app.py

import gradio as gr
from sentence_transformers import SentenceTransformer, util
import pandas as pd
import torch

# -- static dataset (10 items) --
movies = [
    {"title": "Movie_1","overview": "Space mission to explore distant galaxies."},
    {"title": "Movie_2","overview": "Romantic love story in college."},
    {"title": "Movie_3","overview": "Detective solves mysterious murder."},
    {"title": "Movie_4","overview": "Superheroes unite to save the world."},
    {"title": "Movie_5","overview": "Scientist travels through wormhole."},
    {"title": "Movie_6","overview": "Underdog team wins the championship."},
    {"title": "Movie_7","overview": "Horror story in a haunted house."},
    {"title": "Movie_8","overview": "Entrepreneur builds a startup."},
    {"title": "Movie_9","overview": "Historical war drama based on events."},
    {"title": "Movie_10","overview": "AI becomes self-aware and challenges humanity."}
]

df = pd.DataFrame(movies)
df["title_lower"] = df["title"].str.lower()

# -- load sentence transformer model --
MODEL_NAME = "sentence-transformers/all-MiniLM-L6-v2"
model = SentenceTransformer(MODEL_NAME)

# -- precompute embeddings --
movie_embeddings = model.encode(df["overview"].tolist(), convert_to_tensor=True)

# -- recommendation function --
def recommend_movie(movie_title):
    movie_title = movie_title.strip().lower()
    if movie_title not in df["title_lower"].values:
        return "Movie not found. Try Movie_1 to Movie_10."
    
    idx = df.index[df["title_lower"] == movie_title][0]
    query_emb = movie_embeddings[idx]

    # cosine similarity
    similarity = util.cos_sim(query_emb, movie_embeddings)[0]

    # get top 6 indices (including the movie itself)
    top_idxs = torch.topk(similarity, k=6).indices.tolist()

    # exclude the original movie
    recs = [df.iloc[i]["title"] for i in top_idxs if i != idx]

    return "\n".join(recs)

# -- Gradio interface --
interface = gr.Interface(
    fn=recommend_movie,
    inputs=gr.Textbox(label="Enter Movie title"),
    outputs="text",
    title="Movie Recommender",
    description="Enter a title (e.g., Movie_10) to get similar movies."
)

if __name__ == "__main__":
    interface.launch()