                    <property name='fsname'>input.csv</property>
                    <property name='transactional'>true</property>
                    <property name='blocksize'>1</property>
                  </properties>
                </connector>
              </connectors>
            </window-source>

            <window-calculate name='w_calculate' algorithm='MRR' index='pi_EMPTY'>
            <description> 
		This calculate window receives data events from the source window. It 
		publishes the calculated moving relating range. In this example, the
		following algorithm parameters govern the MRR algorithm: 

		rangeWindowLength:	Specifies the window length to calculate the 
					range (M). For a time series whose mean is 
					changing quickly, a lower value should be 
					specified. 
		expRangeWindowLength:	Specifies the window length to calculate the
					moving relative range (K). For a time series 
					whose variance is changing quickly, a lower
					value should be specified. 

		Additionally, the following input-map and output-map properties define
		this calculate window: 

		input:			Specifies the analysis variable name in the
					input stream. 
		timeId:			Specifies the time ID variable name in the
					input stream. 

		timeIdOut:		Specifies the time ID (key) variable name in the
					output stream. 
		erangeOut:		Specifies the expected range variable name in the
					output stream. 
		rangeOut:		Specifies the range variable name in the output
					stream. 

		The resulting output is written to the file result.out via a file socket
		connector. 
	    </description>
              <schema>
                <fields>
                  <field name='id' type='int64' key='true'/>
		  <field name='rangeOut' type='double'/>
                  <field name='erangeOut' type='double'/>
                </fields>
              </schema>
              <parameters>
                <properties>
                  <property name='rangeWindowLength'>3</property>
                  <property name='expRangeWindowLength'>3</property>
                </properties>
              </parameters>
                <input-map>
                <properties>
                  <property name="input">x1</property>
                  <property name="timeId">id</property>
                </properties>
              </input-map>
              <output-map>
                 <properties>
                  <property name="timeIdOut">id</property>
                  <property name="erangeOut">erangeOut</property>
                  <property name="rangeOut">rangeOut</property>
                </properties>
              </output-map>
              <connectors>
                <connector class='fs' name='sub'>
                  <properties>
                    <property name='type'>sub</property>
                    <property name='fstype'>csv</property>
                    <property name='header'>true</property>
                    <property name='fsname'>result.out</property>
                    <property name='snapshot'>true</property> 
                    <property name='header'>true</property>
                  </properties>
                </connector>
              </connectors>
            </window-calculate>
          </windows>
          <edges>
            <edge source='w_data'    target='w_calculate' role='data'/>
          </edges>
        </contquery>
      </contqueries>
    </project>