Spaces:
Running
Running
File size: 20,196 Bytes
b8277c4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 | """
โ ๏ธ DEPRECATED - DO NOT USE IN PRODUCTION โ ๏ธ
This file (oldAgent.py) is DEPRECATED and maintained only for reference.
MIGRATION GUIDE:
- Use agent.py for all new development
- agent.py provides the same functionality with:
* Production-grade API integration
* Multi-tenant isolation via session state
* No direct database connections
* Centralized data access through data_sources API
* Better error handling and logging
* API key authentication support
This file will be REMOVED in a future release.
Please migrate all code to use agent.py instead.
Last Updated: October 2025
Deprecation Status: ACTIVE - Do not use for new features
Removal Target: Next major release
"""
import os
import json
import re
import logging
from typing import Dict, Set, Any, List, Optional
from collections import defaultdict
from functools import lru_cache
from textwrap import dedent
from sqlalchemy import create_engine, text
from tenacity import retry, stop_after_attempt, wait_exponential
from urllib.parse import quote_plus
import pandas as pd
from datetime import datetime
# Updated imports for comprehensive tracking
from agno.db.sqlite import SqliteDb # Changed from InMemoryDb for persistence
from agno.agent import Agent
from agno.models.google import Gemini
from agno.tools import Toolkit
from agno.tools.reasoning import ReasoningTools
from agno.run.response import RunContext
# Your existing database configuration
DB_DIALECT = "mysql+pymysql"
DB_USER = "root"
DB_PASSWORD = "bwgadmin@2023"
DB_HOST = "65.0.127.253"
DB_PORT = "3306"
DB_NAME = "bookwedgo"
SCHEMA_PATH = r"/content/database_schema.json"
if "GOOGLE_API_KEY" not in os.environ:
print("๐ด WARNING: GOOGLE_API_KEY environment variable not set. The agent will fail.")
encoded_password = quote_plus(DB_PASSWORD)
DB_URL = f"{DB_DIALECT}://{DB_USER}:{encoded_password}@{DB_HOST}:{DB_PORT}/{DB_NAME}"
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)
# Your existing SchemaManager and DatabaseToolkit classes remain the same
class _SchemaManager:
def __init__(self, schema_path: str):
if not os.path.exists(schema_path):
raise FileNotFoundError(f"Schema file not found: {schema_path}. Please upload it.")
with open(schema_path, 'r') as f:
self.schema_data = json.load(f)
# Build the new data structures on initialization
self.relationship_graph = self._build_relationship_graph()
self.keyword_to_columns = self._create_keyword_mappings()
logger.info("SchemaManager initialized with Relationship Graph and Column-Level Mappings.")
def _build_relationship_graph(self) -> Dict[str, Set[str]]:
graph = defaultdict(set)
tables = self.schema_data.get('tables', [])
if isinstance(self.schema_data, list): tables = self.schema_data[0].get('tables', [])
for table in tables:
table_name = table.get('table_name')
if not table_name: continue
for rel in table.get('relationships', []):
referenced_table = rel.get('referenced_table')
if referenced_table:
# Add a two-way connection for easy lookup
graph[table_name].add(referenced_table)
graph[referenced_table].add(table_name)
return graph
@lru_cache(maxsize=None)
def _create_keyword_mappings(self) -> Dict[str, Dict[str, Set[str]]]:
mappings = defaultdict(lambda: defaultdict(set))
tables = self.schema_data.get('tables', [])
if isinstance(self.schema_data, list): tables = self.schema_data[0].get('tables', [])
def split_on_case_and_underscore(s: str) -> List[str]:
parts = re.findall(r'[A-Z][a-z]*|[a-z]+|\d+', s)
return [p.lower() for p in re.split('_', " ".join(parts))]
for table in tables:
table_name = table.get('table_name', '')
for word in split_on_case_and_underscore(table_name):
mappings[word][table_name].add('__table__') # Special key for table name match
for field in table.get('fields', []):
col_name = field.get('name', '')
for word in split_on_case_and_underscore(col_name):
mappings[word][table_name].add(col_name)
for word in re.findall(r'[a-zA-Z]{3,}', field.get('description', '')):
mappings[word.lower()][table_name].add(col_name)
return mappings
def get_related_tables(self, table_name: str) -> Set[str]:
return self.relationship_graph.get(table_name, set())
# --- DatabaseToolkit leveraging the new _SchemaManager ---
class DatabaseToolkit(Toolkit):
def __init__(self, schema_path: str, db_url: str):
super().__init__(name="database_tools", tools=[self.get_filtered_database_schema, self.execute_sql])
self.schema_manager = _SchemaManager(schema_path)
self.engine = create_engine(db_url)
# UPDATED PART 1: _format_schema_for_llm now includes example rows
def _format_schema_for_llm(self, schema: Dict[str, Any]) -> str:
if not schema.get("tables"): return "No relevant tables found."
output_parts = [f"Database Schema: {schema.get('name', 'N/A')}\n"]
for table in schema["tables"]:
table_name = table.get('table_name', 'N/A')
description = table.get('description', '').strip()
output_parts.append(f"---")
output_parts.append(f"\n**Table: `{table_name}`**")
if description: output_parts.append(f"*Description: {description}*")
output_parts.append("\n**Columns:**")
for field in table.get('fields', []):
col_name, col_type, col_example = field.get('name', 'N/A'), field.get('type', 'N/A'), field.get('example', 'N/A')
col_desc = field.get('description', 'No description.').replace(f"This is the '{col_name}' column of the '{table_name}' table.", "").strip()
output_parts.append(f"- `{col_name}` (type: {col_type}, ex: '{col_example}') - {col_desc}")
if table.get("relationships"):
output_parts.append("\n**Relationships:**")
for rel in table["relationships"]:
col, ref_table, ref_col = rel.get('column'), rel.get('referenced_table'), rel.get('referenced_column')
output_parts.append(f"- `{table_name}.{col}` -> `{ref_table}.{ref_col}`")
# --- NEW: Add example rows to the output ---
if table.get("example_rows"):
output_parts.append("\n**Example Rows:**")
try:
df = pd.DataFrame(table["example_rows"])
output_parts.append(f"```\n{df.to_string(index=False)}\n```")
except Exception:
# Fallback if pandas fails for some reason
for row in table["example_rows"]:
output_parts.append(f"- {row}")
output_parts.append("\n")
return "\n".join(output_parts)
# UPDATED PART 2: get_filtered_database_schema now fetches live data samples
def get_filtered_database_schema(self, keywords: List[str]) -> Dict[str, Any]:
logger.info(f"Tool 'get_filtered_database_schema' called with keywords: {keywords}")
if not keywords:
return {"error": "No keywords provided."}
initial_tables = set()
for keyword in keywords:
clean_keyword = keyword.lower().strip()
if clean_keyword in self.schema_manager.keyword_to_columns:
initial_tables.update(self.schema_manager.keyword_to_columns[clean_keyword].keys())
expanded_tables = set(initial_tables)
for table_name in initial_tables:
related_tables = self.schema_manager.get_related_tables(table_name)
expanded_tables.update(related_tables)
logger.info(f"Initial tables: {initial_tables}. Expanded with related tables: {expanded_tables}")
source_tables = self.schema_manager.schema_data.get('tables', [])
if isinstance(self.schema_manager.schema_data, list):
source_tables = self.schema_manager.schema_data[0].get('tables', [])
final_table_objects = [t for t in source_tables if t.get('table_name') in expanded_tables]
# --- NEW: Inject live data samples into the schema object ---
for table_obj in final_table_objects:
table_name = table_obj.get('table_name')
if not table_name: continue
try:
with self.engine.connect() as connection:
sample_query = text(f"SELECT * FROM `{table_name}` LIMIT 3")
sample_result = connection.execute(sample_query)
sample_rows = [dict(row._mapping) for row in sample_result.fetchall()]
table_obj['example_rows'] = sample_rows
except Exception as e:
logger.warning(f"Could not fetch sample rows for table {table_name}: {e}")
table_obj['example_rows'] = []
if not final_table_objects:
return {"formatted_schema_string": "No relevant tables were found."}
final_schema_json = {
"name": self.schema_manager.schema_data.get("schema_name", "database"),
"tables": final_table_objects
}
formatted_string = self._format_schema_for_llm(final_schema_json)
return {"formatted_schema_string": formatted_string}
# UPDATED PART 3: execute_sql now has a pre-execution sanitization layer
@retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=1, min=2, max=5))
def execute_sql(self, sql_query: str) -> Dict[str, Any]:
logger.info(f"Original SQL query from LLM:\n{sql_query}")
# --- INTELLIGENT SANITIZATION LAYER ---
sanitized_query = sql_query
try:
# Fix the DATE_FORMAT percent-sign escaping issue for MySQL
sanitized_query = re.sub(r"DATE_FORMAT\(([^,]+,)\s*'([^']*)%([^']*)'([^)]*)\)", r"DATE_FORMAT(\1 '\2%%\3'\4)", sanitized_query, flags=re.IGNORECASE)
# Safeguard: Ensure COALESCE is used in aggregations
sanitized_query = re.sub(r'(SUM|AVG|MAX|MIN)\(\s*([a-zA-Z0-9_\.]+)\s*\)', r'\1(COALESCE(\2, 0))', sanitized_query, flags=re.IGNORECASE)
if sanitized_query != sql_query:
logger.info(f"Sanitized SQL query for execution:\n{sanitized_query}")
with self.engine.connect() as connection:
result = connection.execute(text(sanitized_query))
results_list = [dict(row._mapping) for row in result.fetchall()]
return {"sql_results": results_list}
except Exception as e:
logger.error(f"SQL execution error: {e}")
# Provide a more helpful error message back to the agent
error_message = f"SQL Error: {str(e)}. Review the query for syntax issues, especially around date functions and aliases. The failed query was: {sanitized_query}"
return {"error": error_message}
# NEW: Enhanced Tool Hook for Complete Logging
def comprehensive_logging_hook(
run_context: RunContext,
function_name: str,
function_call,
arguments: Dict[str, Any]
) -> Any:
"""
Comprehensive tool execution logging hook that saves:
- Tool name and arguments
- Execution timestamp
- Results
- User context
"""
# Access session_state from run_context (Agno v2 API)
if not run_context.session_state:
run_context.session_state = {}
session_state = run_context.session_state
# Initialize logging structure in session state
if "tool_execution_log" not in session_state:
session_state["tool_execution_log"] = []
# Create execution record
execution_start = datetime.now()
execution_record = {
"tool_name": function_name,
"arguments": arguments,
"timestamp": execution_start.isoformat(),
"execution_id": f"{function_name}_{execution_start.timestamp()}"
}
logger.info(f"๐ง Executing tool: {function_name} with args: {arguments}")
try:
# Execute the actual tool
result = function_call(**arguments)
# Log successful execution
execution_end = datetime.now()
execution_record.update({
"result": str(result)[:1000], # Truncate long results
"status": "success",
"duration_ms": (execution_end - execution_start).total_seconds() * 1000,
"completed_at": execution_end.isoformat()
})
logger.info(f"โ
Tool {function_name} completed successfully in {execution_record['duration_ms']:.2f}ms")
except Exception as e:
# Log failed execution
execution_end = datetime.now()
execution_record.update({
"error": str(e),
"status": "failed",
"duration_ms": (execution_end - execution_start).total_seconds() * 1000,
"completed_at": execution_end.isoformat()
})
logger.error(f"โ Tool {function_name} failed: {str(e)}")
raise # Re-raise the exception
finally:
# Always save the execution record
session_state["tool_execution_log"].append(execution_record)
return result
# Enhanced system prompt with logging awareness
system_prompt = dedent("""
You are Sirus, an expert data scientist with comprehensive execution tracking.
**EXECUTION TRACKING:**
- All your tool executions are automatically logged with timestamps, arguments, and results
- Session state maintains a complete audit trail of your analysis process
- Each query execution is tracked for performance and debugging
**GUIDING PRINCIPLES:**
1. **Be a Business Analyst:** Hide technical complexity from users
2. **Be Resilient & Self-Correct:** Use `think` to diagnose and retry on failures
3. **Prioritize Source of Truth:** Choose the best tables for analysis
4. **Decompose Complex Questions:** Break down into simple sub-problems
**MANDATORY THOUGHT PROCESS:**
For complex questions, solve ONE sub-problem at a time:
a) **`think`**: State the sub-problem and plan
b) **`execute_sql`**: Run the query
c) **`think`**: Review results and plan next steps
**SQL REQUIREMENTS:**
- Use MySQL syntax with single % in DATE_FORMAT
- Always use COALESCE in aggregations
- Test with simple queries first if complex ones fail
**FINAL RESPONSE:**
- Start with key numbers in **bold**
- Provide business insights
- Explain methodology simply
- Suggest logical next steps
""")
print("โ
Configuration set. Initializing enhanced agent with comprehensive logging...")
# Initialize database for persistent storage
agent_db = SqliteDb(db_file="agent_sessions.db")
# Initialize toolkits
db_toolkit = DatabaseToolkit(schema_path=SCHEMA_PATH, db_url=DB_URL)
reasoning_tools = ReasoningTools(add_instructions=True, enable_analyze=True, enable_think=True)
# Create enhanced agent with comprehensive tracking
gemini_sql_agent = Agent(
model=Gemini(
id="gemini-2.5-flash",
system_prompt=system_prompt,
thinking_budget=24000,
include_thoughts=True,
),
tools=[db_toolkit, reasoning_tools],
tool_hooks=[comprehensive_logging_hook], # Add the logging hook
# Database and session management
db=agent_db,
add_history_to_context=True,
num_history_runs=3,
read_chat_history=True,
# Session state for tracking
session_state={
"tool_execution_log": [],
"user_context": {},
"analysis_metadata": {}
},
add_session_state_to_context=True, # Make session state available to tools
# Response formatting
markdown=True,
add_datetime_to_context=True,
stream_intermediate_steps=True,
# Error handling
exponential_backoff=True,
delay_between_retries=10
)
# Enhanced execution function with user tracking
def execute_query_with_tracking(question: str, user_id: str, session_id: Optional[str] = None):
"""
Execute a query with comprehensive user and session tracking
"""
print(f"\n๐ User {user_id} asking: '{question}'\n")
# Execute with session tracking
response = gemini_sql_agent.print_response(
question,
stream=True,
session_id=session_id
)
# Get final session state with all execution logs
final_session_state = gemini_sql_agent.get_session_state(session_id)
print(f"\n๐ Execution Summary for User {user_id}:")
print(f" Session ID: {gemini_sql_agent.session_id}")
print(f" Tools executed: {len(final_session_state.get('tool_execution_log', []))}")
# Print tool execution summary
for i, log_entry in enumerate(final_session_state.get('tool_execution_log', [])[-3:], 1):
print(f" Tool {i}: {log_entry['tool_name']} - {log_entry['status']} ({log_entry.get('duration_ms', 0):.1f}ms)")
return response, final_session_state
# Function to retrieve session history and analytics
def get_session_analytics(session_id: str):
"""
Retrieve comprehensive session analytics
"""
session = gemini_sql_agent.get_session(session_id)
if session:
print(f"\n๐ Session Analytics for {session_id}:")
print(f" Created: {session.created_at}")
print(f" Updated: {session.updated_at}")
print(f" Total runs: {len(session.runs) if hasattr(session, 'runs') else 'N/A'}")
# Get session state
session_state = gemini_sql_agent.get_session_state(session_id)
tool_logs = session_state.get('tool_execution_log', [])
print(f" Total tool executions: {len(tool_logs)}")
# Tool usage statistics
tool_usage = {}
for log in tool_logs:
tool_name = log['tool_name']
tool_usage[tool_name] = tool_usage.get(tool_name, 0) + 1
print(" Tool usage breakdown:")
for tool, count in tool_usage.items():
print(f" - {tool}: {count} times")
return session
# Example usage with user tracking
if __name__ == "__main__":
import time
# Example 1: First user query
user_1_response, user_1_state = execute_query_with_tracking(
question="What were our total bookings",
user_id="user_001",
session_id="business_analysis_session_1"
)
time.sleep(30)
# Example 2: Follow-up query in same session
user_1_followup, user_1_final_state = execute_query_with_tracking(
question="Give me the total revenue from these bookings",
user_id="user_001",
session_id="business_analysis_session_1"
)
time.sleep(30)
# Example 3: Different user, new session
user_2_response, user_2_state = execute_query_with_tracking(
question="total no of unique vendors",
user_id="user_002",
session_id="vendor_analysis_session_1"
)
# Get comprehensive analytics
print("\n" + "="*50)
print("COMPREHENSIVE SESSION ANALYTICS")
print("="*50)
get_session_analytics("business_analysis_session_1")
get_session_analytics("vendor_analysis_session_1")
# Export execution logs to JSON for external analysis
def export_execution_logs(session_id: str, filename: str):
session_state = gemini_sql_agent.get_session_state(session_id)
execution_logs = {
"session_id": session_id,
"export_timestamp": datetime.now().isoformat(),
"tool_execution_log": session_state.get('tool_execution_log', []),
"user_context": session_state.get('user_context', {}),
"analysis_metadata": session_state.get('analysis_metadata', {})
}
with open(filename, 'w') as f:
json.dump(execution_logs, f, indent=2)
print(f"๐ Execution logs exported to {filename}")
export_execution_logs("business_analysis_session_1", "user_001_execution_log.json")
export_execution_logs("vendor_analysis_session_1", "user_002_execution_log.json") |