Compare commits

...

6 Commits

Author SHA1 Message Date
0834ec5c91 5.9.2
Some checks failed
Default (tags) / security (push) Successful in 32s
Default (tags) / test (push) Successful in 3m3s
Default (tags) / release (push) Failing after 51s
Default (tags) / metadata (push) Successful in 57s
2025-04-18 15:10:04 +00:00
6a2a708ea1 fix(documentation): Update search API documentation to replace deprecated searchWithLucene examples with the unified search(query) API and clarify its behavior. 2025-04-18 15:10:03 +00:00
1d977986f1 5.9.1
Some checks failed
Default (tags) / security (push) Successful in 39s
Default (tags) / test (push) Successful in 3m2s
Default (tags) / release (push) Failing after 51s
Default (tags) / metadata (push) Successful in 58s
2025-04-18 14:56:11 +00:00
e325b42906 fix(search): Refactor search tests to use unified search API and update text index type casting 2025-04-18 14:56:11 +00:00
1a359d355a 5.9.0
Some checks failed
Default (tags) / security (push) Successful in 39s
Default (tags) / test (push) Failing after 6m57s
Default (tags) / release (push) Has been skipped
Default (tags) / metadata (push) Has been skipped
2025-04-18 11:25:39 +00:00
b5a9449d5e feat(collections/search): Improve text index creation and search fallback mechanisms in collections and document search methods 2025-04-18 11:25:39 +00:00
7 changed files with 126 additions and 93 deletions

View File

@ -1,5 +1,24 @@
# Changelog
## 2025-04-18 - 5.9.2 - fix(documentation)
Update search API documentation to replace deprecated searchWithLucene examples with the unified search(query) API and clarify its behavior.
- Replaced 'searchWithLucene' examples with 'search(query)' in the README.
- Updated explanation to detail field-specific exact match, partial word regex search, multi-word literal matching, and handling of empty queries.
- Clarified guidelines for creating MongoDB text indexes on searchable fields for optimized search performance.
## 2025-04-18 - 5.9.1 - fix(search)
Refactor search tests to use unified search API and update text index type casting
- Replaced all calls from searchWithLucene with search in test/search tests
- Updated text index specification in the collection class to use proper type casting
## 2025-04-18 - 5.9.0 - feat(collections/search)
Improve text index creation and search fallback mechanisms in collections and document search methods
- Auto-create a compound text index on all searchable fields in SmartdataCollection with a one-time flag to prevent duplicate index creation.
- Refine the search method in SmartDataDbDoc to support exact field matches and safe regex fallback for non-Lucene queries.
## 2025-04-17 - 5.8.4 - fix(core)
Update commit metadata with no functional code changes

View File

@ -1,6 +1,6 @@
{
"name": "@push.rocks/smartdata",
"version": "5.8.4",
"version": "5.9.2",
"private": false,
"description": "An advanced library for NoSQL data organization and manipulation using TypeScript with support for MongoDB, data validation, collections, and custom data types.",
"main": "dist_ts/index.js",

View File

@ -18,7 +18,7 @@ A powerful TypeScript-first MongoDB wrapper that provides advanced features for
- **Enhanced Cursors**: Chainable, type-safe cursor API with memory efficient data processing
- **Type Conversion**: Automatic handling of MongoDB types like ObjectId and Binary data
- **Serialization Hooks**: Custom serialization and deserialization of document properties
- **Powerful Search Capabilities**: Lucene-like query syntax with field-specific search, advanced operators, and fallback mechanisms
- **Powerful Search Capabilities**: Unified `search(query)` API supporting field:value exact matches, multi-field regex searches, case-insensitive matching, and automatic escaping to prevent regex injection
## Requirements
@ -218,43 +218,31 @@ class Product extends SmartDataDbDoc<Product, Product> {
const searchableFields = getSearchableFields('Product'); // ['name', 'description', 'category']
// Basic search across all searchable fields
const iPhoneProducts = await Product.searchWithLucene('iPhone');
const iphoneProducts = await Product.search('iPhone');
// Field-specific search
const electronicsProducts = await Product.searchWithLucene('category:Electronics');
// Field-specific exact match
const electronicsProducts = await Product.search('category:Electronics');
// Search with wildcards
const macProducts = await Product.searchWithLucene('Mac*');
// Partial word search (regex across all fields)
const laptopResults = await Product.search('laptop');
// Search in specific fields with partial words
const laptopResults = await Product.searchWithLucene('description:laptop');
// Multi-word literal search
const paperwhite = await Product.search('Kindle Paperwhite');
// Search is case-insensitive
const results1 = await Product.searchWithLucene('electronics');
const results2 = await Product.searchWithLucene('Electronics');
// results1 and results2 will contain the same documents
// Using boolean operators (requires text index in MongoDB)
const wirelessOrLaptop = await Product.searchWithLucene('wireless OR laptop');
// Negative searches
const electronicsNotSamsung = await Product.searchWithLucene('Electronics NOT Samsung');
// Phrase searches
const exactPhrase = await Product.searchWithLucene('"high-speed blender"');
// Grouping with parentheses
const complexQuery = await Product.searchWithLucene('(wireless OR bluetooth) AND Electronics');
// Empty query returns all documents
const allProducts = await Product.search('');
```
The search functionality includes:
- `@searchable()` decorator for marking fields as searchable
- `getSearchableFields()` to retrieve all searchable fields for a class
- `search()` method for basic search (requires MongoDB text index)
- `searchWithLucene()` method with robust fallback mechanisms
- Support for field-specific searches, wildcards, and boolean operators
- Automatic fallback to regex-based search if MongoDB text search fails
- `getSearchableFields()` to list searchable fields for a model
- `search(query: string)` method supporting:
- Field-specific exact matches (`field:value`)
- Case-insensitive partial matches across all searchable fields
- Multi-word literal matching
- Empty queries returning all documents
- Automatic escaping of special characters to prevent regex injection
### EasyStore
@ -549,10 +537,10 @@ class Order extends SmartDataDbDoc<Order, Order> {
### Search Optimization
- Create MongoDB text indexes for collections that need advanced search operations
- Use `searchWithLucene()` for robust searches with fallback mechanisms
- Prefer field-specific searches when possible for better performance
- Use simple term queries instead of boolean operators if you don't have text indexes
- (Optional) Create MongoDB text indexes on searchable fields to speed up full-text search
- Use `search(query)` for all search operations (field:value, partial matches, multi-word)
- Prefer field-specific exact matches when possible for optimal performance
- Avoid unnecessary complexity in query strings to keep regex searches efficient
### Performance Optimization

View File

@ -104,21 +104,21 @@ tap.test('should search products by basic search method', async () => {
}
});
tap.test('should search products with searchWithLucene method', async () => {
tap.test('should search products with search method', async () => {
// Using the robust searchWithLucene method
const wirelessResults = await Product.searchWithLucene('wireless');
console.log(
`Found ${wirelessResults.length} products matching 'wireless' using searchWithLucene`,
);
const wirelessResults = await Product.search('wireless');
console.log(
`Found ${wirelessResults.length} products matching 'wireless' using search`,
);
expect(wirelessResults.length).toEqual(1);
expect(wirelessResults[0].name).toEqual('AirPods');
});
tap.test('should search products by category with searchWithLucene', async () => {
tap.test('should search products by category with search', async () => {
// Using field-specific search with searchWithLucene
const kitchenResults = await Product.searchWithLucene('category:Kitchen');
console.log(`Found ${kitchenResults.length} products in Kitchen category using searchWithLucene`);
const kitchenResults = await Product.search('category:Kitchen');
console.log(`Found ${kitchenResults.length} products in Kitchen category using search`);
expect(kitchenResults.length).toEqual(2);
expect(kitchenResults[0].category).toEqual('Kitchen');
@ -127,7 +127,7 @@ tap.test('should search products by category with searchWithLucene', async () =>
tap.test('should search products with partial word matches', async () => {
// Testing partial word matches
const proResults = await Product.searchWithLucene('Pro');
const proResults = await Product.search('Pro');
console.log(`Found ${proResults.length} products matching 'Pro'`);
// Should match both "MacBook Pro" and "professionals" in description
@ -136,7 +136,7 @@ tap.test('should search products with partial word matches', async () => {
tap.test('should search across multiple searchable fields', async () => {
// Test searching across all searchable fields
const bookResults = await Product.searchWithLucene('book');
const bookResults = await Product.search('book');
console.log(`Found ${bookResults.length} products matching 'book' across all fields`);
// Should match "MacBook" in name and "Books" in category
@ -145,8 +145,8 @@ tap.test('should search across multiple searchable fields', async () => {
tap.test('should handle case insensitive searches', async () => {
// Test case insensitivity
const electronicsResults = await Product.searchWithLucene('electronics');
const ElectronicsResults = await Product.searchWithLucene('Electronics');
const electronicsResults = await Product.search('electronics');
const ElectronicsResults = await Product.search('Electronics');
console.log(`Found ${electronicsResults.length} products matching lowercase 'electronics'`);
console.log(`Found ${ElectronicsResults.length} products matching capitalized 'Electronics'`);
@ -166,14 +166,14 @@ tap.test('should demonstrate search fallback mechanisms', async () => {
// Use a simpler term that should be found in descriptions
// Avoid using "OR" operator which requires a text index
const results = await Product.searchWithLucene('high');
const results = await Product.search('high');
console.log(`Found ${results.length} products matching 'high'`);
// "High-speed blender" contains "high"
expect(results.length).toBeGreaterThan(0);
// Try another fallback example that won't need $text
const powerfulResults = await Product.searchWithLucene('powerful');
const powerfulResults = await Product.search('powerful');
console.log(`Found ${powerfulResults.length} products matching 'powerful'`);
// "Powerful laptop for professionals" contains "powerful"
@ -192,6 +192,35 @@ tap.test('should explain the advantages of the integrated approach', async () =>
expect(true).toEqual(true);
});
// Additional robustness tests
tap.test('should search exact name using field:value', async () => {
const nameResults = await Product.search('name:AirPods');
expect(nameResults.length).toEqual(1);
expect(nameResults[0].name).toEqual('AirPods');
});
tap.test('should throw when searching non-searchable field', async () => {
let error: Error;
try {
await Product.search('price:129');
} catch (err) {
error = err as Error;
}
expect(error).toBeTruthy();
expect(error.message).toMatch(/not searchable/);
});
tap.test('empty query should return all products', async () => {
const allResults = await Product.search('');
expect(allResults.length).toEqual(8);
});
tap.test('should search multi-word term across fields', async () => {
const termResults = await Product.search('iPhone 12');
expect(termResults.length).toEqual(1);
expect(termResults[0].name).toEqual('iPhone 12');
});
tap.test('close database connection', async () => {
await testDb.mongoDb.dropDatabase();
await testDb.close();

View File

@ -3,6 +3,6 @@
*/
export const commitinfo = {
name: '@push.rocks/smartdata',
version: '5.8.4',
version: '5.9.2',
description: 'An advanced library for NoSQL data organization and manipulation using TypeScript with support for MongoDB, data validation, collections, and custom data types.'
}

View File

@ -1,7 +1,7 @@
import * as plugins from './plugins.js';
import { SmartdataDb } from './classes.db.js';
import { SmartdataDbCursor } from './classes.cursor.js';
import { SmartDataDbDoc, type IIndexOptions } from './classes.doc.js';
import { SmartDataDbDoc, type IIndexOptions, getSearchableFields } from './classes.doc.js';
import { SmartdataDbWatcher } from './classes.watcher.js';
import { CollectionFactory } from './classes.collectionfactory.js';
@ -128,6 +128,8 @@ export class SmartdataCollection<T> {
public smartdataDb: SmartdataDb;
public uniqueIndexes: string[] = [];
public regularIndexes: Array<{field: string, options: IIndexOptions}> = [];
// flag to ensure text index is created only once
private textIndexCreated: boolean = false;
constructor(classNameArg: string, smartDataDbArg: SmartdataDb) {
// tell the collection where it belongs
@ -153,6 +155,16 @@ export class SmartdataCollection<T> {
console.log(`Successfully initiated Collection ${this.collectionName}`);
}
this.mongoDbCollection = this.smartdataDb.mongoDb.collection(this.collectionName);
// Auto-create a compound text index on all searchable fields
const searchableFields = getSearchableFields(this.collectionName);
if (searchableFields.length > 0 && !this.textIndexCreated) {
// Build a compound text index spec
const indexSpec: Record<string, 'text'> = {};
searchableFields.forEach(f => { indexSpec[f] = 'text'; });
// Cast to any to satisfy TypeScript IndexSpecification typing
await this.mongoDbCollection.createIndex(indexSpec as any, { name: 'smartdata_text_index' });
this.textIndexCreated = true;
}
}
}

View File

@ -61,6 +61,10 @@ export function getSearchableFields(className: string): string[] {
}
return Array.from(searchableFieldsMap.get(className));
}
// Escape user input for safe use in MongoDB regular expressions
function escapeForRegex(input: string): string {
return input.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
}
/**
* unique index - decorator to mark a unique index
@ -326,57 +330,38 @@ export class SmartDataDbDoc<T extends TImplements, TImplements, TManager extends
}
/**
* Search documents using Lucene query syntax
* @param luceneQuery Lucene query string
* Search documents by text or field:value syntax, with safe regex fallback
* @param query A search term or field:value expression
* @returns Array of matching documents
*/
public static async search<T>(
this: plugins.tsclass.typeFest.Class<T>,
luceneQuery: string,
query: string,
): Promise<T[]> {
const filter = (this as any).createSearchFilter(luceneQuery);
return await (this as any).getInstances(filter);
}
/**
* Search documents using Lucene query syntax with robust error handling
* @param luceneQuery The Lucene query string to search with
* @returns Array of matching documents
*/
public static async searchWithLucene<T>(
this: plugins.tsclass.typeFest.Class<T>,
luceneQuery: string,
): Promise<T[]> {
try {
const className = (this as any).className || this.name;
const searchableFields = getSearchableFields(className);
if (searchableFields.length === 0) {
console.warn(
`No searchable fields defined for class ${className}, falling back to simple search`,
);
return (this as any).searchByTextAcrossFields(luceneQuery);
}
// Simple term search optimization
if (
!luceneQuery.includes(':') &&
!luceneQuery.includes(' AND ') &&
!luceneQuery.includes(' OR ') &&
!luceneQuery.includes(' NOT ')
) {
return (this as any).searchByTextAcrossFields(luceneQuery);
}
// Try to use the Lucene-to-MongoDB conversion
const filter = (this as any).createSearchFilter(luceneQuery);
return await (this as any).getInstances(filter);
} catch (error) {
console.error(`Error in searchWithLucene: ${error.message}`);
return (this as any).searchByTextAcrossFields(luceneQuery);
const className = (this as any).className || this.name;
const searchableFields = getSearchableFields(className);
if (searchableFields.length === 0) {
throw new Error(`No searchable fields defined for class ${className}`);
}
// field:value exact match (case-sensitive for non-regex fields)
const fv = query.match(/^(\w+):(.+)$/);
if (fv) {
const field = fv[1];
const value = fv[2];
if (!searchableFields.includes(field)) {
throw new Error(`Field '${field}' is not searchable for class ${className}`);
}
return await (this as any).getInstances({ [field]: value });
}
// safe regex across all searchable fields (case-insensitive)
const escaped = escapeForRegex(query);
const orConditions = searchableFields.map((field) => ({
[field]: { $regex: escaped, $options: 'i' },
}));
return await (this as any).getInstances({ $or: orConditions });
}
/**
* Search by text across all searchable fields (fallback method)
* @param searchText The text to search for in all searchable fields