If you want your project to be featured in the blogpost on the Google OSPO blog, kindly comment back with a short, catchy description of the project along with a screenshot.

In T109307#1559719, @NiharikaKohli wrote:

@Phoenix303, there are a bunch of other links you need to provide here. :)

According to Niklas, language search should index all the words(using default analyzer), therefore, we do not need language analyzer now.

With some more testing and digging deeper into the features of ElasticSearch, I have found few conclusions we made earlier not to be completely true.

• ElasticSearch does not score phrase matches higher( not found any document to support ), but scores AND higher than OR. In other words, even with the default 'OR' search ES would score documents which match all the query words high. However, this is only true for each shard. ES does not score based on the entire index (globally) Ref.
• An example in translatewiki.net and with the offset=150.

Something like this. I have used an alphabet 'X' here. Does the positioning look good? what about the color, I think a different color like red would look better?

In T92929#1527730, @NiharikaKohli wrote:

A few questions (for both mentors and student):

Are you confident in completing the project on time?

Yes, the project is almost complete. It is available here

By when do you think you can merge the code, if at all?

The code review is in progress by @Nikerabbit.

This is taken care of by the existing update().

Thanks @Nemo_bis for the feedback.

See T106931

I would like to propose a design for Cross language search.

Searching any word in the string is the default behaviour which is already supported in the product.

Thanks @Nemo_bis
We have been able to provide a user to search in the source language which defaults to the wiki's content language and get translations in their language. We have not come up yet with any interface to let users select a source language or tell user what source language is searched in, however, a user can force the search in any source language with URL parameter 'sourcelanguage'.
One solution to find the source language is to detect the language of the search string using a language detection plugin, if one already exists? The another solution can be to let a user select it while typing a search string, see the attached mockup.

@Nemo_bis: Source language is English for http://translation-search.wmflabs.org/w/index.php?title=Special:SearchTranslations&filter=translated&language=ksh&query=edit&string=edit&uselang=en
and Finnish for http://translation-search.wmflabs.org/w/index.php?title=Special:SearchTranslations&filter=translated&language=ksh&query=edit&string=edit&uselang=fi right?

Thank you so much @Nemo_bis @Nikerabbit :)

@Nemo_bis: Thank you. Will recheck all the smaller bugs found.

How about meeting this Wednesday?

Hi @NiharikaKohli, I didn't realize that this task was not up-to-date. Thanks for letting me know :)

Patch https://gerrit.wikimedia.org/r/#/c/216958/ submitted.

@Nemo_bis: http://translation-search.wmflabs.org/w/index.php?title=Special:SearchTranslations&filter=translated&language=it&query=edit currently displays text only in the source language. I will make a change to support this feature too.

The problem is that MessageGroups::getGroupStructure(); returns aggregate groups for only wiki 1 (relation between group and subgroups) and do not know anything about wiki 2. Since translations for wiki 2 exist in the shard, It would be possible to list filtered message groups but don't know of any possible way to map groups and subgroups relationship.

@Nemo_bis @Nikerabbit: As per discussion I have updated the timeline. Please let me know if I have missed any point.