Skip to content

python.regex_compile

Performance Medium

Detects regex patterns that are compiled repeatedly inside loops or frequently-called functions instead of being compiled once at module level.

Why It Matters

Compiling regex patterns repeatedly:

Wastes CPU cycles — Regex compilation is expensive
Increases latency — Each request pays the compilation cost
Ignores caching — Python’s re module caches patterns, but cache is limited
Degrades under load — Performance impact scales with request rate

Example

# ❌ Before (compiled every call)
def validate_email(email):
    pattern = re.compile(r'^[\w\.-]+@[\w\.-]+\.\w+$')
    return pattern.match(email) is not None

def process_items(items):
    for item in items:
        if re.match(r'\d{4}-\d{2}-\d{2}', item.date):  # Compiled per iteration
            process(item)

# ✅ After (compiled once)
EMAIL_PATTERN = re.compile(r'^[\w\.-]+@[\w\.-]+\.\w+$')
DATE_PATTERN = re.compile(r'\d{4}-\d{2}-\d{2}')

def validate_email(email):
    return EMAIL_PATTERN.match(email) is not None

def process_items(items):
    for item in items:
        if DATE_PATTERN.match(item.date):
            process(item)

What Unfault Detects

re.compile() inside functions or loops
re.match(), re.search(), re.findall() with literal patterns in loops
Repeated pattern compilation in hot paths

Auto-Fix

Unfault generates patches that move regex compilation to module level:

# Patched: moved to module level
_PATTERN_1 = re.compile(r'\d{4}-\d{2}-\d{2}')

def process_dates(items):
    for item in items:
        if _PATTERN_1.match(item.date):
            process(item)

Performance Impact

Pattern	10,000 matches
Compiled in loop	~150ms
Pre-compiled	~12ms