As Dave suggested, I tried to build the index with fuseki binaries, from
same assembler file :

java -cp fuseki-server.jar jena.textindexer
--desc="/home/nps/text-config.ttl"
INFO  2 (2 per second) properties indexed

But again, jena complains he cannot find the text index. Config are same as
previous email & directly come from documentation.

Apache jena is 2.13
jena-text is 1.1.2



2016-02-21 15:57 GMT+01:00 Nicolas Paris <[email protected]>:

> Well,
>
> Here are the actualized gist (and simplified example):
> my code :
> https://gist.github.com/parisni/53cab71322fadf9ee94b
> my Assemblar:
> https://gist.github.com/parisni/24a8fbeecbfc1317312c
> my log:
> https://gist.github.com/parisni/9db5dd9f3881ba61b8b8
>
> When I add new statements to my TDB, I get logs :
> 2016-02-21 15:49:04 DEBUG TextIndexLucene:195 - Add entity: 1 : {text=A}
> 2016-02-21 15:49:04 DEBUG TextIndexLucene:195 - Add entity: 2 : {text=B}
>
> Moreover, I get new files in the lucene repository:
> -rw-rw-r-- 1 nps nps   20 févr. 21 15:49 segments.gen
> -rw-rw-r-- 1 nps nps   81 févr. 21 15:49 segments_2
> -rw-rw-r-- 1 nps nps  268 févr. 21 15:49 _0.cfe
> -rw-rw-r-- 1 nps nps  736 févr. 21 15:49 _0.cfs
> -rw-rw-r-- 1 nps nps  265 févr. 21 15:49 _0.si
> -rw-rw-r-- 1 nps nps    0 févr. 21 15:49 write.lock
>
>
> Sadly I still got the same warning:
> 2016-02-21 15:49:04 WARN  TextQueryPF:78 - Failed to find the text index :
> tried context and as a text-enabled dataset
> 2016-02-21 15:49:04 WARN  TextQueryPF:78 - No text index - no text search
> performed
>
> Moreover, the results from lucene contains error (I would expect "B"
> before "A")
> ---------------
> | s   | label |
> ===============
> | <1> | "A"   |
> | <2> | "B"   |
> ---------------
>
> So this looks like my config is not completely working, but I cannot
> figure where is the problem.
>
> 2016-02-21 13:00 GMT+01:00 Dave Reynolds <[email protected]>:
>
>> On 21/02/16 11:20, Nicolas Paris wrote:
>>
>>> @Dave:
>>> I though this section was fuseki specific, right ?
>>>
>>
>> No. It just uses the Fuseki jar as a convenient way to get access to the
>> text indexer.
>>
>> You either have to load your data through your TDB + text configuration
>> (as others have said) or, if you already have the TDB, then run the text
>> indexer on it.
>>
>> Dave
>>
>>
>

Reply via email to