You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jena.apache.org by Andy Seaborne <an...@apache.org> on 2021/02/10 15:13:09 UTC
Next release: Jena4
The next release of Jena will be version 4.0.0.
The development builds are already 4.0.0-SNAPSHOT.
* Java11 runtime is required
* Removed
jena-sdb, jena-text-es, jena-elephas
* Upgrade jena-text to Lucene 8.8.x
* IRIs (see below)
* The default git branches are 'main'
* Remove many deprecated methods
* Rename core.RDFReader -> RDFReaderI and core.RDFWriter -> RDFWriterI
See for JIRA tickets so far:
https://s.apache.org/jena-4.0.0-jira
----
Lucene:
This is a major version number upgrade. Apache Lucene policy is
generally "two major versions" so if your text index was created awhile
ago, you may need to reload even if you have been upgrading Jena itself.
https://lucene.apache.org/core/8_8_0/MIGRATE.html
IRIs:
There is a new internal API for working with different IRI
implementations. The one in Jena 4.0.0 will be the current Jena3 jena-iri.
There a some areas to be aware of:
+ %ZZ (%-and non-hex digits) are illegal syntax in IRIs.
It has never been legal in the RFCs but could be read in.
+ urn: The correct form of a URN, by the RFC definition, is
urn:NID:NSS where NID is at least two characters and NSS is at least
one character.
<urn:x:abcd> is illegal as is <urn:abc:>
These are not rejected by Jena4 but be warned!
If you have custom URIs schemes, please ensure they match the syntax of
RFC 3986/3987.
It would be good to know about such schemes so representative examples
can be added to the Jena test suite.
jena-iri will remain in Jena. It is useful for more detailed checking
and analysis of IRIs.
Please do try development builds with your system - snapshot builds for
Jena 4.0.0 are now available from
https://repository.apache.org/content/groups/snapshots/ .
The hoped for release date is mid/late March.
This is depend on feedback and contributions.
Andy
Re: Next release: Jena4
Posted by Andy Seaborne <an...@apache.org>.
Hi Sorin - thanks for getting in-touch.
Like any open source project, there is a limited amount of engineering
resource for Jena. Having extra modules has a certain cost, including
dealing with security CVE's, and when evolving the system as a whole
The question is always "what is Jena?". One answer is that it is a core
RDF system. With Jena4, there is a natural point to clear up. It seems
better to make changes in a considered way before they become a burden
and a hindrance.
The original contributor of jena-text-es has previously said they aren't
able to continue supporting it.
We haven't heard much from jena-text-es users except for one recently
that seems to have been about configuring ElasticSearch and not related
to the Jena module. While responding to questions is not guaranteed, the
community has typically done so.
jena-text-es would benefit from some attention:
* The build logging swamps the build output. That makes it hard to see
if other modules have unexpected output logged.
* It depends on Lucene version so we have a linkage that affects
jena-text especially difficult because old indexes may become invalid
due to the changes in supported file format versions.
The codebase for jena-text-es was brought up-to-date with the changes
for Jena 4.0.0 at the time it was retired.
What use to do you make if it? It can fulfil certain roles:
1/ It can be updated directly
2/ Scale of index
3/ Sharing operations with an existing deployment.
Are you, or others, in a position to help with it?
Andy
FWIW:
SDB was retired and then came back. I've now worked with the downstream
open source project team and they have migrated to TDB now and we've
agreed that SDB can be retired again.
Andy
On 10/02/2021 16:09, Sorin Gheorghiu wrote:
> Hi Andy,
>
> would v.4.0.0 drop the support for full text search via ElasticSearch?
>
> Regards,
> Sorin
>
> Am 10.02.2021 um 16:13 schrieb Andy Seaborne:
>> The next release of Jena will be version 4.0.0.
>> The development builds are already 4.0.0-SNAPSHOT.
>>
>> * Java11 runtime is required
>> * Removed
>> jena-sdb, jena-text-es, jena-elephas
>> * Upgrade jena-text to Lucene 8.8.x
>> * IRIs (see below)
>> * The default git branches are 'main'
>> * Remove many deprecated methods
>> * Rename core.RDFReader -> RDFReaderI and core.RDFWriter -> RDFWriterI
>>
>> See for JIRA tickets so far:
>> https://s.apache.org/jena-4.0.0-jira
>>
>> ----
>>
>> Lucene:
>>
>> This is a major version number upgrade. Apache Lucene policy is
>> generally "two major versions" so if your text index was created
>> awhile ago, you may need to reload even if you have been upgrading
>> Jena itself.
>>
>> https://lucene.apache.org/core/8_8_0/MIGRATE.html
>>
>> IRIs:
>>
>> There is a new internal API for working with different IRI
>> implementations. The one in Jena 4.0.0 will be the current Jena3
>> jena-iri.
>>
>> There a some areas to be aware of:
>>
>> + %ZZ (%-and non-hex digits) are illegal syntax in IRIs.
>> It has never been legal in the RFCs but could be read in.
>> + urn: The correct form of a URN, by the RFC definition, is
>> urn:NID:NSS where NID is at least two characters and NSS is at least
>> one character.
>>
>> <urn:x:abcd> is illegal as is <urn:abc:>
>> These are not rejected by Jena4 but be warned!
>>
>> If you have custom URIs schemes, please ensure they match the syntax
>> of RFC 3986/3987.
>>
>> It would be good to know about such schemes so representative examples
>> can be added to the Jena test suite.
>>
>> jena-iri will remain in Jena. It is useful for more detailed checking
>> and analysis of IRIs.
>>
>>
>> Please do try development builds with your system - snapshot builds
>> for Jena 4.0.0 are now available from
>> https://repository.apache.org/content/groups/snapshots/ .
>>
>> The hoped for release date is mid/late March.
>> This is depend on feedback and contributions.
>>
>> Andy
>
Re: Next release: Jena4
Posted by Sorin Gheorghiu <so...@uni-konstanz.de>.
Hi Andy,
would v.4.0.0 drop the support for full text search via ElasticSearch?
Regards,
Sorin
Am 10.02.2021 um 16:13 schrieb Andy Seaborne:
> The next release of Jena will be version 4.0.0.
> The development builds are already 4.0.0-SNAPSHOT.
>
> * Java11 runtime is required
> * Removed
> jena-sdb, jena-text-es, jena-elephas
> * Upgrade jena-text to Lucene 8.8.x
> * IRIs (see below)
> * The default git branches are 'main'
> * Remove many deprecated methods
> * Rename core.RDFReader -> RDFReaderI and core.RDFWriter -> RDFWriterI
>
> See for JIRA tickets so far:
> https://s.apache.org/jena-4.0.0-jira
>
> ----
>
> Lucene:
>
> This is a major version number upgrade. Apache Lucene policy is
> generally "two major versions" so if your text index was created
> awhile ago, you may need to reload even if you have been upgrading
> Jena itself.
>
> https://lucene.apache.org/core/8_8_0/MIGRATE.html
>
> IRIs:
>
> There is a new internal API for working with different IRI
> implementations. The one in Jena 4.0.0 will be the current Jena3
> jena-iri.
>
> There a some areas to be aware of:
>
> + %ZZ (%-and non-hex digits) are illegal syntax in IRIs.
> It has never been legal in the RFCs but could be read in.
> + urn: The correct form of a URN, by the RFC definition, is
> urn:NID:NSS where NID is at least two characters and NSS is at least
> one character.
>
> <urn:x:abcd> is illegal as is <urn:abc:>
> These are not rejected by Jena4 but be warned!
>
> If you have custom URIs schemes, please ensure they match the syntax
> of RFC 3986/3987.
>
> It would be good to know about such schemes so representative examples
> can be added to the Jena test suite.
>
> jena-iri will remain in Jena. It is useful for more detailed checking
> and analysis of IRIs.
>
>
> Please do try development builds with your system - snapshot builds
> for Jena 4.0.0 are now available from
> https://repository.apache.org/content/groups/snapshots/ .
>
> The hoped for release date is mid/late March.
> This is depend on feedback and contributions.
>
> Andy
--
Sorin Gheorghiu Tel: +49 7531 88-3198
Universität Konstanz Raum: B705
78464 Konstanz sorin.gheorghiu@uni-konstanz.de
- KIM: Abteilung IT-Dienste Forschung und Lehre -