search

BIDS-collaborative / destress

Helping @peparedes with text analysis of livejournal data
ISC License
7 stars 2 forks source link

Give access to Lucene implementation on stout #12

Closed peparedes closed 9 years ago

peparedes commented 9 years ago

John - how could we provide access to at least some folks working on queries to the lucene search implementation on the livejournal data on stout?

davclark commented 9 years ago

Hey @peparedes, if you "@-mention" people, it will automatically show up in their git inbox (and maybe trigger an email, depending on settings). John hasn't accepted his invite to this repo, though - so maybe we should ask / remind him to.

But more importantly, is this still necessary - or are we able to simply use Lucene with the index files I copied to mercury?

peparedes commented 9 years ago

Seems we could use the I deres you copied... Gene is looking into it.

John offered to make it work as well... Buy yes, he needs to accept first.

P On Mar 3, 2015 10:38 AM, "Dav Clark" notifications@github.com wrote:

Hey @peparedes https://github.com/peparedes, if you "@-mention" people, it will automatically show up in their git inbox (and maybe trigger an email, depending on settings). John hasn't accepted his invite to this repo, though - so maybe we should ask / remind him to.

But more importantly, is this still necessary - or are we able to simply use Lucene with the index files I copied to mercury?

— Reply to this email directly or view it on GitHub https://github.com/berkeley-dsc/destress/issues/12#issuecomment-77006233 .

geneyoo commented 9 years ago

I'm looking into it right now -- I'll let you know if I find success into using the indices (it looks likely, at least for the original format). But, another thing I'm worried about is the new concatenated files that we would be using? Since pierre now combined the files... the indices for the new searches will not be valid.

davclark commented 9 years ago

It's not clear that we need to be looking at the same files for all cases... I think it's fine for now if the lucene index returns the smaller / older files.

geneyoo commented 9 years ago

Just tested a few searches using some queries with the copied /index/ directory, and it seems valid/functional. It looks like we won't need to re-create indices.

peparedes commented 9 years ago

Excellent!

Good job!

P On Mar 4, 2015 3:27 AM, "geneyoo" notifications@github.com wrote:

Just tested a few searches using some queries with the copied /index/ directory, and it seems valid/functional. It looks like we won't need to re-create indices.

— Reply to this email directly or view it on GitHub https://github.com/berkeley-dsc/destress/issues/12#issuecomment-77142074 .

coryschillaci commented 9 years ago

Can we close this issue? Seems like the remaining work is in #11 .