🎬 MovieSearch

A Smart Movie Recommendation App Without Machine Learning

MovieSearch is a full-stack web application that recommends the top 5 movies similar to the one entered by the user using a simple yet effective Bag-of-Words technique. Built without any machine learning library, it uses intelligent text similarity matching and leverages real-time data from TheMovieDB API for movie metadata.

🌐 Live App: MovieSearch

---

🍿 Features

• 🔍 Search for any movie by typing its name
• ✏️ Auto-completion for movie titles using fuzzy matching
• 🎞️ Movie results include:
  • Title
  • Description (via TheMovieDB API)
  • Poster image (via TheMovieDB API)
• 🌄 Dynamic background that changes based on:
  • 🌙 Time of Day (Morning, Afternoon, Evening, Night)
  • 📍 User Location (via browser)

---

🧱 Tech Stack

Frontend	Backend	Recommendation Logic	External API	UI Feature
HTML, CSS, JS	Python (Flask)	Bag-of-Words (TextSim)	TheMovieDB	Time- Based BG
	Jinja2	Fuzzy Matching	IP-based Geolocation

---

🧠 Colab Notebook

Bag of words technique was used to acheive the results.

📓 Google Colab Notebook

---

🌿 System Overview

---

🗂️ Project Structure

MovieSearch/
├── static/                   # Static assets (CSS, JS, Backgrounds)
│   └── image/                # Time-based folders (morning, afternoon, evening, night)
│   └── autocomplete.js       # Responsible for auto-completion
│   └── background_setter.js  # Responsible for dynamic background setting
│   └── index_style.css       # Home page styling
│   └── result_style.css      # Result page styling
├── templates/                # HTML templates (Home, Result)
├── app.py                    # Flask backend logic
├── recommender.py            # Responsible for recommendation logic and providing movies detailes (Name, icon, info)
├── movie_names.csv           # Contains all movies names
├── Finalized_file.csv        # Contains the processed file from colab
├── similarity.npy            # Contains similarity value [0-1] of every movie with every other movie
├── requirements.txt          # Python dependencies
└── README.md                 # You're here!

---

⚙️ How It Works

🧠 Movie Similarity (No ML)

1. User enters a movie title.
2. Auto-correction fixes spelling errors using fuzzy logic.
3. Backend processes request using a Bag-of-Words technique.
4. Top 5 most similar movies are selected from dataset.
5. Metadata (overview + poster) is fetched from TMDB API.
6. Results are shown with dynamic background.

🌇 Background Image Logic

1. Browser detects user’s IP address and local time.
2. Script checks current time:
   • 🌅 Morning (6–12)
   • ☀️ Afternoon (12–17)
   • 🌆 Evening (17–20)
   • 🌃 Night (20–6)
3. Background image is picked from the corresponding folder.

---

🛠️ Installation & Setup

📦 Prerequisites

• Python 3.9+
• Flask
• An API key from TheMovieDB

---

🔌 Clone the Repo

git clone https://github.com/your-username/MovieSearch.git
cd MovieSearch

---

⚙️ Setup Instructions

1. Configure Backend

pip install -r requirements.txt

2. Set API key

Create a .env file:

TMDB_API_KEY=your_tmdb_api_key

3. Run the App

python app.py

🧠 Learnings & Highlights

• Designed a non-ML movie recommendation system using Bag-of-Words
• Implemented real-time spell correction with fuzzy matching
• Used TheMovieDB API to enrich results with live data
• Built a dynamic day/night themed UI based on local time and location
• Developed a modular, scalable Flask application structure

---

📈 Future Enhancements

• 🔐 Add user login and personalized recommendations
• 🗑️ Clear history or save favorite movies
• 🔍 Search by genre or actor
• 🌐 Improve location detection using browser geolocation

---

🤝 Contributing

Pull requests are welcome! For major changes, please open an issue first to discuss what you would like to change.

---

📄 License

MIT Licence

---

👨‍💻 Author

Aryan Tiwari
📫 LinkedIn
💻 GitHub

---