SOLR-18037 Remove "local" tika extraction backend from branch_9x

[janhoy]

This is a backport of #3784 to branch_9x, targeting Solr 9.11. We normally don't remove features in a minor version, but it was decided to do an exception due to Tika 1.x being EOL, see https://issues.apache.org/jira/browse/SOLR-18037 for details.

Since we still have XLSXResponseWriter in 9.x, using Apache POI, we cannot get rid of as many dependencies as in 10.0.

[Copilot]

Pull request overview

This PR backports the removal of the deprecated "local" Tika extraction backend from Solr 10.0 to Solr 9.11 (branch_9x). The change addresses security concerns by eliminating the vulnerable Tika 1.x library, requiring users to use an external Tika Server instead.

Key changes include:

• Upgrading Tika from 1.28.5 to 3.2.3 and Apache POI from 5.2.2 to 5.5.1
• Removing the LocalTikaExtractionBackend and related code
• Removing numerous Tika 1.x parser dependencies (PDFBox, NetCDF, SIS, etc.)
• Making tikaserver.url a required configuration parameter
• Adding clear error messages for removed configuration options

Reviewed changes

Copilot reviewed 198 out of 206 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
versions.props, versions.lock	Updated Tika to 3.2.3, POI to 5.5.1, removed ~60 Tika 1.x parser dependencies
ExtractingRequestHandler.java	Removed local backend support, added validation requiring tikaserver.url, provides clear error messages for deprecated configs
TikaServerExtractionBackend.java	Minor error message improvement, TODO comment about removing Tika dependency
ParseContextConfig.java, LocalTikaExtractionBackend.java	Completely removed (local backend implementation files)
indexing-with-tika.adoc	Updated documentation to reflect tikaserver-only approach, Docker setup instructions
test files	Removed local backend tests, updated integration tests to use Docker-based Tika Server
sample configs	Updated techproducts config to point to localhost:9998 Tika Server
license files	Removed ~100 license files for removed Tika 1.x parser dependencies

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The documentation incorrectly states "As of Solr 10" when this change is being backported to Solr 9.11 (branch_9x). This should be updated to "As of Solr 9.11" to accurately reflect when this feature was removed.

[epugh]

LGTM.. Thank you for taking on what feels like an important yet somewhat thankless task @janhoy.

I appreciate the documentation edits you made.

[epugh]

@janhoy I approved it, but obviosly there are test fixes needed... The thrust of the PR looks great. Let me know if you need additional manual testing.

[janhoy]

@janhoy I approved it, but obviosly there are test fixes needed... The thrust of the PR looks great. Let me know if you need additional manual testing.

I got XLSXResponsewriter working. But of course there is also langid module that uses Tika1.x LanguageIdentifier. Tika3 has a completely new LanguageDetector API with an OptimizeLangDetector implementation quite different from the old Tika detector. I'm doing a shot at switching our TikaLanguageIdentifierUpdateProcessor to the new one, and if most existing tests pass, call it a good temporary solution for 9.x.

EDIT: Turned out that the old LanguageIdentifier class is still present in Tika3, so it ended up as a minor change and no user facing changes.

[janhoy]

Appreciate any review, but especially ref-guide changes (so it is clear to users this is a breaking change).