Adding formal processor semver-min=2.3.0 to expath-pkg.xml

[ljo]

No description provided.

[ljo]

Adding nio.2 to cover API change for eXist-db 3.0b2

[joewiz]

Would this PR make the package compatible with eXist 3.0 (Feb 2017 release)? If so, would it be possible to accept the PR and post this?

I ask because the current app listed in the public-repo installs without error in eXist 3.0, but the demo doesn't work, and function documentation lookups in eXide fail with this:

<exception>
  <path>/db/apps/eXide/modules/docs.xql</path>
  <message>exerr:ERROR Cannot find module class from EXPath repository: 
    org.exist.xquery.ner.StanfordNERModule [at line 46, column 46, source: /db/apps/eXide/modules/docs.xql] 
    In function:
    local:get-matching-functions(xs:string) [174:13:/db/apps/eXide/modules/docs.xql]</message>
</exception>

[joewiz]

@wolfgangmm I have tested this PR from @ljo in eXist 4.4.0 and confirms that it works. In contrast, installing the current release .xar fails with a "Package requires eXistdb version > 2.2.1 < 3.0.1. Installed version is 4.4.0" error.

If you merge this, I'll be happy to submit a PR to bump the version and publish a new release to the public-repo.

[joewiz]

I also noticed that the eXist package is based on stanford-ner v3.5.2. I replaced the jar and classifiers with v3.9.1, and all my tests passed. (See the improvements here https://nlp.stanford.edu/software/CRF-NER.shtml#History.)

So if you'd like I could include this in my follow-on PRs, so we could benefit from the ~3 years of development that've gone into NER.

[joewiz]

My WIP branch is at https://github.com/joewiz/exist-stanford-ner/tree/prepare-new-release. But we should merge this one in first.

[duncdrum]

@joewiz just out of curiosity isn't all the NER stuff rolled into core-nlp?

[joewiz]

@duncdrum Yup, NER is one component in CoreNLP. From https://stanfordnlp.github.io/CoreNLP/index.html:

Stanford CoreNLP integrates many of Stanford’s NLP tools, including the part-of-speech (POS) tagger, the named entity recognizer (NER), the parser, the coreference resolution system, sentiment analysis, bootstrapped pattern learning, and the open information extraction tools.

[duncdrum]

and its available via https://github.com/ljo/exist-stanford-corenlp