Struct Document 

#[non_exhaustive]
pub struct Document {
    pub id: DocumentId,
    pub content: String,
    pub metadata: Value,
    pub source: Source,
    pub lineage: Option<Lineage>,
    pub namespace: Namespace,
}

The unit a RAG pipeline moves around — content plus everything downstream needs to know about where it came from.

Splitters consume one Document and emit several. Chunkers consume a sequence and emit a transformed sequence (typically the same length, with mutated content or metadata). Loaders produce them; ingestion pipelines consume them.

metadata is operator-defined free-form JSON for filtering at the vector-store layer — explicit fields above (source, lineage, namespace) are the SDK-stamped boundary that every pipeline must preserve.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

id: DocumentId

Stable identifier for this document within its namespace.

content: String

The textual content. Splitters slice this; chunkers may rewrite it (Contextual Retrieval prepends a generated context prefix).

metadata: Value

Operator-supplied free-form metadata. Vector stores typically expose this as a filterable JSON column.

source: Source

Origin of this document or its top-level ancestor.

lineage: Option<Lineage>

Split / chunk ancestry. None on the loader-produced root; Some on every chunked descendant.

namespace: Namespace

Multi-tenant boundary (invariant 11). Every persistent boundary the pipeline crosses respects this — silent cross-tenant leakage is structurally impossible because every loader, splitter, and pipeline takes a Namespace at construction time.

Implementations

impl Document

pub fn root( id: impl Into<DocumentId>, content: impl Into<String>, source: Source, namespace: Namespace, ) -> Self

Construct a fresh root document — the shape a crate::DocumentLoader emits before any splitter has run. lineage is None; chunked descendants populate it.

pub fn with_metadata(self, metadata: Value) -> Self

Builder-style metadata setter.

pub fn child(&self, content: impl Into<String>, lineage: Lineage) -> Self

Derive a child document for a chunk. The splitter supplies the per-chunk content + lineage; the new id is the parent’s id with a :{chunk_index} suffix so leaves carry hierarchical identifiers stable across re-runs.

Trait Implementations

impl Clone for Document

fn clone(&self) -> Document

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Document

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Document

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for Document

fn eq(&self, other: &Document) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for Document

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for Document

impl StructuralPartialEq for Document