🧠 NLP Entity Extractor

A simple Named Entity Recognition (NER) application built using Spring Boot and Stanford CoreNLP, with a lightweight frontend for extracting entities like Person, City, Country, etc.

🚀 Features

• Extract named entities from text

• Supports multiple entity types:

  • 👤 Person
  • 🏙 City
  • 🌍 Country
  • 📍 State/Province
  • ✉️ Email
  • 🏷 Title

• Smart input preprocessing (capitalization handling)

• Clean and minimal frontend UI

• REST API-based architecture

🏗 Tech Stack

Backend

• Java
• Spring Boot
• Stanford CoreNLP

Frontend

• HTML, CSS, JavaScript (Vanilla)

📂 Project Structure

├── mvnw
├── mvnw.cmd
├── pom.xml
├── src/
│   ├── main/
│   │   ├── java/
│   │   │   └── com/
│   │   │       └── nlp/
│   │   │           └── NLP/
│   │   │               ├── NlpApplication.java
│   │   │               ├── controller/
│   │   │               │   └── NERController.java
│   │   │               ├── core/
│   │   │               │   └── Pipeline.java
│   │   │               └── model/
│   │   │                   └── Type.java
│   │   └── resources/
│   │       └── application.properties
│   └── test/
│       └── java/
│           └── com/
│               └── nlp/
│                   └── NLP/
│                       └── NlpApplicationTests.java
└── .mvn/
    └── wrapper/
        └── maven-wrapper.properties

⚙️ Setup Instructions

1️⃣ Clone the repository

git clone https://github.com/Himanshux19/NaturalLanguageProcessing.git
cd NLP

2️⃣ Run Backend

mvn clean install
mvn spring-boot:run

Backend will start at:

http://localhost:8080

3️⃣ Run Frontend

• Open index.html directly OR
• Use Live Server (recommended)

🔗 API Endpoint

POST /api/v1/ner

Query Parameter:

type = PERSON | CITY | COUNTRY | STATE_OR_PROVINCE | EMAIL | TITLE

Request Body:

Plain text input

Example:

curl -X POST "http://localhost:8080/api/v1/ner?type=PERSON" \
-H "Content-Type: text/plain" \
-d "Elon Musk lives in USA"

🧪 Example

Input:

elon musk met tim cook in usa

Processed:

Elon Musk Met Tim Cook In USA

Output:

Elon, Musk, Tim, Cook

🧠 How It Works

1. Input text is received via REST API

2. Text is preprocessed (capitalization normalization)

3. Stanford CoreNLP pipeline performs:

   • Tokenization
   • POS tagging
   • Named Entity Recognition

4. Entities are filtered based on selected type

5. Results are returned as a set

⚠️ Important Notes

• Backend expects:

Content-Type: text/plain

• CORS is enabled for frontend:

http://localhost:63342

• Proper capitalization improves NER accuracy

🔥 Future Improvements

• Highlight entities in UI
• Multi-entity detection (all types at once)
• Medical NLP integration (Apache cTAKES / BioBERT)
• True-casing using ML models
• Deployment (Render / AWS / Docker)

👨‍💻 Author

Himanshu Singh

⭐ Acknowledgements

• Stanford NLP Group
• Spring Boot Framework