2024-08-30 17:47:35 +01:00
|
|
|
|
{
|
|
|
|
|
|
pattern: {
|
|
|
|
|
|
name: "override-recursive-chunker",
|
|
|
|
|
|
icon: "✂️🪚",
|
|
|
|
|
|
title: "Replace default chunker with recursive chunker",
|
|
|
|
|
|
description: "The default chunker used in Trustgraph core is a token-based chunker. This pattern replaces that with a recursive chunker, and allows ou to configure the chunking parameters.",
|
|
|
|
|
|
requires: ["pulsar", "trustgraph"],
|
|
|
|
|
|
features: [],
|
|
|
|
|
|
args: [
|
|
|
|
|
|
{
|
|
|
|
|
|
name: "chunk-size",
|
2024-09-01 21:29:29 +01:00
|
|
|
|
type: "integer",
|
2024-08-30 17:47:35 +01:00
|
|
|
|
description: "Chunk size value",
|
|
|
|
|
|
default: 2000,
|
|
|
|
|
|
required: true,
|
|
|
|
|
|
},
|
|
|
|
|
|
{
|
|
|
|
|
|
name: "chunk-overlap",
|
2024-09-01 21:29:29 +01:00
|
|
|
|
type: "integer",
|
2024-08-30 17:47:35 +01:00
|
|
|
|
description: "Overlap size value",
|
|
|
|
|
|
default: 100,
|
|
|
|
|
|
required: true,
|
|
|
|
|
|
}
|
2024-09-01 21:29:29 +01:00
|
|
|
|
],
|
|
|
|
|
|
category: [ "chunking" ],
|
2024-08-30 17:47:35 +01:00
|
|
|
|
},
|
|
|
|
|
|
module: "components/cassandra.jsonnet",
|
|
|
|
|
|
}
|