Create app.py

app.py

@@ -0,0 +1,102 @@
+import os
+import gradio as gr
+from smolagents import CodeAgent, HfEngine, tool
+from qdrant_client import QdrantClient
+from qdrant_client.models import PointStruct, VectorParams, Distance, Filter, FieldCondition, MatchValue
+from sentence_transformers import SentenceTransformer
+
+# --- 1. Setup ---
+client = QdrantClient(url=os.getenv("QDRANT_URL"), api_key=os.getenv("QDRANT_API_KEY"))
+embed_model = SentenceTransformer("all-MiniLM-L6-v2")
+collection_name = "cricket_analysis"
+
+# Starter Match Config
+STARTER_URL = "https://www.espncricinfo.com/series/icc-men-s-t20-world-cup-2024-1411166/south-africa-vs-india-final-1415755/ball-by-ball-commentary"
+STARTER_ID = "T20_WC_Final_2024"
+
+if not client.collection_exists(collection_name):
+    client.create_collection(
+        collection_name=collection_name,
+        vectors_config=VectorParams(size=384, distance=Distance.COSINE)
+    )
+
+# --- 2. Professional Search Tool ---
+@tool
+def search_cricket_database(query: str, active_match_id: str) -> str:
+    """
+    Search the ball-by-ball database for match events.
+    Args:
+        query: The user's question.
+        active_match_id: The ID of the match to filter results by.
+    """
+    query_vector = embed_model.encode(query).tolist()
+    
+    results = client.search(
+        collection_name=collection_name,
+        query_vector=query_vector,
+        query_filter=Filter(
+            must=[FieldCondition(key="match_id", match=MatchValue(value=active_match_id))]
+        ),
+        limit=8
+    )
+    
+    if not results:
+        return "No specific match events found in the database for this query."
+    
+    return "\n\n".join([f"[Over {r.payload.get('over_number')}] {r.payload['text']}" for r in results])
+
+# --- 3. Professional Agent ---
+# The prompt ensures a high-level analytical tone
+system_prompt = """
+You are a Lead Cricket Strategy Analyst. Your tone is professional, technical, and data-driven.
+When answering questions about the match:
+1. Use the 'search_cricket_database' tool to retrieve facts.
+2. Focus on key turning points, strike rates, and bowling variations mentioned in the commentary.
+3. If the data is missing, state: 'The provided commentary does not contain specific details on this event.'
+"""
+
+agent = CodeAgent(
+    tools=[search_cricket_database], 
+    model=HfEngine("meta-llama/Llama-3-70B-Instruct"),
+    system_prompt=system_prompt
+)
+
+# --- 4. Logic & UI ---
+def chat_fn(message, history, active_match_id):
+    prompt = f"Using context from match {active_match_id}: {message}"
+    response = agent.run(prompt)
+    history.append((message, response))
+    return "", history
+
+def ingest_ui_handler(url, match_id):
+    # This would call your ingestion function (web scraping + upserting)
+    return f"Processing match {match_id} from {url}...", match_id
+
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🏏 Elite Cricket Analytics Portal")
+    
+    with gr.Row():
+        active_id = gr.Textbox(label="Active Match Focus", value=STARTER_ID, interactive=False)
+    
+    with gr.Accordion("Ingest New Match Data", open=False):
+        url_input = gr.Textbox(label="Cricinfo URL")
+        new_id_input = gr.Textbox(label="New Match ID")
+        load_btn = gr.Button("Initialize Analysis")
+        status = gr.Markdown()
+
+    chatbot = gr.Chatbot(label="Analysis Feed")
+    msg = gr.Textbox(label="Query the Analyst")
+
+    msg.submit(chat_fn, [msg, chatbot, active_id], [msg, chatbot])
+    load_btn.click(ingest_ui_handler, [url_input, new_id_input], [status, active_id])
+
+    # Starter Match Loader
+    def check_starter():
+        count = client.count(collection_name=collection_name).count
+        if count == 0:
+            print("Pre-loading Starter Match...")
+            # Here you would trigger the initial scraping for STARTER_URL
+    
+    demo.load(check_starter)
+
+demo.launch()