Create app.py

app.py

@@ -0,0 +1,49 @@
+import gradio as gr
+import pandas as pd
+import pickle
+from sentence_transformers import SentenceTransformer, util
+
+mdl_name = 'sentence-transformers/all-distilroberta-v1'
+model = SentenceTransformer(mdl_name)
+
+embedding_cache_path = ""
+with open(embedding_cache_path, "rb") as fIn:
+    cache_data = pickle.load(fIn)
+
+
+def user_query_recommend(query, min_p, max_p, embedding_table = cache_data["embeddings"], reviews = cache_data["data"]):
+    # Embed user query
+    embedding = model.encode(query)
+
+    # Calculate similarity with all reviews
+    sim_scores = util.cos_sim(embedding, embedding_table)
+    #print(sim_scores.shape)
+
+    # Recommend
+    recommendations = reviews.copy()
+    recommendations['price'] =recommendations.price.apply(lambda x: re.findall("\d+", x.replace(",","").replace(".00","").replace("$",""))[0]).astype('int')
+    recommendations['sim'] = sim_scores.T
+    recommendations = recommendations.sort_values('sim', ascending=False)
+    recommendations = recommendations.loc[(recommendations.price >= min_p) &
+                                          (recommendations.price <= max_p),
+                                          ['name', 'category', 'price', 'description', 'sim']]
+
+    return recommendations
+    
+interface = gr.Interface(
+    user_query_recommend, 
+    inputs=[gr.inputs.Textbox(),
+            gr.inputs.Slider(minimum=1, maximum=100, default=30, label='Min Price'),
+            gr.inputs.Slider(minimum=1, maximum=1000, default=70, label='Max Price')],
+    outputs=[
+        gr.outputs.Textbox(label="Recommendations"),
+    ],
+    title = "Scotch Recommendation",
+    examples=["very sweet with lemons and oranges and marmalades", "smoky peaty earthy and spicy"],
+    theme="huggingface",
+)
+
+interface.launch(
+    enable_queue=True,
+    cache_examples=True,
+)