🔍 Implementing Search in APIs

🧐 Introduction

Search functionality is essential in modern APIs, enabling users to efficiently locate and retrieve relevant data.
An effective search system: ✅ Enhances user experience – Users can find data quickly.
✅ Improves data accessibility – Allows seamless navigation of large datasets.
✅ Optimizes performance – Reduces server load with efficient querying.

In this lesson, we will:

• Understand different types of search techniques.
• Learn about fuzzy search and how it handles typos.
• Explore how Elasticsearch works for advanced searching.
• Implement a search feature in an API.

---

📌 1. Types of Search Techniques 🧩

APIs implement search in different ways, depending on data structure, query complexity, and performance requirements.

✅ 1️⃣ Simple Query-Based Search

🔹 Searches performed using query parameters (GET /items?q=laptop).
🔹 Works well for structured data and small datasets.
🔹 Can be implemented using a database WHERE clause.

Example (SQL Query):

SELECT * FROM products WHERE name LIKE '%laptop%';

---

✅ 2️⃣ Full-Text Search

🔹 Performs comprehensive text searches across large text fields.
🔹 Supports partial matches, phrase searches, and logical operations.
🔹 Typically implemented using Elasticsearch, Solr, or PostgreSQL full-text search.

Example (PostgreSQL Full-Text Search):

SELECT * FROM products WHERE to_tsvector(name) @@ to_tsquery('gaming & laptop');

---

✅ 3️⃣ Fuzzy Search (Handling Typos)

🔹 Allows approximate matching, useful for handling: ✔ Typos
✔ Misspellings
✔ Variations in user input

🔹 Uses Levenshtein distance (edit distance) to calculate similarity between words.

Example: Searching for "exmple" should return "example".

Fuzzy Search in SQL (PostgreSQL - pg_trgm extension):

SELECT * FROM products WHERE similarity(name, 'laptop') > 0.5 ORDER BY similarity(name, 'laptop') DESC;

---

📌 2. Deep Dive into Fuzzy Search 📌

How Does Fuzzy Search Work?

Fuzzy search is based on Levenshtein distance, which calculates how many character changes are needed to match two words.

Word 1	Word 2	Edit Distance
example	exmple	1 (missing "a")
laptop	lapotp	1 (swapped "t" and "o")
gaming	gamong	1 (typo in "n")

Example using Elasticsearch Fuzzy Search:

{
  "query": {
    "fuzzy": {
      "name": {
        "value": "exmple",
        "fuzziness": "AUTO"
      }
    }
  }
}

✔ Fuzziness Levels:

• "AUTO": Adjusts allowed typos based on query length.
• 1: Allows 1-character typo.
• 2: Allows 2-character typos.

---

📌 3. Understanding Elasticsearch 📚

What is Elasticsearch?

Elasticsearch is a distributed search engine designed for fast and scalable full-text search.

🔹 How Elasticsearch Works:

1️⃣ Indexing → Converts documents into an efficient searchable format.
2️⃣ Analyzers → Breaks text into tokens for better matching.
3️⃣ Queries → Allows complex searches (match, fuzzy, wildcard, etc.).
4️⃣ Scoring → Ranks results based on relevance to the search query.

---

📌 4. Implementing Search with Elasticsearch in an API

✅ Step 1: Install Elasticsearch Client

Run:

npm install @elastic/elasticsearch express

---

✅ Step 2: Set Up Elasticsearch Client in Node.js

const { Client } = require('@elastic/elasticsearch');
const client = new Client({ node: 'http://localhost:9200' });

async function createIndex() {
  await client.indices.create({
    index: 'products',
    body: {
      mappings: {
        properties: {
          name: { type: 'text' },
          description: { type: 'text' }
        }
      }
    }
  });
}

createIndex().catch(console.error);

✔ Creates an Elasticsearch index for storing product data.

---

✅ Step 3: Index Sample Data in Elasticsearch

async function addProducts() {
  await client.index({
    index: 'products',
    body: {
      name: 'Gaming Laptop',
      description: 'A powerful laptop for gaming and work'
    }
  });
}

addProducts().catch(console.error);

✔ Adds a product document to Elasticsearch.

---

✅ Step 4: Implement Search Route in Express.js

const express = require('express');
const app = express();

app.get('/search', async (req, res) => {
  const { q } = req.query;

  try {
    const { body } = await client.search({
      index: 'products',
      body: {
        query: {
          multi_match: {
            query: q,
            fields: ['name', 'description'],
            fuzziness: "AUTO"
          }
        }
      }
    });

    res.json(body.hits.hits);
  } catch (error) {
    res.status(500).json({ error: error.message });
  }
});

app.listen(3033, () => console.log('Server running on port 3033'));

✔ Searches for products using fuzzy matching and multi-field search.

---

📌 5. Other Search Mechanisms 🔧

✅ Prefix Search

• Finds words starting with a specific prefix.

{
  "query": {
    "prefix": { "name": "lap" }
  }
}

✅ Wildcard Search

• Matches words with pattern-based queries.

{
  "query": {
    "wildcard": { "name": "lap*op" }
  }
}

✅ Geospatial Search

• Used for location-based searches (e.g., find nearby stores).

{
  "query": {
    "geo_distance": {
      "distance": "10km",
      "location": { "lat": 40.7128, "lon": -74.0060 }
    }
  }
}

---

📌 6. Best Practices for Search Performance 🌟

✅ 1️⃣ Proper Indexing

• Use correct field types (text, keyword, geo_point).
• Preprocess data to remove stop words and normalize text.

✅ 2️⃣ Optimized Queries

• Use filters instead of queries for faster lookups.
• Combine full-text search with keyword filters.

✅ 3️⃣ Security Measures

• Sanitize user input to prevent query injection.
• Implement rate limiting for API search endpoints.

✅ 4️⃣ Pagination & Sorting

• Use from & size in Elasticsearch to paginate results.

{
  "from": 0,
  "size": 10,
  "query": { "match": { "name": "laptop" } }
}

---

🎯 Summary

✅ Implemented different search techniques (simple, full-text, fuzzy).
✅ Learned how Elasticsearch indexes and retrieves data efficiently.
✅ Built an API with search capabilities using Express.js & Elasticsearch.
✅ Explored advanced search mechanisms (wildcard, prefix, geospatial).