Semantic Search Engine using FAISS

Demo

A simple Semantic Search Engine built with Python, Sentence Transformers, and FAISS.

The application reads a PDF document, converts the text into embeddings, stores them in a FAISS vector index, and retrieves the most semantically relevant chunks for a user's query.

---

Features

• Read PDF documents
• Extract text page by page
• Split text into sentences
• Create overlapping text chunks
• Generate embeddings using Sentence Transformers
• Store embeddings in FAISS
• Perform semantic similarity search
• Display page number and chunk ID for retrieved results

---

Technologies Used

• Python
• PyMuPDF
• Sentence Transformers
• FAISS
• NumPy
• PyTorch

---

Project Structure

Semantic_Search_Engine/
│
├── app.py
├── requirements.txt
├── README.md
├── .gitignore
└── data/
    └── ai_ml.pdf

---

Installation

Clone the repository:

git clone (https://github.com/ridazahra-ai/semantic-search-engine)

Install dependencies:

pip install -r requirements.txt

Run the application:

python app.py

---

Example

Ask a question:

What is Machine Learning?

Top Results

Page: 5
Chunk ID: 12

Machine Learning is a subset of Artificial Intelligence...

---

What I Learned

While building this project, I learned:

• PDF processing with PyMuPDF
• Text chunking strategies
• Sentence embeddings
• Vector similarity search
• FAISS indexing
• Metadata management
• Building a semantic search pipeline

---

Note

Place your own PDF inside the data/ folder.

Update the file path in app.py if necessary.

Author

Rida Zahra

AI Engineering Learning Journey