import pandas as pd
import numpy as np
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import gradio as gr

# Load dataset
dataset = pd.read_csv("https://huggingface.co/datasets/Pabl1nho/movies-dataset/resolve/main/9000plus.csv")
dataset['full_text'] = dataset['Title'] + " : " + dataset['Overview']

# Encode dataset
model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
embeddings = model.encode(dataset['full_text'], convert_to_tensor=True, show_progress_bar=True)
dataset['embeddings'] = embeddings.cpu().numpy().tolist()

# Define function
def recommend_movies(description):
    input_emb = model.encode([description])
    similarities = cosine_similarity(input_emb, np.array(dataset['embeddings']))[0]
    top_indices = similarities.argsort()[::-1][:5]
    results = []
    for idx in top_indices:
        title = dataset.iloc[idx]['Title']
        overview = dataset.iloc[idx]['Overview']
        results.append(f"**{title}**: {overview}")
    return "\n\n".join(results)

# Gradio interface
iface = gr.Interface(
    fn=recommend_movies,
    inputs=gr.Textbox(lines=2, placeholder="Describe the movie you're looking for..."),
    outputs="markdown",
    title="🎬 Movie Recommendation App",
    description="Get 5 movies similar to your description based on semantic search!"
)

iface.launch()