Create app.py

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+import pandas as pd
+from datasets import load_dataset
+
+# 1. Load MovieLens 100K from the Hub
+ml = load_dataset("movielens", "100k")            # train + test splits
+df = pd.concat([
+    ml["train"].to_pandas(), 
+    ml["test"].to_pandas()
+], ignore_index=True)
+
+# 2. Extract year and prepare genres
+df["year"] = pd.to_datetime(df["timestamp"], unit="s").dt.year
+# movieId → title/genres mapping is in the "movies" config
+movies = load_dataset("movielens", "100k", split="train") \
+            .to_pandas()[["movieId","title","genres"]].drop_duplicates()
+df = df.merge(movies, on="movieId", how="left")
+
+# 3. Deduplicate for metadata
+metadata = df[["title","genres","year"]].drop_duplicates()
+
+def recommend_by_genre_year(genre, year, top_k=5):
+    mask_genre = metadata["genres"].str.lower().str.contains(genre.lower())
+    mask_year  = metadata["year"] >= year
+    candidates = metadata[mask_genre & mask_year]
+    if candidates.empty:
+        return f"No {genre.title()} movies found from {year} onward."
+    picks = candidates.sample(min(top_k, len(candidates)))
+    return "\n".join(f"• {row.title} ({row.year})" for _, row in picks.iterrows())
+
+iface = gr.Interface(
+    fn=recommend_by_genre_year,
+    inputs=[
+        gr.Textbox(label="Genre", placeholder="e.g. Action, Romance"),
+        gr.Number(label="Release Year (≥)", value=2010),
+        gr.Slider(1, 10, step=1, label="Number of Recommendations", value=5)
+    ],
+    outputs="text",
+    title="🎬 Online MovieLens Recommender",
+    description="""
+Pulls MovieLens 100K live from Hugging Face Datasets—no local files needed.
+Filters by genre substring and release year (inferred from timestamp).
+"""
+)
+
+if __name__ == "__main__":
+    iface.launch()