RustUse/use-nosql

use-nosql is a RustUse facade workspace for small, focused Rust 2024 primitive crates around NoSQL-style data modeling.

RustUse is a collection of composable sets of primitive Rust utility crates for fellow crustaceans.

use-nosql covers document stores, key-value stores, wide-column stores, graph stores, cache stores, search indexes, vector stores, time-series stores, partitioning, consistency labels, and change events.

Experimental

use-nosql is experimental while the workspace remains below 0.3.0. Expect small API refinements during the first release wave.

Non-goals

use-nosql is not:

• a database driver
• a database client
• an ORM
• a connection pool
• a query execution engine
• a migration tool
• an async networking abstraction
• a vendor SDK wrapper

The crates avoid network calls, database connections, query execution, driver behavior, ORM behavior, migrations, and vendor-specific adapters.

Relationships

• use-nosql models generic NoSQL-style data modeling primitives.
• use-sql models generic SQL language, value, query, and schema primitives.
• use-postgres models PostgreSQL-specific metadata primitives.
• use-data models generic data format helpers.
• use-graph remains the home for graph-theory algorithms and general graph primitives.

Workspace crates

Crate	Path	Purpose
use-nosql	crates/use-nosql/	Feature-gated facade over the focused NoSQL primitive crates
use-document-store	crates/use-document-store/	Document-store identifiers, metadata, and patch operation primitives
use-key-value-store	crates/use-key-value-store/	Key, namespace, bucket, entry, pattern, and range primitives
use-wide-column	crates/use-wide-column/	Wide-column keyspace, family, row, column, and clustering primitives
use-graph-store	crates/use-graph-store/	Property-graph store vertex, edge, label, and property primitives
use-cache-store	crates/use-cache-store/	Cache key, namespace, entry, TTL, expiration, status, and eviction primitives
use-search-index	crates/use-search-index/	Search index document, field, term, analyzer, query-shape, sort, and filter primitives
use-vector-store	crates/use-vector-store/	Vector ID, embedding, dimension, metric, metadata, and record primitives
use-timeseries-store	crates/use-timeseries-store/	Series, metric, timestamp, point, retention, sampling, and aggregation primitives
use-document-path	crates/use-document-path/	Dot-path, segment, selector, and parse-error primitives
use-partition-key	crates/use-partition-key/	Partition, shard, routing, sort, composite key, and strategy primitives
use-consistency	crates/use-consistency/	Consistency, read concern, write concern, durability, replication, and quorum labels
use-change-event	crates/use-change-event/	Change event, event ID, cursor, resume token, sequence, and changed-document primitives

Installation

Use the workspace directly or depend on a Git revision until the first crates.io release is published.

[dependencies]
use-nosql = { git = "https://github.com/RustUse/use-nosql", rev = "<commit>" }

After publication, choose the narrowest focused crate that matches your use case or use the facade when one dependency is more convenient.

[dependencies]
use-nosql = "0.1.0"

Basic usage

use use_nosql::{CollectionName, DocumentId, DocumentPath, PatchOperation, PatchSet};

let collection = CollectionName::new("customers");
let document_id = DocumentId::new("customer_123");
let patch = PatchSet::new(vec![PatchOperation::set(
    DocumentPath::new("profile.display_name"),
    "Joshua Whalen",
)]);

assert_eq!(collection.as_str(), "customers");
assert_eq!(document_id.as_str(), "customer_123");
assert_eq!(patch.operations().len(), 1);

use use_nosql::{CacheKey, CacheNamespace};

let key = CacheKey::builder()
    .namespace(CacheNamespace::new("reviews"))
    .segment("google-business-profile")
    .segment("location")
    .segment("fort-wayne")
    .segment("summary")
    .build();

assert_eq!(key.to_string(), "reviews:google-business-profile:location:fort-wayne:summary");

use use_nosql::{Embedding, SimilarityMetric, VectorDimension, VectorId, VectorRecord};

let record = VectorRecord::new(
    VectorId::new("review_789_embedding"),
    Embedding::new(vec![0.012, -0.032, 0.481]),
)
.with_dimension(VectorDimension::new(3))
.unwrap()
.with_similarity_metric(SimilarityMetric::Cosine);

assert_eq!(record.dimension(), Some(VectorDimension::new(3)));

Development

cargo fmt --all -- --check
cargo clippy --workspace --all-targets --all-features -- -D warnings
cargo test --workspace --all-features
cargo test --workspace --no-default-features
cargo check --workspace --all-features --examples
cargo doc --workspace --all-features --no-deps

License

Licensed under either of the following, at your option:

• Apache License, Version 2.0
• MIT license