NEWVectors or files. Pick a path.Start →
    Transformation

    JSON Transform

    Transform document payloads using JSONPath expressions and mapping rules

    Why do anything?

    Document schemas may not match consumer expectations. Transformation enables schema adaptation.

    Why now?

    Different consumers need different views of the same data.

    Why this feature?

    Transform document payloads with JSONPath selection, field mapping, and computed fields.

    How It Works

    JSON transform modifies document payloads during retrieval.

    1

    Rule Parsing

    Parse transformation rules

    2

    Field Selection

    Select fields with JSONPath

    3

    Mapping

    Apply field mappings and renames

    4

    Computation

    Compute derived fields

    Why This Approach

    JSONPath provides powerful selection. In-pipeline transformation avoids post-processing.

    Integration

    retriever = client.retrievers.create(stages=[{"type": "json_transform", "mappings": {...}}])