Update app.py

app.py

@@ -1,95 +1,66 @@
-import gradio as gr
 import pandas as pd
-import re
-from huggingface_hub import InferenceClient
+import gradio as gr
 
-# Load inventory dataset
+# Load inventory
 df = pd.read_csv("inventory.csv")
 
-# Hugging Face LLM client
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+# Extract category list
+categories = sorted(df['Category'].dropna().unique())
 
-# 🧞 RetailGenie logic
-def retailgenie_query(user_input: str) -> str:
-    user_input = user_input.lower()
-
-    if "where can i find" in user_input:
-        product = user_input.split("find")[-1].strip()
-        match = df[df['Product Name'].str.lower().str.contains(product)]
-        if not match.empty:
-            return f"{product.title()} is in {match.iloc[0]['Aisle']}."
-        else:
-            return "Sorry, I couldn't find that product."
+# Step 1: Ask for category
+def ask_category():
+    return "Please select a product category:", gr.Dropdown(choices=categories, label="Product Category"), None
 
-    elif "in stock" in user_input:
-        product = re.sub(r"(is|in stock|\?)", "", user_input).strip()
-        match = df[df['Product Name'].str.lower().str.contains(product)]
-        if not match.empty:
-            return f"{product.title()} is {'available' if match.iloc[0]['In Stock'].lower() == 'yes' else 'out of stock'}."
-        else:
-            return "Product not found."
-
-    elif "suggest" in user_input and "under" in user_input:
-        category = user_input.split("suggest a")[-1].split("under")[0].strip()
-        try:
-            price = int(re.findall(r"\d+", user_input)[-1])
-        except:
-            return "Please enter a valid price."
-        matches = df[(df['Category'].str.lower().str.contains(category)) & (df['Price (rupees)'] <= price)]
-        if not matches.empty:
-            return "Here are some options: " + ", ".join(matches['Product Name'].values[:3])
-        else:
-            return "No items found in that range."
+# Step 2: Based on category, ask for product
+def ask_product(category):
+    products = df[df['Category'] == category]['Product Name'].unique().tolist()
+    return (
+        f"Now select a product from {category}:",
+        gr.Dropdown(choices=products, label="Product Name"),
+        category
+    )
 
-    elif "deal" in user_input or "deals" in user_input:
-        electronics = df[df['Category'].str.lower() == "electronics"]
-        deals = electronics[electronics['Price (rupees)'] < 5000]
-        if not deals.empty:
-            return "Deals: " + ", ".join(deals['Product Name'].values[:3])
+# Step 3: Show product info
+def show_result(product, category):
+    match = df[(df['Category'] == category) & (df['Product Name'] == product)]
+    
+    if not match.empty:
+        row = match.iloc[0]
+        if row["In Stock"].strip().lower() == "yes":
+            return f"✅ Your **{product}** is in **{row['Aisle']}**. Quantity available: **{row['Quantity']}**, Price: ₹{row['Price (rupees)']}."
         else:
-            return "No current deals found in electronics."
+            return f"❌ Sorry, **{product}** is currently **not available** in stock."
+    else:
+        return "Product not found."
 
-    return None  # Not a RetailGenie-style query
+# Gradio interface with state
+with gr.Blocks(title="🧞 RetailGenie Guided Assistant") as demo:
+    gr.Markdown("## 🛍️ RetailGenie – Smart In-Store Assistant")
+    gr.Markdown("Start by selecting a product category:")
 
-# 🔁 Unified response function
-def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
-    # First try product-aware logic
-    rg_reply = retailgenie_query(message)
-    if rg_reply:
-        yield rg_reply
-        return
+    state_category = gr.State()
+    
+    # Step 1
+    cat_btn = gr.Button("Select Category")
+    cat_out = gr.Textbox(label="Message")
+    cat_dropdown = gr.Dropdown(choices=categories, visible=False)
+    
+    # Step 2
+    prod_out = gr.Textbox(label="Message")
+    prod_dropdown = gr.Dropdown(visible=False)
+    
+    # Step 3
+    final_out = gr.Textbox(label="Result")
 
-    # Else use LLM
-    messages = [{"role": "system", "content": system_message}]
-    for user, assistant in history:
-        if user: messages.append({"role": "user", "content": user})
-        if assistant: messages.append({"role": "assistant", "content": assistant})
-    messages.append({"role": "user", "content": message})
+    # Step 1 callback
+    def show_cat_options():
+        return "Please choose a product category:", gr.update(visible=True), None
+    cat_btn.click(show_cat_options, outputs=[cat_out, cat_dropdown, state_category])
 
-    response = ""
-    for msg in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = msg.choices[0].delta.content
-        response += token
-        yield response
+    # Step 2 callback
+    cat_dropdown.change(ask_product, inputs=cat_dropdown, outputs=[prod_out, prod_dropdown, state_category])
 
-# 🧠 Gradio Chat UI
-demo = gr.ChatInterface(
-    respond,
-    title="🧞 RetailGenie",
-    description="Ask me where to find products, check availability, get price-based suggestions, or talk to an AI.",
-    additional_inputs=[
-        gr.Textbox(value="You are a helpful retail assistant.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
-    ]
-)
+    # Step 3 callback
+    prod_dropdown.change(show_result, inputs=[prod_dropdown, state_category], outputs=final_out)
 
-if __name__ == "__main__":
-    demo.launch()
+demo.launch()