uql-orm 0.2.10 → 0.3.1

package/CHANGELOG.md

@@ -3,13 +3,12 @@
 All notable changes to this project will be documented in this file.
 See [Conventional Commits](https://conventionalcommits.org) for commit guidelines.
 
-## [0.2.10](https://github.com/rogerpadilla/uql/compare/uql-orm@0.2.9...uql-orm@0.2.10) (2026-03-11)
+## [0.3.1](https://github.com/rogerpadilla/uql/compare/uql-orm@0.3.0...uql-orm@0.3.1) (2026-03-12)
 
 
 ### Features
 
-* Add SVG drawing animations to the full logo and update the README to display it. ([4515203](https://github.com/rogerpadilla/uql/commit/45152032021b2a21ee4bbdf74cc3a48097003cb4))
-* enhance $size operator to support comparison objects for array lengths and relation counts. ([02037fb](https://github.com/rogerpadilla/uql/commit/02037fbb6113efab9d23c9809370704b2de3747a))
+* implement MongoDB vector search functionality in dialect and querier ([2987574](https://github.com/rogerpadilla/uql/commit/2987574eaa10efc201ec17d0b6f72a144a1b49fb))
 
 
 
@@ -21,6 +20,28 @@ All notable changes to this project will be documented in this file. Please add
 
 date format is [yyyy-mm-dd]
 
+## [0.3.1] - 2026-03-12
+### New Features
+- **MongoDB Atlas Vector Search**: Semantic search now supports MongoDB via the [`$vectorSearch`](https://www.mongodb.com/docs/atlas/atlas-vector-search/vector-search-stage/) aggregation pipeline stage. Same `$sort` API — UQL internally translates to Atlas's native vector search with optimal pre-filtering (`$where` → `$vectorSearch.filter`), score projection via `$meta: 'vectorSearchScore'`, and secondary sort support. Configure with `@Index(['embedding'], { type: 'vectorSearch', name: 'my_index' })`.
+
+## [0.3.0] - 2026-03-12
+### New Features
+- **Semantic Search**: First-class vector similarity search across PostgreSQL (pgvector), MariaDB, and SQLite. Query via `$sort` on vector fields:
+  ```ts
+  const results = await querier.findMany(Article, {
+    $sort: { embedding: { $vector: queryVec, $distance: 'cosine' } },
+    $limit: 10,
+  });
+  ```
+  Supports 5 distance metrics (`cosine`, `l2`, `inner`, `l1`, `hamming`), distance projection via `$project`, and the `WithDistance<E>` utility type. Each dialect generates native SQL: Postgres operators (`<=>`, `<->`), MariaDB (`VEC_DISTANCE_COSINE()`), SQLite (`vec_distance_cosine()`).
+- **Vector Field Types**: `@Field({ type: 'vector', dimensions: 1536 })` for standard 32-bit embeddings, plus Postgres-specific `'halfvec'` (16-bit, 50% storage savings) and `'sparsevec'` (for SPLADE-style sparse embeddings). Automatic SQL mapping across dialects.
+- **Vector Indexes**: `@Index()` supports HNSW and IVFFlat index types with `distance`, `m`, `efConstruction`, and `lists` options. Generates pgvector operator classes for Postgres, inline `VECTOR INDEX` for MariaDB, and standard indexes for SQLite.
+- **Auto Extension Creation**: Schema generator automatically emits `CREATE EXTENSION IF NOT EXISTS vector` for Postgres tables containing vector columns.
+
+### Architecture
+- **Schema Generator Dialect Config Refactor**: Eliminated all `this.dialect ===` branches from `schemaGenerator.ts` by adding new declarative `dialectConfig` properties (`columnComment`, `vectorIndexStyle`, `dropIndexSyntax`, `renameTableSyntax`, `booleanLiteral`, `alterColumnStrategy`, `vectorOpsClass`, `vectorExtension`). All dialect-specific behavior is now config-driven.
+- **Unified CREATE INDEX**: `generateCreateIndexFromNode` now delegates to `generateCreateIndex`, eliminating duplicated SQL assembly and ensuring consistent vector index handling across both code paths.
+
 ## [0.2.7] - 2026-03-11
 ### New Features
 - **More `$size` Comparison Operators**: `$size` now accepts comparison operator objects in addition to exact numbers — e.g. `{ $size: { $gte: 2 } }`, `{ $size: { $gt: 0, $lte: 5 } }`, `{ $size: { $between: [1, 10] } }`. Supported operators: `$eq`, `$ne`, `$gt`, `$gte`, `$lt`, `$lte`, `$between`.

package/README.md

@@ -2,7 +2,7 @@
 
 <a href="https://uql-orm.dev"><img src="assets/logo.svg" alt="uql" width="80" /></a>
 
-[![tests](https://github.com/rogerpadilla/uql/actions/workflows/tests.yml/badge.svg)](https://github.com/rogerpadilla/uql) [![Coverage Status](https://coveralls.io/repos/github/rogerpadilla/uql/badge.svg?branch=main)](https://coveralls.io/github/rogerpadilla/uql?branch=main) [![license](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/rogerpadilla/uql/blob/main/LICENSE) [![npm version](https://img.shields.io/npm/v/uql-orm.svg)](https://www.npmjs.com/package/uql-orm)
+[![tests](https://github.com/rogerpadilla/uql/actions/workflows/tests.yml/badge.svg)](https://github.com/rogerpadilla/uql) [![Coverage Status](https://coveralls.io/repos/github/rogerpadilla/uql/badge.svg?branch=main)](https://coveralls.io/github/rogerpadilla/uql?branch=main) [![license](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/rogerpadilla/uql/blob/main/LICENSE) [![npm version](https://img.shields.io/npm/v/uql-orm.svg)](https://www.npmjs.com/package/uql-orm) [![Discord](https://img.shields.io/badge/Discord-Join%20Us-5865F2?logo=discord&logoColor=white)](https://discord.gg/DHJYp6MDS7)
 
 **[UQL](https://uql-orm.dev)** is the [smartest ORM](https://medium.com/@rogerpadillac/in-search-of-the-perfect-orm-e01fcc9bce3d) for TypeScript. It is engineered to be **fast**, **safe**, and **universally compatible**.
 
@@ -31,7 +31,8 @@ const users = await querier.findMany(User, {
 | **[Declarative Transactions](https://uql-orm.dev/querying/transactions)**  | Standard `@Transactional()` and `@InjectQuerier()` decorators for NestJS/DI.                                                |
 | **[Lifecycle Hooks](https://uql-orm.dev/entities/lifecycle-hooks)**| `@BeforeInsert`, `@AfterLoad` and 5 more decorators for validation, timestamps, and computed fields.                        |
 | **[Aggregate Queries](https://uql-orm.dev/querying/aggregate)** | `GROUP BY`, `HAVING`, `COUNT`, `SUM`, `AVG`, `MIN`, `MAX`, and `DISTINCT` across all dialects. |
-| **[Modern &amp; Versatile](https://uql-orm.dev/entities/virtual-fields)** | **Pure ESM**, high-res timing, [Soft-delete](https://uql-orm.dev/entities/soft-delete), and **Vector/JSONB/JSON** support. |
+| **[Semantic Search](https://uql-orm.dev/querying/semantic-search)** | Vector similarity via `$sort` with `$vector`/`$distance`. Supports `vector`, `halfvec`, `sparsevec` types, HNSW/IVFFlat indexes, and 5 distance metrics across Postgres, MariaDB, SQLite, and MongoDB Atlas. |
+| **[Modern & Versatile](https://uql-orm.dev/entities/virtual-fields)** | **Pure ESM**, high-res timing, [Soft-delete](https://uql-orm.dev/entities/soft-delete), and **JSONB/JSON** support. |
 | **[Database Migrations](https://www.uql-orm.dev/migrations)**          | Built-in [Entity-First synchronization](https://uql-orm.dev/migrations#3-entity-first-synchronization-development) and a robust CLI for version-controlled schema evolution. |
 | **[Logging & Monitoring](https://www.uql-orm.dev/logging)**               | Professional-grade monitoring with slow-query detection and colored output.                                                     |
 
@@ -101,7 +102,7 @@ UQL separates the **intent** of your data from its **storage**. Both properties
 
 | Property | Purpose | Values |
 | :--- | :--- | :--- |
-| **`type`** | **Logical Type** (Abstraction). Used for runtime behavior and automatic SQL mapping. | `String`, `Number`, `Boolean`, `Date`, `BigInt`, or semantic strings: `'uuid'`, `'json'`, `'vector'`. |
+| **`type`** | **Logical Type** (Abstraction). Used for runtime behavior and automatic SQL mapping. | `String`, `Number`, `Boolean`, `Date`, `BigInt`, or semantic strings: `'uuid'`, `'json'`, `'vector'`, `'halfvec'`, `'sparsevec'`. |
 | **`columnType`** | **Physical Type** (Implementation). **Highest Priority**. Bypasses UQL's inference for exact SQL control. | Raw SQL types: `'varchar(100)'`, `'decimal(10,2)'`, `'smallint'`, etc. |
 
 ```ts
@@ -590,6 +591,7 @@ Learn more about UQL at [uql-orm.dev](https://uql-orm.dev) for details on:
 
 - [Complex Logical Operators](https://uql-orm.dev/querying/logical-operators)
 - [Aggregate Queries (GROUP BY, HAVING, DISTINCT)](https://uql-orm.dev/querying/aggregate)
+- [Semantic Search (Vector Similarity)](https://uql-orm.dev/querying/semantic-search)
 - [Relationship Mapping (1-1, 1-M, M-M)](https://uql-orm.dev/querying/relations)
 - [Lifecycle Hooks](https://uql-orm.dev/entities/lifecycle-hooks)
 - [Soft Deletes &amp; Auditing](https://uql-orm.dev/entities/soft-delete)