improve pom.xml parsing; use XML parser instead of regexp

[mbohun]

At the moment the script generate-github-travis-build-summary-with-grails.sh used to generate the "build status" table (github gist) embedded inside https://atlasoflivingaustralia.github.io/ is using clumsy regexp/grep/sed/etc to extract:

   <groupId />
   <artifactId />
   <version />

from ALA's maven pom.xml files.
This had it's obvious limitations:

• <parent /> (if present) has to be the very first sub-element of <project />
• <version /> and <groupId /> of the project had to be the first "version" and "groupId" element-s after the closing </parent>
• etc. recently <parent> pom was added, some "common", "shared" settings/properties like the JDK version, groupId moved to the parent pom - ALL changes that are next to impossible to process/parse with regexp.

The solution is to EITHER:

• parse pom.xml with a XML parser, examples:

  xmllint --xpath '/*[local-name()="project"]/*[local-name()="parent"]/*[local-name()="groupId"]/text()' ala-cas-2.0_pom.xml; echo;
  au.org.ala
  
  xmllint --xpath '/*[local-name()="project"]/*[local-name()="parent"]/*[local-name()="groupId"]/text()' ala-logger_pom.xml; echo;
  au.org.ala
  
  xmllint --xpath '/*[local-name()="project"]/*[local-name()="version"]/text()' ala-cas-2.0_pom.xml; echo;
  2.0.11-SNAPSHOT
  
  xmllint --xpath '/*[local-name()="project"]/*[local-name()="version"]/text()' ala-logger_pom.xml; echo;
  1.4.1-SNAPSHOT
  
  xmllint --xpath '/*[local-name()="project"]/*[local-name()="artifactId"]/text()' ala-cas-2.0_pom.xml; echo;
  ala-cas
  
  xmllint --xpath '/*[local-name()="project"]/*[local-name()="artifactId"]/text()' ala-logger_pom.xml; echo;
  ala-logger

  OR

• do full git clone of each processed project and then use mvn on the clone to extract desired properties