Design issues

[jai.thomas]

Hi,
We are getting ready to integrate Texis on our system and are having a couple of design issues.

1. All page-related information is stored in Oracle (8.1.6) tables and I assume we need to copy (part of) this data to a Texis database to enable searching. Now, we let privileged users to change page information using our content management tool and changes (add, change, delete) happen on an on-going basis. My question is how do we keep the Texis database in sync. with the Oracle database. All content management actions are performed by EJB objects and I am looking for a solution which has least impact on the system.
What is the easiest way to do this?

2. We also maintain a list of external links (ie. links to other sites with business relevance). How can we enable users to search on these external links?

Thanks
Jai Thomas

[John]

1. Probably the simplest approach to synchronization would be to write Vortex functions corresponding to the add, change and delete you might perform, and have the EJB invoke those with an HTTP request.

2. You would probably want to use code based on Webinator. If you are searching those individually then a simple Webinator crawl would suffice, otherwise you may want to modify the crawler to store the contents of those links in the table you are populating with your data.

[jai.thomas]

Regarding issue 2, we won't be searching them individually, but as one search topic. How do I copy the contents of a link to a table. fromfile/frommfiletext doesn't seem to work on links.

Thanks
Jai

[John]

The <fetch> function in Vortex will fetch a link. $ret will contain the raw HTML, and you can use <urltext> function to extract the text from the HTML, which you would then insert into the table, e.g.

<fetch $link>
<urltext>
<SQL "insert into table values($link, $ret)"></SQL>