question

richardm-3341 avatar image
0 Votes"
richardm-3341 asked Grmacjon-MSFT commented

Azure Search: How to add unsupported Language support

So, with ElasticSearch Managed service you can add your own plugins or bundles - these can be stemmers, dictionaries etc and this functionality enables you to support FTS for languages that have no support out of the box.

Does Azure Cognitive Search support any way to add new language support - in a similar way to this, or any other way?

azure-cognitive-search
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Grmacjon-MSFT avatar image
0 Votes"
Grmacjon-MSFT answered Grmacjon-MSFT commented

Hi @richardm-3341,
Just to clarify are you referring to natural language support? for Unsupported language deployments please see this doc: https://docs.microsoft.com/en-us/azure/cognitive-services/translator/custom-translator/unsupported-language-deployments


-Grace




· 3
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

We are not looking at Natural Language support - we are looking to implement a FTS system on AZURE search, in a fairly standard manner.

To accomplish this for ElasticSearch we would write a stemmer for example, and be able to deploy it as a plugin (see: https://www.elastic.co/guide/en/cloud/current/ec-custom-bundles.html#ec-custom-bundles ) for the ElasticSearch instance. My question is about whether AZURE Search supports similar or other methods to add support and functionality for an unsupported language.

0 Votes 0 ·

Hey @Grmacjon-MSFT - do you have a response to the above?

0 Votes 0 ·

Hi @richardm-3341,

Just checking to see if the below response helped answered your question. If you have additional questions please let us know.

-Grace

0 Votes 0 ·
Grmacjon-MSFT avatar image
0 Votes"
Grmacjon-MSFT answered

Hi @richardm-3341,

Thanks for your patience and for providing more clarity about your scenario.

Although Azure Cognitive Search doesn’t support custom plugin integrations, there is rich language support for lexical analysis of content. This includes 35 language analyzers from Lucene and 50 in-house Microsoft ones which are building using Office and Bing NLP technology. These analyzers can be defined as field properties in the index schema letting the user how the content is indexed and queried. You can find the complete list of languages here - Add language analyzers to string fields - Azure Cognitive Search | Microsoft Docs

In addition, you can customize the analyzers by combining different analysis components like tokenizers, char filters, and token filters to achieve the exact scenario required for the application. Please see more details here - Add custom analyzers to string fields - Azure Cognitive Search | Microsoft Docs

If this solution doesn't meet the requirements for your scenario, let us know so we can take this conversation offline. Please send us an email to azcommunity@microsoft.com with the link to this question and the subject "Attn: Grace".

Best,
Grace

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.