Hi,

I agree with you Vasyl Kaigorodov. Even I had the same feeling when I saw the mail about the new search engine. We could use the existing search engines. They are doing a good job.

Regards, *seran.*

On 29 August 2015 at 18:17, Vasyl Kaigorodov vasyl@redhat.com wrote:

Hi folks,

I did follow the discussion during the meeting, and was hesitating to ask: why exactly do we want own search implemented? From my experience, Google or DuckDuck do index Fedora resources just fine, and whenever I want to find something - I just use those. Own search engine adds to the cost of support and maintenance, with questionable benefits. Unless we have some public resources that have tricky robots.txt and are not indexed by other search engines - this looks like a waste of resources to me. I might miss something though, so any insight is welcome.

PS: there're 2 messages in the list moderation queue from me that could be deleted - was fighting with Zimbra identities, sorry for spam.

Thanks.

vk

----- Original Message ----- From: "Paweł Kłos" pawel@pawelklos.eu To: "Fedora Infrastructure" infrastructure@lists.fedoraproject.org Sent: Thursday, August 27, 2015 9:15:09 PM Subject: Fedora Search Engine

Hi,

The idea of Fedora Search engine has came back with discussion about https://fedorahosted.org/fedora-infrastructure/ticket/1055 ticket on the meeting.

Many people was truly interested in participating in that project.

So, if somebody is still interested I propose to make some brainstorm about that matter, so we could reach better results than years ago :)

First we can discuss in that mail thread and move then to irc or some other place.

Regards, Pawel Klos _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org

http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject.... _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org

http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

_______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

On Sat, 29 Aug 2015 11:08:22 -0300 Emiliano Dalla Verde Marcozzi edvm@fedoraproject.org wrote:

Custom Search Engine from Google. It lets you integrate Google search engine into your site and tweak it a bit, more info at: https://cse.google.com/cse/all The integration is super easy, just some lines of javascript. Best, Emiliano.

Well, thats fine, except it's not open source and it's not free (in either sense). So, its a no go.

kevin

_______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

On Thu, 3 Sep 2015 06:45:20 +0800 Christopher Meng i@cicku.me wrote:

On Thursday, September 3, 2015, Kevin Fenzi kevin@scrye.com wrote:

...

Well, thats fine, except it's not open source and it's not free (in either sense). So, its a no go.

Yeah but Firefox uses it still.

Sure. Not sure what that has to do with Fedora Infrastructure tho.

It's their project, they are welcome to do whatever they like.

(also, newer firefox versions default to yahoo search)

kevin

_______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

Hi,

How about using elasticsearch ( https://www.elastic.co/ ). It is built on top of apache lucene with a powerful json based query system.

On Fri, Sep 4, 2015 at 11:31 PM, Zach Villers greyman@zanshin.xyz wrote:

Here are a few possibilities I've found so far;

Apache Solr - http://lucene.apache.org/solr/

  Features - "Solr is a standalone enterprise search server with a

REST-like API. You put documents in it (called "indexing") via JSON, XML, CSV or binary over HTTP. You query it via HTTP GET and receive JSON, XML, CSV or binary results."

  It's a java servlet, which may be a mark against it. There are

docker images for it. To try out the Velocity Search engine, move the Velocity .jar files into the server library, use the techproducts solrconfig.xml, and start the server with bin/solr start -e techproducts. You can copy in some html files to the exampledocs directory using wget to mirror a site, then index the files with bin/post command.

Apache Spark - http://spark.apache.org/

"Apache Spark™ is a fast and general engine for large-scale data

processing."

 Here is a use case for someone using it as a search engine;

http://www.maana.io/company/press-release/the-first-and-only-big-data-search...

 Just in scanning the docs, you can run it with java or python.

OpenSearchServer - http://www.opensearchserver.com/

 Also runs on java - I haven't tried it out yet.

Sphinx - Free open-source SQL full-text search engine -

 http://sphinxsearch.com/docs/current.html
 (Haven't tried this either)

Here is a link to a recent tech blog with few other possibilities; http://www.mytechlogy.com/IT-blogs/8685/tech-blogs-top-5-open-source-search-...

-- irc #aikidouke _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org

http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

_______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org http://lists.fedoraproject.org/postorius/infrastructure@lists.fedoraproject....

On Sat, 29 Aug 2015, Kevin Fenzi wrote:

free/open source service. Also, if we want to adjust results (say to favor current releases) we may not have control to do that.

Hm, adjusting results is the feature I did not think of, that indeed might be useful.

The parts of the project that really expressed interest in a local search engine are docs and wiki. We could see if the new wiki version we are going to move to soon has a better search. The docs folks also are looking at moving to a new platform, perhaps it would have better search too...

So, what is that new wiki version will be, and what platform are docs folks are moving to? We could probably do some homework and see what those have in term of search, and then we can decide if we need a separate engine, or we can somehow tune wiki/docs to have better search.