TIKA-4831: Add content-based detection and a parser for GeoGebra files (ggb, ggs, ggt)

[dschmidt]

Adds detection and parsing support for the GeoGebra file formats.

Issue: https://issues.apache.org/jira/browse/TIKA-4831

Detection

• application/vnd.geogebra.file (*.ggb) and application/vnd.geogebra.tool (*.ggt) are now sub-class-of application/zip, so the filename hint survives magic detection instead of being discarded in favor of plain application/zip
• new mime types: application/vnd.geogebra.slides (*.ggs, zip-based) and application/vnd.geogebra.pinboard (*.ggp, JSON-based)
• a new GeoGebraDetector (zip container detector, ZipFile and streaming mode) recognizes the formats without a filename by their well-known entries: geogebra.xml (worksheet), structure.json + _slideN/geogebra.xml (Notes/Slides), geogebra_macro.xml (tool). Since a worksheet with macros contains both geogebra.xml and geogebra_macro.xml, the decision is made after all entry names have been seen
• ZipParser.ZIP_SPECIALIZATIONS is kept in sync with the new registry entries

Parser

A new GeoGebraParser (miscoffice module) for ggb/ggs/ggt:

• metadata: construction title/author/date become dc:title/dc:creator/geogebra:date, plus geogebra:appName, geogebra:appVersion, geogebra:formatVersion, geogebra:id; Notes/Slides set xmpTPg:NPages
• content: user-visible text as XHTML paragraphs - string-literal expression elements of text objects, rich-text content runs (JSON), element captions, macro names and help texts; Notes/Slides emit one div per slide in structure.json order
• the representative rendering (geogebra_thumbnail.png at the root, or the first slide's thumbnail) is emitted as an embedded document marked embeddedResourceType=THUMBNAIL, following the existing convention in the OOXML, ODF and iWork parsers, so /unpack/all sidecars identify the preview image; other slides' thumbnails are redundant renderings and are skipped
• any other embedded file (e.g. inserted pictures) is emitted as an embedded document

XML is parsed through XMLReaderUtils.parseSAX like the other parsers in the module; structure.json and rich-text runs use Jackson (new jackson-databind dependency in the miscoffice module, version managed by the existing BOM import).

Testing

• new unit tests: GeoGebraDetectionTest (zip-commons) and GeoGebraParserTest (miscoffice) with crafted ggb/ggs/ggt fixtures; the ggs fixture deliberately orders _slide1 before _slide0 in structure.json to pin the ordering behavior
• full test suites of tika-core, tika-parser-zip-commons, tika-parser-miscoffice-module and tika-parser-pkg-module pass, plus CompositeZipContainerDetectorTest in the integration tests
• verified end-to-end against a real-world GeoGebra Notes file: detected as vnd.geogebra.slides, text extracted, thumbnail emitted with the THUMBNAIL marker