Update app.py

app.py

@@ -1,22 +1,17 @@
 import gradio as gr
 import pandas as pd
-from transformers import AutoTokenizer, AutoModelForCausalLM
-from sentence_transformers import SentenceTransformer
 import faiss
 import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from sentence_transformers import SentenceTransformer
 
 # -------------------------------
 # Load dataset
 # -------------------------------
-file_path = "marketing-campaigns.csv"
+file_path = "marketing-campaigns.csv"  # file uploaded in Hugging Face space
 df = pd.read_csv(file_path)
-
-if df.empty:
-    raise ValueError("CSV is empty. Please provide a dataset with campaign info.")
-
-# Join all columns to form knowledge text
-df = df.dropna()
-df["text"] = df.astype(str).agg(" | ".join, axis=1)
+df = df.dropna()  # drop completely empty rows
+df["text"] = df.astype(str).agg(" | ".join, axis=1)  # merge all cols into text
 
 # -------------------------------
 # Embeddings + FAISS
@@ -30,91 +25,155 @@ index = faiss.IndexFlatL2(d)
 index.add(embeddings_np)
 
 # -------------------------------
-# Load LLM (Phi-4-mini)
+# Load LLM (default Phi-4-mini)
 # -------------------------------
-model_name = "microsoft/phi-4-mini"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float32, device_map="auto")
+model_choices = {
+    "openai/gpt-oss-120b": "microsoft/phi-4-mini",
+    "openai/gpt-oss-20b": "microsoft/phi-4-mini"  # placeholder, can map to another HF model
+}
+
+current_model_id = "openai/gpt-oss-120b"
+hf_model = model_choices[current_model_id]
+
+tokenizer = AutoTokenizer.from_pretrained(hf_model)
+model = AutoModelForCausalLM.from_pretrained(
+    hf_model, torch_dtype=torch.float32, device_map="auto"
+)
 
 # -------------------------------
-# RAG functions
+# RAG Functions
 # -------------------------------
 def retrieve_context(query, k=3):
-    """Retrieve top-k similar rows from dataset"""
-    if not query.strip():
-        return []
     query_vec = embed_model.encode([query], convert_to_tensor=True).cpu().numpy()
-    D, I = index.search(query_vec, min(k, len(df)))
+    D, I = index.search(query_vec, k)
     results = [df.iloc[i]["text"] for i in I[0]]
     return results
 
-def generate_with_rag(prompt, k=3, temperature=0.7):
-    if not prompt.strip():
-        return "⚠️ Please enter a campaign idea or theme."
-    
-    # Step 1: Retrieve supporting facts
-    context = retrieve_context(prompt, k)
-    if not context:
-        return "⚠️ No relevant context found in dataset."
-    context_str = "\n".join(context[:k])
-
-    # Step 2: Build grounded structured prompt
+def generate_with_rag(prompt, temperature=0.8, max_tokens=250):
+    # Step 1: Retrieve context
+    context = retrieve_context(prompt, k=3)
+    context_str = "\n".join(context)
+
+    # Step 2: Construct grounded prompt
     rag_prompt = f"""
-    You are a top-tier creative marketing AI assistant.
-    Use the following supporting dataset entries as context:
+    You are a creative AI campaign assistant.
+    Use the following supporting data to ground your answer:
     {context_str}
 
-    Task: Generate a **structured marketing campaign** for:
+    Task: Generate a unique and creative marketing campaign idea for:
     {prompt}
-
-    Format your answer clearly with:
-    - 📌 Campaign Title
-    - ✨ Tagline
-    - 🧑‍🤝‍🧑 Target Audience
-    - 🎯 Key Selling Points
-    - 🎬 Creative Idea
     """
 
-    # Step 3: LLM Generation
+    # Step 3: Generate
     inputs = tokenizer(rag_prompt, return_tensors="pt").to(model.device)
     outputs = model.generate(
         **inputs,
-        max_length=300,
-        temperature=float(temperature),
-        top_p=0.9,
-        do_sample=True
+        max_length=max_tokens,
+        temperature=temperature,
+        top_p=0.9
     )
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-def search_dataset(query, k=5):
-    """Search dataset and return top matching rows"""
-    context = retrieve_context(query, k)
-    if not context:
-        return "⚠️ No results found."
-    return "\n\n".join(context)
+def switch_model(model_choice):
+    """Switch between available models."""
+    global model, tokenizer, current_model_id
+    hf_model = model_choices.get(model_choice, "microsoft/phi-4-mini")
+    tokenizer = AutoTokenizer.from_pretrained(hf_model)
+    model = AutoModelForCausalLM.from_pretrained(
+        hf_model, torch_dtype=torch.float32, device_map="auto"
+    )
+    current_model_id = model_choice
+    return gr.update(visible=(model_choice=="openai/gpt-oss-120b")), gr.update(visible=(model_choice=="openai/gpt-oss-20b")), model_choice
+
+# -------------------------------
+# Custom CSS
+# -------------------------------
+custom_css = """
+.gradio-container {
+    background: linear-gradient(135deg, #667eea 0%, #764ba2 25%, #f093fb 50%, #4facfe 75%, #00f2fe 100%);
+    background-size: 400% 400%;
+    animation: gradient-animation 15s ease infinite;
+    min-height: 100vh;
+}
+@keyframes gradient-animation {
+    0% { background-position: 0% 50%; }
+    50% { background-position: 100% 50%; }
+    100% { background-position: 0% 50%; }
+}
+.main-container {
+    background-color: rgba(255, 255, 255, 0.95);
+    backdrop-filter: blur(10px);
+    border-radius: 20px;
+    padding: 20px;
+    box-shadow: 0 8px 32px 0 rgba(31, 38, 135, 0.37);
+    border: 1px solid rgba(255, 255, 255, 0.18);
+    margin: 10px;
+}
+"""
 
 # -------------------------------
 # Gradio UI
 # -------------------------------
-with gr.Blocks() as demo:
-    gr.Markdown("# 🤖 RAG-powered Creative Campaign Assistant")
-    gr.Markdown("Generate **smart, creative, and data-grounded campaigns** with retrieval-augmented AI.")
-
-    with gr.Tab("🔎 Explore Dataset"):
-        search_query = gr.Textbox(label="Search dataset by keyword / theme")
-        search_results = gr.Textbox(label="Top Matches", lines=10)
-        search_btn = gr.Button("Search")
-        search_btn.click(search_dataset, inputs=search_query, outputs=search_results)
-
-    with gr.Tab("✍️ Generate Campaign"):
-        with gr.Row():
-            prompt = gr.Textbox(label="Enter campaign idea / theme", lines=3)
-        with gr.Row():
-            k_slider = gr.Slider(1, 10, value=3, step=1, label="Number of supporting facts (k)")
-            temp_slider = gr.Slider(0.3, 1.2, value=0.7, step=0.1, label="Creativity (temperature)")
-        campaign_output = gr.Textbox(label="Generated Campaign", lines=15)
-        gen_btn = gr.Button("Generate with RAG")
-        gen_btn.click(generate_with_rag, inputs=[prompt, k_slider, temp_slider], outputs=campaign_output)
+with gr.Blocks(fill_height=True, theme=gr.themes.Soft, css=custom_css) as demo:
+    # Top badges
+    with gr.Row(elem_classes="badge-container"):
+        gr.HTML("""
+            <div style="display: flex; justify-content: center; align-items: center; gap: 15px;">
+                <a href="https://huggingface.co/spaces/VIDraft/gpt-oss-RAG" target="_blank">
+                    <img src="https://img.shields.io/static/v1?label=gpt-oss-20b&message=RAG&color=%23000080&labelColor=%23ffa500&logo=huggingface&logoColor=white&style=for-the-badge" alt="badge">
+                </a>
+                <a href="https://discord.gg/openfreeai" target="_blank">
+                    <img src="https://img.shields.io/static/v1?label=Discord&message=Openfree%20AI&color=%23000080&labelColor=%23ffa500&logo=discord&logoColor=white&style=for-the-badge" alt="badge">
+                </a>
+            </div>
+        """)
+
+    with gr.Row():
+        # Sidebar
+        with gr.Column(scale=1):
+            with gr.Group(elem_classes="main-container"):
+                gr.Markdown("# 🚀 Inference Provider")
+                model_dropdown = gr.Dropdown(
+                    choices=list(model_choices.keys()),
+                    value=current_model_id,
+                    label="📊 Select Model"
+                )
+                reload_btn = gr.Button("🔄 Apply Model Change", variant="primary")
+
+                with gr.Accordion("⚙️ Advanced Options", open=False):
+                    temperature = gr.Slider(
+                        minimum=0, maximum=2, value=0.8, step=0.1, label="Temperature"
+                    )
+                    max_tokens = gr.Slider(
+                        minimum=50, maximum=1024, value=250, step=10, label="Max Tokens"
+                    )
+
+        # Main chat area
+        with gr.Column(scale=3):
+            with gr.Group(elem_classes="main-container"):
+                gr.Markdown("## 💬 RAG-powered Creative Campaign Assistant")
+
+                query = gr.Textbox(label="Enter campaign idea / theme", lines=2)
+                output = gr.Textbox(label="Generated Campaign Script", lines=10)
+
+                btn = gr.Button("✨ Generate with RAG")
+
+                btn.click(
+                    generate_with_rag,
+                    inputs=[query, temperature, max_tokens],
+                    outputs=output
+                )
+
+                with gr.Column(visible=True) as model_120b_container:
+                    gr.Markdown("### Active Model: openai/gpt-oss-120b")
+                with gr.Column(visible=False) as model_20b_container:
+                    gr.Markdown("### Active Model: openai/gpt-oss-20b")
+
+    reload_btn.click(
+        fn=switch_model,
+        inputs=[model_dropdown],
+        outputs=[model_120b_container, model_20b_container, gr.State(current_model_id)]
+    )
 
 if __name__ == "__main__":
     demo.launch()