tika-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris A. Mattmann (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (TIKA-1319) Translation
Date Wed, 04 Jun 2014 20:51:03 GMT

    [ https://issues.apache.org/jira/browse/TIKA-1319?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14018158#comment-14018158

Chris A. Mattmann commented on TIKA-1319:

Hmm...I'm not so convinced this should be a separate module. When we started Tika, the goal
was for language detection to be part of the core and it is there now - language translation
seems to be close enough of that original intent, I wonder why we wouldn't make it part of
the core. I share [~tallison@apache.org]'s concern of tying core to a web service, which is
why I suggested turning this into an interface.

I've got the following suggestion:

# Make the Translator interface part of tika-core and put in org.apache.tika.language.translate
# Create a tika-translate module that includes the 1 specific implementation linking to Microsoft's
# Bubble up anything that seems generic across implementations in tika-translate to the tika-core


> Translation
> -----------
>                 Key: TIKA-1319
>                 URL: https://issues.apache.org/jira/browse/TIKA-1319
>             Project: Tika
>          Issue Type: New Feature
>            Reporter: Tyler Palsulich
>            Assignee: Chris A. Mattmann
>            Priority: Minor
> I just opened up a review on reviews.apache.org -- https://reviews.apache.org/r/22219/.
I copied the description below. 
> This patch adds basic language translation functionality to Tika. Translation is provided
by a Microsoft API, but accessed through Apache 2 licensed com.memetix.microsoft-translator-java-api
(https://code.google.com/p/microsoft-translator-java-api/ ). If a user wants to use the translation
feature, they have to add a client id and client secret to the tika-core/src/main/resources/org/apache/tika/language/translator.properties
file (see http://msdn.microsoft.com/en-us/library/hh454950.aspx ). I added com.memetix as
a dependency in tika-core. I put the Translator class in org.apache.tika.language. There is
no integration with the server or CLI, yet. Further, only Strings are translated right now
-- if you pass in a full document with xml tags, the structure will be mangled. But, I think
that would be a cool feature -- translate the body, title, subtitle, etc, but not the structural
> There is still more work to do, but I wanted some more eyes on this to make sure I'm
heading in the right direction and this is a desired feature. Let me know what you think!
> There are two simple unit tests for now which translate "hello" to French ("salut").
One for inputting the source and target languages, one for inputing just the target language
(and detecting the source language automatically).

This message was sent by Atlassian JIRA

View raw message