Skip to content

LOGGING.md

Latest commit

 

History

History
114 lines (86 loc) · 3.28 KB

LOGGING.md

File metadata and controls

114 lines (86 loc) · 3.28 KB

Logging System

Overview

The pipeline now uses a centralized logging system with configurable verbosity levels. Logs are written to both console and files.

Usage

Command Line Flags

# Normal mode (default) - Shows key milestones only
python test.py "<query>"

# Quiet mode - Only warnings and errors
python test.py "<query>" --quiet
python test.py "<query>" -q

# Verbose mode - Shows all information
python test.py "<query>" --verbose
python test.py "<query>" -v

# Debug mode - Shows everything including internals
python test.py "<query>" --debug
python test.py "<query>" -d

# Auto-approve mode (for testing)
python test.py "<query>" --auto

# Dump full ToolGeneratorState after the run
python test.py "<query>" --state

# Run integration projection verification
python test.py "<query>" --verify

Log Files

All runs generate a detailed log file in logs/pipeline_{timestamp}.log containing:

  • Full execution trace
  • All print statements and debug info
  • Timestamps for each operation
  • Module and function names

In Code

from src.logger_config import get_logger, log_section, log_success

logger = get_logger(__name__)

# Use logger instead of print
logger.info("Processing complete")
logger.debug("Debug info")
logger.warning("Warning message")
logger.error("Error occurred")

# Convenience functions
log_section(logger, "SECTION TITLE")
log_success(logger, "Operation completed")

Verbosity Levels

Level Console Output File Output Use Case
quiet ⚠️ ❌ only Everything Production, when you only care about problems
normal ✅ 📍 Key steps Everything Default, clean overview of progress
verbose All info Everything When you want to see what's happening
debug Everything Everything Troubleshooting, development

Log Output Examples

Normal Mode (Clean)

📊 Graph visualization saved to: docs\pipeline_graph.png
📍 Intent extracted
📍 Spec generated  
📍 Code generated
✅ Tool promoted to registry

Debug Mode (Detailed)

DEBUG [intent_extraction] Extracting intent from query...
DEBUG [code_generator] Using template: config/prompts/code_generation.txt
INFO [validator] Schema OK: True, Tests OK: True
DEBUG [executor] Module contents: ['FastMCP', 'analyze_...']

Log File (Always Detailed)

2026-02-04 16:30:15 - INFO - pipeline - build_graph:73 - Graph visualization saved
2026-02-04 16:30:16 - DEBUG - intent_extraction - intent_node:105 - Query: Run ANOVA...
2026-02-04 16:30:18 - INFO - spec_generator - spec_generator_node:87 - Spec generated

Benefits

  1. Clean Terminal: Default mode shows only what matters
  2. Full Audit Trail: Everything saved to log files for debugging
  3. Flexible: Choose verbosity based on your needs
  4. Professional: Standard logging practices, easy to extend
  5. Debuggable: Can enable debug mode when issues occur

Migration Notes

All pipeline modules use structured logging via src/logger_config.py:

  • ✅ test.py - CLI arguments and logger setup
  • ✅ src/intent_extraction.py, src/intent_validator.py
  • ✅ src/spec_generator.py, src/code_generator.py
  • ✅ src/validator.py, src/executor.py
  • ✅ src/promoter.py, src/pipeline.py
  • ✅ src/sandbox.py