Hi all, I would like to share with you the integration process of this searching code (Whoosh) in my web2py application. For those who don't know it, Whoosh is a search engine fully written in python, so it is easy to use in a framework such as web2py, but with some little issues. Well, after placing the source code of Whoosh in modules section, and writing the proper imports, we are ready to go using it.
The problem comes when, reading whoosh documentation, we find out that when indexing a document, whoosh lock the index (it is just locked to write, you can read in while indexing). So, how to solve this imposibility of concurrency? (Imagine that two users are performing actions which require an indexation at the same time... one of them would return an error). What I did is to create a queue, and a cron script that looks into that queue and process it. The queue is a simple table in the database, with all the fields to be indexed, that stores a record everytime a user performs an action that needs to index something. Then, the cron script reads that table, locks the index, indexes the records in the queue, and frees the index. I hope this help someone... if you have any doubts, please do not hesitate in asking. regards, Bernardo
