Details
-
Task
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
4.x
-
None
-
Unknown
Description
Currently, when loading data into a VectorDB, applications have to implement chunking on their own. This is typically done as a processor (which then uses LangChain4J Document Splitter features or implement their own).
We could make this simpler by implementing a specialized chunking DSL that abstracts this process and simplify loading the data to those databases.