pfly 0.1.1

A Python logging linter that enforces best practices

Why?

Production incidents are harder to debug when:

• Exception logs lack stack traces (exc_info=True)
• Hot loops spam logs, degrading performance
• Inconsistent logging causes time delay in parsing logs
• Complementary: Use with structlog, loguru — not instead of them

Installation

uv pip install pfly

Usage

pfly path/to/code

pfly path/to/code --config=pfly.yml

Policies

Configuration

Create a pfly.yml:

logger_names: [logger, logging, log]

exception_exc_info:
  levels: [error]
  severity: fail

log_loop:
  levels: [info, debug]
  severity: fail

Examples

LY001: Missing exc_info=True

# Fail
try:
    connect_db()
except Exception:
    logger.error("Connection failed")  # No stack trace!

# Pass
try:
    connect_db()
except Exception:
    logger.error("Connection failed", exc_info=True)

LY002: Logging in hot loops

# Fail
for user in users:
    logger.info(f"Processing {user}")  # Logs every iteration

# Pass
for i, user in enumerate(users):
    if i % 100 == 0:
        logger.info(f"Progress: {i}/{len(users)}")  # Conditional logging

LY003: Inconsistent key-value logging

# Fail
logger.info(f"User {user}")  # Missing key
logger.info("User %s", user)  # Placeholder without key

# Pass
logger.info(f"User user={user}")  # Key-value pair
logger.info("User user=%s", user)  # Key=placeholder format

License

Apache