Is the project active?

Jul 18, 2012 at 3:26 PM

Hello

Seeing that there haven't been any commits for a while and the project status is "Alpha", I'm wondering whether it's still being actively developed and whether it's considered safe to use in production.

Thank you

Developer
Jul 18, 2012 at 6:39 PM

I'm not sure if I would use it in production, but I'm a bit paranoid mind you. The Lucene.NET portion is very sturdy and comes straight out of the oven from the Lucene/Port at the Apache foundation - so no issues there. Any bugs would be located in the LinqToLucene library itself.

The Linq implementation however is far from complete, make sure that if you do use it in production all the functions that you need are already available. I hit the ceiling of my Linq knowledge at the moment and need either a bit of help or to complete some research in order to take it to the next step.

Jul 19, 2012 at 9:49 AM

Are there any particular weakness points that you'd be worried about ? You seem to make good use of unit tests so how bad can it be? :)

We'll have some unit tests also - I mean, they won't cover such an extensive range of scenarios but it's better than nothing.

It's a great project so I would love to help you with LINQ but I very much doubt my knowledge of LINQ is better than yours!
Have you considered getting some help from other projects that heavily use LINQ also?

Jul 19, 2012 at 10:40 AM
Edited Jul 19, 2012 at 10:42 AM

And a quick question - when using the EFDatabaseIndexSet you've made, is it possible to store the file on an Azure blob, like in Demo3?

Jul 19, 2012 at 11:04 AM

OK I found it - please add this ctor to EfDatabaseIndexSet:

public AcceleratorIndexContext(Directory directory, AcceleratorContext dataContext)
            : base(directory, dataContext)
        {
        }

 

This way, I can pass an AzureDirectory in.

Jul 19, 2012 at 12:03 PM

Hm... next question: It seems that EfDatabaseIndexSet.Write<T>() generates the index fine, I can query and everything.

However, if I reload the application, the index gets deleted and I have to recreate it.

Is there any way to override this behavior? I want to generate the index on demand and add objects as I add them to the DB.

Jul 20, 2012 at 7:52 AM

Does anybody interested to contribute to this project ?

We at Agrobytes Planning to create REST APIs over this using WCF and make it similar to Solr implementation.

Please let me know  - we all can work together and create good search product over MS technologies. If somebody needs exact road map of our development - I would be happy to share it. 

Basically we are planning to use this in one of our social network.

 

Thanks,

Developer
Jul 20, 2012 at 2:30 PM

Georgiosd,

I've put in the new constructor.

Developer
Jul 20, 2012 at 2:32 PM

udkadam,

What are you proposing?

I'm a developer for LinqToLucene

Developer
Jul 20, 2012 at 2:43 PM

Georgiosd,

I created item 15433 for the bug you mentioned, I'm trying to find it

Developer
Jul 20, 2012 at 2:48 PM

Georgiosd,

If possible, can you give me the simplest repro-steps so that I can reproduce with minimal effort and just use my time to fix it?

Would much appreciate it - thanks,

Jul 20, 2012 at 3:33 PM

I'm guessing it's probably a design issue and maybe not an implementation issue.

Here's what I do:

A1) Define a EfDatabaseIndexSet with a DbContext

A2) Load some objects in the DbContext

A3) Do index.Write<T>()

A4) Run queries on index

At this point, you can read objects fine.

However, if you let the application finish, and run again with:

B1) Create EfDatabaseIndexSet

B2) Try to query => no results because index was deleted on initialization

So what I figured out I had to do is:

C1) Do A1-A3 to create the index from the existing database objects

C2) Call index.Get<T>().Dispose

C3) Create a new Index<T> with the same AzureDirectory object

C4) Run queries successfully

So, it's possible - just a bit counter-intuitive... I would have expected that the Db Helper can:

a) Index the existing DB objects (does this) the first it is run

b) Read/Update and existing index with manual steps

c) If possible, update the index automatically - probably too chaotic to implement

 

Georgios

Jul 4, 2013 at 9:28 PM
*bump

This project still active (vision / roadmap) ?