Update app.py

app.py

@@ -1,150 +1,105 @@
-import gradio as gr
-from huggingface_hub import InferenceClient
-import psycopg2
-import os
+from sqlalchemy import create_engine, Table, Column, String, Integer, Float, Text, TIMESTAMP, MetaData
+from sqlalchemy.dialects.postgresql import UUID
+from sqlalchemy import text
+from llama_index.core import SQLDatabase
+from llama_index.core.query_engine import NLSQLTableQueryEngine
+from llama_index.llms.huggingface import HuggingFaceLLM
 import logging
-from datetime import datetime, timezone
 
 # Set up logging
 logging.basicConfig(level=logging.DEBUG)
 logger = logging.getLogger(__name__)
 
-# Hugging Face Zephyr model
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+# PostgreSQL DB connection (converted from JDBC)
+engine = create_engine("postgresql+psycopg2://postgres:password@localhost:5434/postgres")
 
-# Use your existing database connection settings
-DB_CONFIG = {
-    "host": "127.0.0.1",
-    "port": 5434,
-    "database": "postgres",
-    "user": "postgres",
-    "password": "password"
-}
+metadata_obj = MetaData()
 
-# Query TimescaleDB with improved error handling
-def query_timescaledb(sql_query):
-    try:
-        logger.debug("Attempting to connect to the database...")
-        with psycopg2.connect(**DB_CONFIG) as conn:
-            logger.debug("Connected to the database successfully.")
-            with conn.cursor() as cur:
-                cur.execute(sql_query)
-                result = cur.fetchall()
-                logger.debug("Query executed successfully.")
-                return result
-    except Exception as e:
-        logger.error(f"DB Error: {e}")
-        return f"DB Error: {e}"
+# Define the machine_current_log table
+machine_current_log_table = Table(
+    "machine_current_log",
+    metadata_obj,
+    Column("mac", Text, primary_key=True),
+    Column("created_at", TIMESTAMP(timezone=True), primary_key=True),
+    Column("CT1", Float),
+    Column("CT2", Float),
+    Column("CT3", Float),
+    Column("CT_Avg", Float),
+    Column("total_current", Float),
+    Column("state", Text),
+    Column("state_duration", Integer),
+    Column("fault_status", Text),
+    Column("fw_version", Text),
+    Column("machineId", UUID),
+    Column("hi", Text),
+)
 
-# Modified to match your table structure
-def get_sql_for_question(message):
-    message = message.lower()
-    
-    if "average current" in message:
-        return """
-            SELECT AVG("CT_Avg") as avg_current 
-            FROM machine_current_log
-            WHERE created_at >= NOW() - INTERVAL '1 day';
-        """, "Here's the average current over the past 24 hours:"
+# Create the table
+metadata_obj.create_all(engine)
 
-    elif "machine status" in message:
-        return """
-            SELECT mac, state, state_duration, fault_status
-            FROM machine_current_log
-            WHERE created_at = (
-                SELECT MAX(created_at) 
-                FROM machine_current_log
-            )
-            LIMIT 5;
-        """, "Here are the latest machine statuses:"
-
-    elif "current readings" in message:
-        return """
-            SELECT mac, created_at, "CT1", "CT2", "CT3", "CT_Avg"
-            FROM machine_current_log
-            ORDER BY created_at DESC
-            LIMIT 5;
-        """, "Here are the latest current readings:"
-    
-    elif "fault status" in message:
-        return """
-            SELECT fault_status, COUNT(*) 
-            FROM machine_current_log
-            WHERE created_at >= NOW() - INTERVAL '1 day'
-            GROUP BY fault_status;
-        """, "Here's the distribution of fault statuses in the last 24 hours:"
-    
-    elif "firmware versions" in message:
-        return """
-            SELECT DISTINCT fw_version, COUNT(*) 
-            FROM machine_current_log
-            GROUP BY fw_version;
-        """, "Here are the firmware versions in use:"
+# Convert to TimescaleDB hypertable
+with engine.connect() as conn:
+    conn.execute(text("SELECT create_hypertable('machine_current_log', 'created_at', if_not_exists => TRUE);"))
+    print("TimescaleDB hypertable created")
+    conn.commit()
 
-    return None, None
+# Query 1: Get all MAC addresses
+print("\nQuerying all MAC addresses:")
+with engine.connect() as con:
+    rows = con.execute(text("SELECT mac from machine_current_log"))
+    for row in rows:
+        print(row)
 
-# Respond using LLM + data if relevant
-def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
-    sql_query, context_prefix = get_sql_for_question(message)
+# Query 2: Get all data and count
+print("\nQuerying all data and count:")
+stmt = text("""
+    SELECT mac, created_at, CT1, CT2, CT3, CT_Avg, 
+           total_current, state, state_duration, fault_status, 
+           fw_version, machineId
+    FROM machine_current_log
+""")
 
-    if sql_query:
-        result = query_timescaledb(sql_query)
-        if isinstance(result, str):  # error case
-            db_info = result
-        elif not result:
-            db_info = "No data found."
-        else:
-            # Clean and format result
-            db_info = "\n".join(str(row) for row in result)
+with engine.connect() as connection:
+    print("hello")
+    count_stmt = text("SELECT COUNT(*) FROM machine_current_log")
+    count = connection.execute(count_stmt).scalar()
+    print(f"Total number of rows in table: {count}")
+    results = connection.execute(stmt).fetchall()
+    print(results)
 
-        message = f"{context_prefix}\n{db_info}\n\nAnswer the user's query based on this information."
+# Set up LlamaIndex natural language querying
+sql_database = SQLDatabase(engine)
 
-    messages = [{"role": "system", "content": system_message}]
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-    messages.append({"role": "user", "content": message})
+llm = HuggingFaceLLM(
+    model_name="HuggingFaceH4/zephyr-7b-beta",
+    context_window=2048,
+    max_new_tokens=256,
+    generate_kwargs={"temperature": 0.7, "top_p": 0.95},
+)
 
-    response = ""
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-        response += token
-        yield response
+query_engine = NLSQLTableQueryEngine(
+    sql_database=sql_database,
+    tables=["machine_current_log"],
+    llm=llm
+)
 
-# Gradio UI
-with gr.Blocks() as demo:
-    gr.Markdown("## 🤖 Machine Monitoring Assistant")
-    gr.Markdown(
-        """
-        Welcome to the **Machine Monitoring Assistant**. You can ask questions about:
-        - Current readings (CT1, CT2, CT3, CT_Avg)
-        - Machine status and state duration
-        - Fault status
-        - Firmware versions
-        """
-    )
-    
-    gr.ChatInterface(
-        respond,
-        additional_inputs=[
-            gr.Textbox(
-                value="You are an expert AI assistant for machine monitoring. Help users understand machine metrics and status using the latest database values.",
-                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)"),
-        ],
-    )
+def natural_language_query(question: str):
+    try:
+        response = query_engine.query(question)
+        return str(response)
+    except Exception as e:
+        logger.error(f"Query error: {e}")
+        return f"Error processing query: {str(e)}"
 
-# Run
 if __name__ == "__main__":
-    demo.launch()
+    # Natural language query examples
+    print("\nNatural Language Query Examples:")
+    questions = [
+        "What is the average CT1 reading?",
+        "Which machine has the highest total current?",
+        "Show me the latest fault status for each machine"
+    ]
+    
+    for question in questions:
+        print(f"\nQuestion: {question}")
+        print("Answer:", natural_language_query(question))