Getting reasoner to execute after base model or schema change

[Howard Burrows <ho...@gmail.com>]

I'm using Fuseki/TDB and use a number of named graphs in a modular way.  Many of these graphs serve as either a base model or schema model for an inference model.  If I make a change directly to a base or schema model I need to restart Fuseki to get those updates reflected in the inference model.  I can update the inference model directly and those changes get written back to the base model but many times I don't want those updates written to the "base" model but rather some other model (i.e. I partitioned my triples into separate graphs to keep things modular and maintainable).

Using Fuseki, is there a way I can modify an inference model's base or schema graphs and then tell (or force) Fuseki to re-run the reasoner for a single (or all) inference model(s)?  

I understand that I can built a custom solution using Jena but I'm really hoping to avoid that and continue to use Fuseki directly which except for this issue is great for my use cases.  I'm also not opposed to modifying the source for Fuseki to make such an enhancement if that's the best way forward (tips on what to do would be very welcome).

Any advice and guidance would be very welcome.

Thanks,

Howard Burrows

Re: Getting reasoner to execute after base model or schema change

[Andy Seaborne <an...@apache.org>]

On 07/10/13 15:35, Howard Burrows wrote:
> Andy,
>
> Thanks for the response.  Do you have any suggestions on the best approach to enhance Fuseki to tell it to re-run the reasoner on an inference model?

If you're not going via the inference model to do the updates, I think 
the only thing that can be done is to restart the server.

	Andy


>
> Howard
>
> On Oct 5, 2013, at 3:17 AM, Andy Seaborne <an...@gmail.com> wrote:
>
>> On 03/10/13 17:01, Howard Burrows wrote:
>>> I'm using Fuseki/TDB and use a number of named graphs in a modular
>>> way.  Many of these graphs serve as either a base model or schema
>>> model for an inference model.  If I make a change directly to a base
>>> or schema model I need to restart Fuseki to get those updates
>>> reflected in the inference model.  I can update the inference model
>>> directly and those changes get written back to the base model but
>>> many times I don't want those updates written to the "base" model but
>>> rather some other model (i.e. I partitioned my triples into separate
>>> graphs to keep things modular and maintainable).
>>>
>>> Using Fuseki, is there a way I can modify an inference model's base
>>> or schema graphs and then tell (or force) Fuseki to re-run the
>>> reasoner for a single (or all) inference model(s)?
>>>
>>> I understand that I can built a custom solution using Jena but I'm
>>> really hoping to avoid that and continue to use Fuseki directly which
>>> except for this issue is great for my use cases.  I'm also not
>>> opposed to modifying the source for Fuseki to make such an
>>> enhancement if that's the best way forward (tips on what to do would
>>> be very welcome).
>>>
>>> Any advice and guidance would be very welcome.
>>>
>>> Thanks,
>>>
>>> Howard Burrows
>>>
>>
>> Hi Howard,
>>
>> I don't think there is a way - updates to an infernece model do need to go in through the inference model.  Even if there were, it would likely need to e a complete reinference because the details of what the changes were would be lost.
>>
>> It might be good to have a Fuseki mechanism to reapply the configuration for a dataset without needing a restart as long as it does not open a DOS attack path.
>>
>> 	Andy
>

Re: Getting reasoner to execute after base model or schema change

[Howard Burrows <ho...@howardburrows.me>]

Andy,

Thanks for the response.  Do you have any suggestions on the best approach to enhance Fuseki to tell it to re-run the reasoner on an inference model?

Howard

On Oct 5, 2013, at 3:17 AM, Andy Seaborne <an...@gmail.com> wrote:

> On 03/10/13 17:01, Howard Burrows wrote:
>> I'm using Fuseki/TDB and use a number of named graphs in a modular
>> way.  Many of these graphs serve as either a base model or schema
>> model for an inference model.  If I make a change directly to a base
>> or schema model I need to restart Fuseki to get those updates
>> reflected in the inference model.  I can update the inference model
>> directly and those changes get written back to the base model but
>> many times I don't want those updates written to the "base" model but
>> rather some other model (i.e. I partitioned my triples into separate
>> graphs to keep things modular and maintainable).
>> 
>> Using Fuseki, is there a way I can modify an inference model's base
>> or schema graphs and then tell (or force) Fuseki to re-run the
>> reasoner for a single (or all) inference model(s)?
>> 
>> I understand that I can built a custom solution using Jena but I'm
>> really hoping to avoid that and continue to use Fuseki directly which
>> except for this issue is great for my use cases.  I'm also not
>> opposed to modifying the source for Fuseki to make such an
>> enhancement if that's the best way forward (tips on what to do would
>> be very welcome).
>> 
>> Any advice and guidance would be very welcome.
>> 
>> Thanks,
>> 
>> Howard Burrows
>> 
> 
> Hi Howard,
> 
> I don't think there is a way - updates to an infernece model do need to go in through the inference model.  Even if there were, it would likely need to e a complete reinference because the details of what the changes were would be lost.
> 
> It might be good to have a Fuseki mechanism to reapply the configuration for a dataset without needing a restart as long as it does not open a DOS attack path.
> 
> 	Andy

Re: Getting reasoner to execute after base model or schema change

[Andy Seaborne <an...@gmail.com>]

On 03/10/13 17:01, Howard Burrows wrote:
> I'm using Fuseki/TDB and use a number of named graphs in a modular
> way.  Many of these graphs serve as either a base model or schema
> model for an inference model.  If I make a change directly to a base
> or schema model I need to restart Fuseki to get those updates
> reflected in the inference model.  I can update the inference model
> directly and those changes get written back to the base model but
> many times I don't want those updates written to the "base" model but
> rather some other model (i.e. I partitioned my triples into separate
> graphs to keep things modular and maintainable).
>
> Using Fuseki, is there a way I can modify an inference model's base
> or schema graphs and then tell (or force) Fuseki to re-run the
> reasoner for a single (or all) inference model(s)?
>
> I understand that I can built a custom solution using Jena but I'm
> really hoping to avoid that and continue to use Fuseki directly which
> except for this issue is great for my use cases.  I'm also not
> opposed to modifying the source for Fuseki to make such an
> enhancement if that's the best way forward (tips on what to do would
> be very welcome).
>
> Any advice and guidance would be very welcome.
>
> Thanks,
>
> Howard Burrows
>

Hi Howard,

I don't think there is a way - updates to an infernece model do need to 
go in through the inference model.  Even if there were, it would likely 
need to e a complete reinference because the details of what the changes 
were would be lost.

It might be good to have a Fuseki mechanism to reapply the configuration 
for a dataset without needing a restart as long as it does not open a 
DOS attack path.

	Andy