[
https://issues.apache.org/jira/browse/CAMEL-21540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Guillaume Nodet updated CAMEL-21540:
------------------------------------
Description:
New camel-pgvector component for vector similarity search using the PostgreSQL
pgvector extension. Provides a lightweight, SQL-native vector database option —
no separate infrastructure needed beyond PostgreSQL.
Example route (YAML DSL):
{code:yaml}
- route:
from:
uri: direct:index
steps:
- setVariable:
name: text
simple: "${body}"
- to:
uri: openai:embeddings
parameters:
embeddingModel: nomic-embed-text
- setHeader:
name: CamelPgVectorAction
constant: UPSERT
- setHeader:
name: CamelPgVectorTextContent
simple: "${variable.text}"
- to: pgvector:documents
{code}
Features:
* Actions: CREATE_TABLE, CREATE_INDEX (HNSW), DROP_TABLE, UPSERT, DELETE,
SIMILARITY_SEARCH
* Distance types: cosine (default), euclidean, inner product
* SQL WHERE clause filtering on text_content and metadata columns
* UUID auto-generation when no record ID is provided
* Upsert with ON CONFLICT DO UPDATE for idempotent writes
* Data type transformers: pgvector:embeddings and pgvector:rag for LangChain4j
integration
* OpenAI integration: direct chaining from openai:embeddings to pgvector
* Cross-documentation with camel-openai and camel-langchain4j-embeddings
was:
Implement Vector search capabilities for PostgreSQL.
Also add the langchain4j Embeddings DataFormat for PostgreSQL + an Integration
test in the langchain4j embeddings directory
[https://github.com/apache/camel/tree/main/components/camel-ai/camel-langchain4j-embeddings/src/test/java/org/apache/camel/component/langchain4j/embeddings]
> Vector Database capabilities - PostgreSQL
> -----------------------------------------
>
> Key: CAMEL-21540
> URL: https://issues.apache.org/jira/browse/CAMEL-21540
> Project: Camel
> Issue Type: New Feature
> Reporter: Zineb Bendhiba
> Assignee: Guillaume Nodet
> Priority: Major
>
> New camel-pgvector component for vector similarity search using the
> PostgreSQL pgvector extension. Provides a lightweight, SQL-native vector
> database option — no separate infrastructure needed beyond PostgreSQL.
> Example route (YAML DSL):
> {code:yaml}
> - route:
> from:
> uri: direct:index
> steps:
> - setVariable:
> name: text
> simple: "${body}"
> - to:
> uri: openai:embeddings
> parameters:
> embeddingModel: nomic-embed-text
> - setHeader:
> name: CamelPgVectorAction
> constant: UPSERT
> - setHeader:
> name: CamelPgVectorTextContent
> simple: "${variable.text}"
> - to: pgvector:documents
> {code}
> Features:
> * Actions: CREATE_TABLE, CREATE_INDEX (HNSW), DROP_TABLE, UPSERT, DELETE,
> SIMILARITY_SEARCH
> * Distance types: cosine (default), euclidean, inner product
> * SQL WHERE clause filtering on text_content and metadata columns
> * UUID auto-generation when no record ID is provided
> * Upsert with ON CONFLICT DO UPDATE for idempotent writes
> * Data type transformers: pgvector:embeddings and pgvector:rag for
> LangChain4j integration
> * OpenAI integration: direct chaining from openai:embeddings to pgvector
> * Cross-documentation with camel-openai and camel-langchain4j-embeddings
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
