[Building Sakai] Search weirdness

[Adrian Fish]

Excellent. Thanks. I'll check out the JIRA.

Cheers,

Adrian.

Stephen Marquard wrote:
> The JIRA in which this was added is
>
> http://jira.sakaiproject.org/browse/SAK-16034
>
> To support storing digested content, it looks like your ECP should also implement StoredDigestContentProducer. Our logic was that it wasn't always necessary to store the digested versions, depending on how expensive it is to digest content. Some files in CHS are known to be very resource-intensive to digest (PDFs, OOXML docs), so that was the primary target for performance gains.
>
> Regards
> Stephen
>  
>   
>>>> Adrian Fish <a.fish at lancaster.ac.uk> 6/16/2010 6:59 AM >>> 
>>>>         
> Hi Stephen,
>
> It's on the trunk that I'm seeing it. I have actually seen the new 
> directory I think (indexwork?). My ECP getContent is definitely getting 
> called on a match though.
>
> Cheers,
>
> Adrian.
>
> Stephen Marquard wrote:
>   
>> Hi Adrian,
>>
>> Search does not store the digested content in the index itself (it just stores the search terms), so when the search tool displays a list of matches, it calls the ECP for each item to get back a digested version show it can show the few lines with matching text.
>>
>> One of the changes between 2.6 and 2.7 is that the digested content now gets stored in a separate filesystem tree on first digesting (at index time), removing the need to re-digest it every time there's a match on the content in the index. Are you seeing behaviour below in 2.6.x or 2.7/trunk?
>>
>> Regards
>> Stephen
>>  
>>   
>>     
>>>>> Adrian Fish <a.fish at lancaster.ac.uk> 6/16/2010 4:45 AM >>> 
>>>>>         
>>>>>           
>> I've implemented a EntityContentProducer in YAFT which works as 
>> expected. The search index builder calls into it when a relevant YAFT 
>> event occurs or when the index is refreshed and all seems okay. 
>> However,when I search for something I know is there, via the search tool 
>> interface, the matches and getContent methods on my ECP get called 
>> again, even though the content should theoretically be in the index. I'm 
>> wondering whether this is the case for all the search enabled tools.
>>
>> Has anybody any idea what is going on?
>>
>> Cheers,
>>
>> Adrian.
>>
>>   
>>     
>
>   

-- 
==================================
Adrian Fish
Software Engineer
Centre for e-Science
Bowland Tower South C Floor
Lancaster University
Lancaster
LA1 4YW
email: a.fish at lancaster.ac.uk

http://confluence.sakaiproject.org/display/YAFT/Yaft
http://confluence.sakaiproject.org/display/BLOG/Home
http://confluence.sakaiproject.org/display/AGORA/Home

-------------- next part --------------
A non-text attachment was scrubbed...
Name: a_fish.vcf
Type: text/x-vcard
Size: 289 bytes
Desc: not available
Url : http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20100616/5f15225b/attachment.vcf