Create app.py

app.py

@@ -0,0 +1,103 @@
+import requests
+import os
+import json
+import gradio as gr
+import logging
+
+# Setup logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Hugging Face API setup
+HF_API_TOKEN = os.getenv("HF_API_TOKEN")
+HF_API_URL = "https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-8B-Instruct"
+hf_headers = {"Authorization": f"Bearer {HF_API_TOKEN}"}
+
+# SAP configuration
+SAP_API_KEY = os.getenv('SAP_API_KEY')
+SAP_BASE_URL = "https://sandbox.api.sap.com/s4hanacloud/sap/opu/odata/sap"
+sap_headers = {
+    "APIKey": SAP_API_KEY,
+    "Accept": "application/json"
+}
+
+def fetch_sales_orders(top=20):
+    url = f"{SAP_BASE_URL}/API_SALES_ORDER_SRV/A_SalesOrder?$top={top}"
+    response = requests.get(url, headers=sap_headers)
+    return response.json().get('d', {}).get('results', [])
+
+def fetch_purchase_orders(top=20):
+    url = f"{SAP_BASE_URL}/API_PURCHASEORDER_PROCESS_SRV/A_PurchaseOrder?$top={top}"
+    response = requests.get(url, headers=sap_headers)
+    return response.json().get('d', {}).get('results', [])
+
+def generate_response(question, context, data_type):
+    context_str = json.dumps(context, indent=2)
+    if len(context_str) > 4000:  # Limit context size
+        context_str = context_str[:4000] + "... (truncated)"
+
+    prompt = f"""
+    You are a helpful SAP data analyst. Answer the user's question based on the provided data.
+
+    Data Type: {data_type}
+
+    Available Data:
+    {context_str}
+
+    User Question: {question}
+
+    Instructions:
+    1. Provide a clear, concise answer based on the data.
+    2. Include specific numbers, dates, or values when relevant.
+    3. If the data doesn't contain enough information, mention this.
+    4. Format your response clearly and concisely.
+
+    Answer:
+    """
+
+    payload = {"inputs": prompt, "parameters": {"max_new_tokens": 500, "temperature": 0.1}}
+    response = requests.post(HF_API_URL, headers=hf_headers, json=payload)
+
+    if response.status_code == 200:
+        output = response.json()
+        generated_text = output[0]['generated_text']
+        return generated_text.replace(prompt, "").strip()
+    else:
+        logger.error(f"Hugging Face API Error: {response.status_code}, {response.text}")
+        return "Error generating response from Hugging Face."
+
+def chat_with_sap(question, history):
+    sales_orders = fetch_sales_orders()
+    purchase_orders = fetch_purchase_orders()
+
+    context = {"sales_orders": sales_orders, "purchase_orders": purchase_orders}
+    data_type = "Sales and Purchase Orders"
+
+    response = generate_response(question, context, data_type)
+    history.append((question, response))
+    return history
+
+# Gradio UI
+with gr.Blocks(title="SAP Order Analytics with LLAMA-3") as demo:
+    gr.Markdown("""
+    # 🚀 SAP Order Analytics Agent (LLAMA-3)
+
+    This agent analyzes SAP Sales and Purchase Orders. Example queries:
+    - "List recent sales orders"
+    - "Total net amount of purchase orders"
+    - "Sales orders for a specific customer"
+    """)
+
+    chatbot = gr.Chatbot(height=500)
+    msg = gr.Textbox(label="Your Question", placeholder="Ask your SAP query here...")
+    submit_btn = gr.Button("Send", variant="primary")
+    clear_btn = gr.Button("Clear")
+
+    submit_btn.click(chat_with_sap, [msg, chatbot], [chatbot])
+    msg.submit(chat_with_sap, [msg, chatbot], [chatbot])
+    clear_btn.click(lambda: [], None, chatbot)
+    submit_btn.click(lambda: "", None, msg)
+    msg.submit(lambda: "", None, msg)
+
+if __name__ == "__main__":
+    demo.launch(debug=True)