app.py

import gradio as gr
import pandas as pd
import os

# 1) Load movies.csv from extracted ml-32m dataset
csv_path = os.path.join("movies.csv")  # Adjust path if needed
df = pd.read_csv(csv_path)

# 2) Normalize genres and extract release year from title
df["genres"] = df["genres"].apply(lambda g: "|".join(g) if isinstance(g, list) else str(g))
df["release_year"] = (
    df["title"]
      .str.extract(r"\((\d{4})\)")[0]
      .astype(pd.Int64Dtype(), errors='ignore')
)

# 3) Deduplicate metadata
metadata = df[["title", "genres", "release_year"]].drop_duplicates()

# 4) Extract unique genres
all_genres = set()
df["genres"].str.split("|").apply(all_genres.update)
genre_list = sorted(all_genres)

# 5) Recommendation function with year range and genre check
def recommend_by_genre_and_year_range(genre: str, start_year: int, end_year: int, top_k: int = 5) -> str:
    if not genre:
        return "⚠️ Please select a genre."

    mask_genre = metadata["genres"].str.lower().str.contains(genre.lower(), na=False)
    year_col = metadata["release_year"].fillna(0)
    mask_year = (year_col >= start_year) & (year_col <= end_year)

    candidates = metadata[mask_genre & mask_year]
    if candidates.empty:
        return f"No '{genre.title()}' movies found between {start_year} and {end_year}."

    picks = candidates.sample(n=min(top_k, len(candidates)))
    return "\n".join(
        f"• {row.title} ({int(row.release_year) if pd.notna(row.release_year) else 'Year N/A'})"
        for _, row in picks.iterrows()
    )

# 6) Gradio interface
iface = gr.Interface(
    fn=recommend_by_genre_and_year_range,
    inputs=[
        gr.Dropdown(choices=genre_list, label="Select Genre", value=None),
        gr.Number(label="Start Year", value=1990, precision=0),
        gr.Number(label="End Year", value=1995, precision=0),
        gr.Number(label="Number of Recommendations", value=5),
    ],
    outputs="text",
    title="🎬 Movie Recommender by Genre & Year Range",
    description="""
Loads local MovieLens metadata (ml-32m), extracts release years from titles,
normalizes genres, and filters by genre and a custom year range.
"""
)

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