Counts v2 format

[rakow]

This PR updates the counts format to v2 introducing more flexibility in the type of data that can be stored. In summary, the counts format can now store:

• Any observations additional to counts (volumes)
• Observations differentiated by mode
• Data at any desired (but fixed) time interval

Example file:

<?xml version="1.0" encoding="UTF-8"?>
<counts xmlns="http://www.matsim.org/files/dtd" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.matsim.org/files/dtd http://www.matsim.org/files/dtd/counts_v2.xsd" source="Bundesanstalt für Straßenwesen" name="BASt Counts" description="Aggregated daily traffic volumes for car and freight traffic." year="2021">

	<location refType="link" refId="2423" id="2423" name="Marienfelde_S">

		<measurements type="volumes" networkMode="car" interval="3600">
				<value t="0" val="56.28772173437474" />
				<value t="3600" val="34.67126398386528" />
				<value t="7200" val="37.18496486930159" />
<!-- Other values are left out... -->
		</measurements>
		<measurements type="volumes" networkMode="truck" interval="3600">
				<value t="0" val="11.100505172776847" />
				<value t="3600" val="9.51377028264906" />
				<value t="7200" val="12.541180784315472" />
<!-- Other values are left out... -->
				<value t="82800" val="14.506341987151561" />
		</measurements>
	</location>

The new API involves using createAndAddMeasureLocation, MeasurementLocation and Measurable (see javadoc on usage) instead of existing Count and Volume classes.

All changes are supposed to be backwards compatible, since the old Count and Volume classes are still present and functional.
They will internally use volume measurements of a one-hour interval for the mode car.

Closes #2780

[kainagel]

Question: If it says "interval = 3600" and "t=3600", do you mean the time period from 0 to 3599, or from 3600 to 7199? I think that it should be the latter. (Otherwise, the second that is mentioned in the "t" entry would not even be part of the interval.) If this is already the case, it would help to have the examples start at "t=0". If this is not the case, then I really really think we should change it.

We had lots of problems with the fact that the old format started at 1 and not at 0. @mrieser , I think that you said the same two days ago.