[
https://issues.apache.org/jira/browse/ATLAS-5416?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rachid Azgaou updated ATLAS-5416:
---------------------------------
Description:
*Goal* Integrate a vector-based semantic search to enable natural language
search queries alongside existing keyword matching.
*Key Requirements*
* *Vector Embedding Generation:* Generate embeddings for core content entities
on creation and update using an embedding model API
* *Vector Database Setup:* Set up and configure a vector store (OpenSearch)
for embedding storage and fast similarity retrieval.
* *Search Integration:* Integrate with existing workflow by translating
incoming natural language queries into embeddings, running a k-NN (k-nearest
neighbors) lookup against OpenSearch, and returning ranked results strictly
based on vector distance scores
* *API Endpoint Update:* Expose semantic search capabilities via the search
API.
* New exposed {*}APIs{*}:
{code:java}
api/atlas/v2/search/semantic
api/atlas/v2/search/similar
{code}
was:
*Goal* Integrate a vector-based semantic search to enable natural language
search queries alongside existing keyword matching.
*Key Requirements*
* *Vector Embedding Generation:* Generate embeddings for core content entities
on creation and update using an embedding model API
* *Vector Database Setup:* Set up and configure a vector store (OpenSearch)
for embedding storage and fast similarity retrieval.
* *Search Integration:* Integrate with existing workflow by translating
incoming natural language queries into embeddings, running a k-NN (k-nearest
neighbors) lookup against OpenSearch, and returning ranked results strictly
based on vector distance scores
* *API Endpoint Update:* Expose semantic search capabilities via the search
API.
> Introduce semantic search
> -------------------------
>
> Key: ATLAS-5416
> URL: https://issues.apache.org/jira/browse/ATLAS-5416
> Project: Atlas
> Issue Type: New Feature
> Components: atlas-intg
> Reporter: Rachid Azgaou
> Priority: Major
>
> *Goal* Integrate a vector-based semantic search to enable natural language
> search queries alongside existing keyword matching.
> *Key Requirements*
> * *Vector Embedding Generation:* Generate embeddings for core content
> entities on creation and update using an embedding model API
> * *Vector Database Setup:* Set up and configure a vector store (OpenSearch)
> for embedding storage and fast similarity retrieval.
> * *Search Integration:* Integrate with existing workflow by translating
> incoming natural language queries into embeddings, running a k-NN (k-nearest
> neighbors) lookup against OpenSearch, and returning ranked results strictly
> based on vector distance scores
> * *API Endpoint Update:* Expose semantic search capabilities via the search
> API.
> * New exposed {*}APIs{*}:
>
> {code:java}
> api/atlas/v2/search/semantic
> api/atlas/v2/search/similar
> {code}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
