You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@couchdb.apache.org by Ronan Jouchet <ro...@cadensimaging.com> on 2017/01/26 17:06:51 UTC
How to prep my CouchDB 2 for *text* indexes/queries?
Hi!
Following mailing list discussion [1], I am looking at the
performance of various Mango ways to fetch a bunch of docs.
Following Garren links, I stumbled upon a blog post called
"Understanding Mango View-Based Indexes vs. Search-Based Indexes" [2].
There, the author mentions that:
> If users don't know in advance what queries will be executed,
> then Mango search-based indexes are the way to go.
> This flexibility, however, comes with its own tradeoffs.
That sounds interesting for my use case, so I'm off to Fauxton
creating test docs and trying to declare my first `text` index:
> {
> "index": {
> "fields": [{"name": "test_textindexed", "type": "number"}]
> },
> "type": "text"
> }
To which Fauxton responds with error "text", and a 503 shows up
in the CouchDB log. Looks like I simply don't have the Lucene
engine plugged, but I'm at a loss as to how to enable it in a
post-CouchDB2-release world:
- I couldn't find any mention of it in the docs [3].
- I'm unsure what part of this Cloudant work is now merged
into CouchDB 2 vs. what still needs manual installation,
and my Googling yields mostly old posts/documentation.
* Should I follow the readme of couchdb-lucene? [4]
* Is this 2015 post "Enable FTS in CouchDB" [5] still relevant?
- I'm currently running couchdb2 via the official Docker image.
Should I give klaemo/couchdb-lucene [6] a try?
Thanks for your help :)
[1]
https://lists.apache.org/thread.html/1cd9fa2d063ca0f476d703e4f36fe62d366f9a2fd8910954dcd4798e@%3Cuser.couchdb.apache.org%3E
[2] https://cloudant.com/blog/mango-json-vs-text-indexes/
[3] http://docs.couchdb.org/en/2.0.0/
[4] https://github.com/rnewson/couchdb-lucene
[5] https://developer.ibm.com/dwblog/2015/text-search-apache-couchdb/
[6] https://github.com/klaemo/docker-couchdb-lucene
Re: How to prep my CouchDB 2 for *text* indexes/queries?
Posted by Ronan Jouchet <ro...@cadensimaging.com>.
On 2017-01-26 22:51, Joan Touzet wrote:
> I'm afraid you're going to have to custom compile CouchDB to
> use Lucene with it. More information on how to do this within
> the Docker environment is available at
>
> https://github.com/klaemo/docker-couchdb/issues/48#issuecomment-274323668
>
> which discusses an effort to add clouseau to the CouchDB 2.0
> Docker image and pair it with a secondary image that actually
> runs dreyfus and Lucene.
Fantastic, thanks for the pointer <insert xkcd joke here> and
confirmation that a custom compile is still the way to go.
> FYI there may be an "official Docker image" that has our blessing,
> but it's not a direct work product of Apache CouchDB -- Clemens,
> a long standing member of this community and a CouchDB committer,
> maintains this project on behalf of the community.
Yup, that was clear to me \U0001f44d. Thanks!
Re: How to prep my CouchDB 2 for *text* indexes/queries?
Posted by Joan Touzet <wo...@apache.org>.
Hi Ronan,
I'm afraid you're going to have to custom compile CouchDB to use
Lucene with it. More information on how to do this within the Docker
environment is available at
https://github.com/klaemo/docker-couchdb/issues/48#issuecomment-274323668
which discusses an effort to add clouseau to the CouchDB 2.0 Docker
image and pair it with a secondary image that actually runs dreyfus and
Lucene.
FYI there may be an "official Docker image" that has our blessing,
but it's not a direct work product of Apache CouchDB -- Clemens,
a long standing member of this community and a CouchDB committer,
maintains this project on behalf of the community.
-Joan
----- Original Message -----
> From: "Ronan Jouchet" <ro...@cadensimaging.com>
> To: "Joan Touzet" <wo...@apache.org>, "couchdb-user" <us...@couchdb.apache.org>
> Sent: Thursday, January 26, 2017 2:02:20 PM
> Subject: Re: How to prep my CouchDB 2 for *text* indexes/queries?
>
> On 2017-01-26 13:41, Joan Touzet wrote:
> > See https://cloudant.com/blog/open-sourcing-cloudant-search/
> > for all the details.
> Hi Joan.
>
> Thanks, that's helpful as a historical perspective. But that's not
> what
> I'm looking for. My email provided some context, but it looks like it
> must have muddied the core question. Let me re-iterate in a drier
> way.
>
> I'm currently using couchdb2 via the official Docker image. What's
> the
> easiest way to plug Lucene? Is there end-to-end documentation?
>
Re: How to prep my CouchDB 2 for *text* indexes/queries?
Posted by Ronan Jouchet <ro...@cadensimaging.com>.
On 2017-01-26 13:41, Joan Touzet wrote:
> See https://cloudant.com/blog/open-sourcing-cloudant-search/
> for all the details.
Hi Joan.
Thanks, that's helpful as a historical perspective. But that's not what
I'm looking for. My email provided some context, but it looks like it
must have muddied the core question. Let me re-iterate in a drier way.
I'm currently using couchdb2 via the official Docker image. What's the
easiest way to plug Lucene? Is there end-to-end documentation?
Re: How to prep my CouchDB 2 for *text* indexes/queries?
Posted by Joan Touzet <wo...@apache.org>.
Hi Roman,
See https://cloudant.com/blog/open-sourcing-cloudant-search/ for
all the details.
-Joan
----- Original Message -----
> From: "Ronan Jouchet" <ro...@cadensimaging.com>
> To: "couchdb-user" <us...@couchdb.apache.org>
> Sent: Thursday, January 26, 2017 12:06:51 PM
> Subject: How to prep my CouchDB 2 for *text* indexes/queries?
>
> Hi!
>
> Following mailing list discussion [1], I am looking at the
> performance of various Mango ways to fetch a bunch of docs.
>
> Following Garren links, I stumbled upon a blog post called
> "Understanding Mango View-Based Indexes vs. Search-Based Indexes"
> [2].
> There, the author mentions that:
>
> > If users don't know in advance what queries will be executed,
> > then Mango search-based indexes are the way to go.
> > This flexibility, however, comes with its own tradeoffs.
>
> That sounds interesting for my use case, so I'm off to Fauxton
> creating test docs and trying to declare my first `text` index:
>
> > {
> > "index": {
> > "fields": [{"name": "test_textindexed", "type": "number"}]
> > },
> > "type": "text"
> > }
>
> To which Fauxton responds with error "text", and a 503 shows up
> in the CouchDB log. Looks like I simply don't have the Lucene
> engine plugged, but I'm at a loss as to how to enable it in a
> post-CouchDB2-release world:
>
> - I couldn't find any mention of it in the docs [3].
>
> - I'm unsure what part of this Cloudant work is now merged
> into CouchDB 2 vs. what still needs manual installation,
> and my Googling yields mostly old posts/documentation.
> * Should I follow the readme of couchdb-lucene? [4]
> * Is this 2015 post "Enable FTS in CouchDB" [5] still relevant?
>
> - I'm currently running couchdb2 via the official Docker image.
> Should I give klaemo/couchdb-lucene [6] a try?
>
> Thanks for your help :)
>
> [1]
> https://lists.apache.org/thread.html/1cd9fa2d063ca0f476d703e4f36fe62d366f9a2fd8910954dcd4798e@%3Cuser.couchdb.apache.org%3E
>
> [2] https://cloudant.com/blog/mango-json-vs-text-indexes/
>
> [3] http://docs.couchdb.org/en/2.0.0/
>
> [4] https://github.com/rnewson/couchdb-lucene
>
> [5] https://developer.ibm.com/dwblog/2015/text-search-apache-couchdb/
>
> [6] https://github.com/klaemo/docker-couchdb-lucene
>