Crate nodedb_types 

Shared NodeDB type definitions: strong IDs (TenantId, VShardId, RequestId, Lsn, Surrogate), error types (NodeDbError + ErrorCode + ErrorDetails), wire types for cross-plane and cross-node messages, value types, configuration schemas, and the approximate-aggregate sketches.

Every NodeDB workspace crate depends on this. Items here form the stable cross-crate vocabulary; types specific to one engine live in that engine’s crate.

Re-exports

pub use approx::CountMinSketch;

pub use approx::HyperLogLog;

pub use approx::SpaceSaving;

pub use approx::TDigest;

pub use array_cell::ArrayCell;

pub use array_route::ShardBucket;

pub use ascii::find_ascii_case_insensitive;

pub use ascii::find_ascii_case_insensitive_from;

pub use ascii::rfind_ascii_case_insensitive;

pub use ascii::starts_with_ascii_case_insensitive;

pub use ascii::strip_prefix_ascii_case_insensitive;

pub use audit_dml::AuditDmlMode;

pub use bbox::BoundingBox;

pub use bbox::geometry_bbox;

pub use clone::CloneOrigin;

pub use clone::CloneStatus;

pub use clone::MAX_CLONE_DEPTH;

pub use collection::CollectionType;

pub use collection::CollectionTypeParseError;

pub use collection_config::KeySpec;

pub use collection_config::PartitionStrategy;

pub use collection_config::PayloadAtom;

pub use collection_config::PayloadIndexKind;

pub use collection_config::PrimaryEngine;

pub use collection_config::VectorPrimaryConfig;

pub use columnar::ColumnDef;

pub use columnar::ColumnType;

pub use columnar::ColumnarProfile;

pub use columnar::ColumnarSchema;

pub use columnar::DocumentMode;

pub use columnar::SchemaError;

pub use columnar::StrictSchema;

pub use config::TuningConfig;

pub use crdt_preview::CrdtPreviewResult;

pub use datetime::NdbDateTime;

pub use datetime::NdbDateTimeError;

pub use datetime::NdbDuration;

pub use document::Document;

pub use dropped_collection::DroppedCollection;

pub use error::NodeDbError;

pub use filter::EdgeFilter;

pub use filter::MetadataFilter;

pub use graph::Direction;

pub use graph::GraphStats;

pub use hlc::ClockSkew;

pub use hlc::Hlc;

pub use hlc::HlcClock;

pub use hlc::MAX_CLOCK_SKEW_NS;

pub use hnsw::HnswCheckpoint;

pub use hnsw::HnswNodeSnapshot;

pub use hnsw::HnswParams;

pub use id::CollectionId;

pub use id::CollectionKey;

pub use id::CollectionKeyError;

pub use id::DatabaseId;

pub use id::DocumentId;

pub use id::EdgeId;

pub use id::EdgeIdParseError;

pub use id::IdError;

pub use id::IdType;

pub use id::NodeId;

pub use id::QualifiedCollection;

pub use id::ShapeId;

pub use id::TenantId;

pub use identity::KeyRepr;

pub use json_msgpack::InstantKind;

pub use json_msgpack::JsonValue;

pub use json_msgpack::MsgpackError;

pub use json_msgpack::MsgpackResult;

pub use json_msgpack::NativeCell;

pub use json_msgpack::json_from_msgpack;

pub use json_msgpack::json_to_msgpack;

pub use json_msgpack::json_to_msgpack_or_empty;

pub use json_msgpack::msgpack_to_json_string;

pub use json_msgpack::read_instant;

pub use json_msgpack::value_from_msgpack;

pub use json_msgpack::value_to_msgpack;

pub use json_msgpack::write_instant;

pub use kv::KV_DEFAULT_INLINE_THRESHOLD;

pub use kv::KvConfig;

pub use kv::KvTtlPolicy;

pub use kv::is_valid_kv_key_type;

pub use lsn::Lsn;

pub use mirror::MirrorLagRecord;

pub use mirror::MirrorMode;

pub use mirror::MirrorOrigin;

pub use mirror::MirrorStatus;

pub use multi_vector::MultiVector;

pub use multi_vector::MultiVectorError;

pub use multi_vector::MultiVectorScoreMode;

pub use namespace::Namespace;

pub use path_component::is_plain_path_component;

pub use quota::PriorityClass;

pub use quota::PriorityClassParseError;

pub use quota::QuotaRecord;

pub use quota::QuotaSpec;

pub use quota::QuotaValidationError;

pub use result::QueryResult;

pub use result::SearchResult;

pub use result::SubGraph;

pub use rls_write_check::RlsWriteCheck;

pub use rls_write_check::WriteGateDecision;

pub use row_identity::DEFAULT_IDENTITY_COLUMN;

pub use row_identity::HEADLESS_SENTINEL_PREFIX;

pub use row_identity::ROWID_COLUMN;

pub use row_identity::RowIdentity;

pub use row_identity::StorageKey;

pub use row_identity::declared_key;

pub use row_identity::extract_pk_value;

pub use row_identity::value_to_pk_string;

pub use sparse_vector::SparseVector;

pub use sparse_vector::SparseVectorError;

pub use sql_quote::quote_ident;

pub use sql_quote::quote_literal;

pub use surrogate::Surrogate;

pub use surrogate_bitmap::SurrogateBitmap;

pub use sync::compensation::CompensationHint;

pub use sync::shape::ShapeDefinition;

pub use sync::shape::ShapeType;

pub use sync::violation::ViolationType;

pub use sync::wire::SyncFrame;

pub use sync::wire::SyncMessageType;

pub use temporal::BitemporalFilter;

pub use temporal::BitemporalInterval;

pub use temporal::LsnTimeAnchor;

pub use temporal::LsnTimeError;

pub use temporal::LsnTimeMap;

pub use temporal::MAX_POSITIONS_PER_EPOCH;

pub use temporal::NANOS_PER_MS;

pub use temporal::OPEN_UPPER;

pub use temporal::OrdinalClock;

pub use temporal::SystemTimeScope;

pub use temporal::ValidTimePredicate;

pub use temporal::calvin_txn_ordinal;

pub use temporal::ms_to_ordinal_upper;

pub use temporal::ordinal_to_ms;

pub use text_search::Bm25Params;

pub use text_search::QueryMode;

pub use text_search::TextColumnFault;

pub use text_search::TextSearchParams;

pub use trace::SpanId;

pub use trace::TraceId;

pub use typeguard::TypeGuardFieldDef;

pub use value::NotScalar;

pub use value::Value;

pub use value::scalar_to_raw_bytes;

pub use vector_ann::VectorAnnOptions;

pub use vector_ann::VectorQuantization;

pub use vector_dtype::VectorStorageDtype;

pub use vector_index_params::StoredVectorIndexParams;

pub use vector_index_stats::VectorIndexQuantization;

pub use vector_index_stats::VectorIndexStats;

pub use vector_index_stats::VectorIndexType;

pub use vector_index_stats::VectorIvfStats;

pub use vector_model::VectorModelEntry;

pub use vector_model::VectorModelMetadata;

pub use volatility::Volatility;

pub use write_version::ShardVersion;

pub use write_version::WriteVersion;

Modules

approx

array_cell

Typed payload for Value::ArrayCell.

array_route

Routing bucket and vShard of an array tile.

ascii

Byte-stable ASCII matching for untrusted text.

audit_dml

AuditDmlMode — per-database DML audit level.

backup_envelope

bbox

Axis-aligned bounding box for 2D geometries (WGS-84).

calvin

Primitive Calvin scheduling types shared between nodedb-physical (the physical-plan IR layer) and nodedb-cluster (the distributed Calvin sequencer / scheduler).

calvin_passive

Identity of one row a passive Calvin participant reads.

clone

Clone catalog types shared between the catalog layer and the control plane.

collection

Collection type enum shared between Origin and Lite.

collection_config

Vector-primary collection configuration types, plus partition-strategy metadata.

columnar

config

conversion

Conversions between serde_json::Value and nodedb_types::Value.

crdt_preview

Wire result for a bounded, non-mutating CRDT delta preview.

datetime

First-class DateTime and Duration types.

decode_bounds

Dependency-neutral allocation bounds for untrusted binary decoders.

diagnostic

Canonical diagnostic-layer taxonomy for end-to-end probes.

document

Document type: the universal data container for CRDT-backed records.

dropped_collection

DroppedCollection — row shape for _system.dropped_collections.

error

Standardized error types for the NodeDB public API.

fail_point

Minimal feature-gated fail-point framework for crash-injection tests.

filter

Query filter types for vector search and graph traversal.

geometry

GeoJSON-compatible geometry types.

graph

Shared graph types used by both Origin and Lite CSR engines.

hlc

Hybrid Logical Clock.

hnsw

Shared HNSW types used by both Origin and Lite vector engines.

id

id_gen

First-class ID type generation and validation.

identity

Plane-neutral row identity.

json_msgpack

kv

KV engine configuration types.

kv_parsing

Shared KV DDL parsing helpers used by both Origin (pgwire) and Lite.

lsn

Log Sequence Number — monotonically increasing identifier for WAL records.

mirror

Mirror catalog types shared between the catalog layer and the control plane.

multi_vector

Multi-vector type for ColBERT-style late interaction retrieval.

namespace

Storage namespace identifiers for the blob KV store.

numeric_cmp

Numeric readings and their total order, shared by every comparison path.

path_component

Validation for names that become a single filesystem path component.

pg_compat

PostgreSQL wire-compatibility version constants. NodeDB advertises compatibility with a fixed PostgreSQL major so libpq-based clients gate features correctly. These are the single source of truth for version(), the server_version_num runtime parameter, and startup ParameterStatus.

protocol

quota

Database and tenant resource quota types.

result

Query result types returned by the NodeDb trait methods.

rls_write_check

The compiled row-level-security WRITE predicate carried on a write plan.

row_identity

This module owns both encodings of a row’s surrogate identity.

sparse_vector

Sparse vector type for learned sparse retrieval (SPLADE, SPLADE++).

sql_quote

SQL-standard quoting for identifiers and string literals.

surrogate

Surrogate — a global, stable, monotonically-allocated u32 row identity.

surrogate_bitmap

sync

temporal

text_search

Text search parameter types shared across the NodeDb trait boundary.

timeseries

Shared timeseries types for multi-model database engines.

trace

W3C-compatible 128-bit trace identifiers and 64-bit span identifiers.

typeguard

Type guard definitions for schemaless collection write-time validation.

value

vector_ann

Runtime ANN tuning options threaded from SQL planner → Data Plane.

vector_distance

Shared scalar distance metric implementations.

vector_dtype

Vector storage precision tag.

vector_index_params

Durable per-collection/column vector-index parameters.

vector_index_stats

Vector index statistics returned by SHOW VECTOR INDEX status ON collection.column.

vector_model

Per-column vector model metadata for embedding tracking and migration.

volatility

Function volatility: how far a call’s result can be reused.

wire_version

Single source of truth for the WIRE_FORMAT_VERSION constant and the WIRE_BUILD_ID build identity, shared between every crate that needs to stamp or interpret them.

write_version

The version a committed write gives the rows it writes.

Macros

fail_point

Inject a fail point. Does nothing at runtime without the failpoints feature.

fail_point_err

Inject a fail point that can abort the enclosing function with an error.