PyHeal

A self-healing execution library for Python with automatic error recovery, intelligent debugging, and runtime repair.

---

Features

PyHeal wraps Python execution in a recovery layer:

• Execute code
• Detect errors
• Analyze traceback
• Apply fixes (if available)
• Retry execution
• Recover or restart

---

Installation

pip install pyheal

---

Configuration

configure(autofix=False, debug_ai=False, retries=1)

Updates the global configuration settings.

Parameter	Type	Description
autofix	bool	Enables automatic fixing mode
debug_ai	bool	Enables AI-style debug suggestions
retries	int	Default retry count for healing

configure(
    autofix=True,
    debug_ai=True,
    retries=3
)

---

Context Managers

skip()

Suppresses all exceptions inside the block.

Useful when you want code to continue even if errors occur.

with skip():
    x = 1 / 0

print("Program continues")

[pyheal] Ignored error: division by zero
Program continues

---

strict()

Normal Python behavior.

Exceptions are NOT suppressed.

with strict():
    x = 1 / 0

Program crashes normally with traceback.

---

heal(retries=1)

Attempts to "heal" after an exception occurs.

Currently acts as a retry placeholder system.

Parameter	Type	Description
retries	int	Number of healing attempts

with heal(retries=3):
    raise ValueError("Something broke")

[PyHeal] healing attempt 1

If healing fails:

[PyHeal] failed to heal

---

engine

High-level runtime protection context manager.

Suppresses crashes and logs recovery messages.

with engine():
    x = 1 / 0

[PyHeal ENGINE] started
[PyHeal ENGINE] crash recovered: division by zero

If no error occurs:

[PyHeal ENGINE] finished cleanly

---

debug_ai()

Provides debugging hints based on exception types.

Error Type	Suggestion
ZeroDivisionError	Avoid division by zero
IndexError	Check list bounds
FileNotFoundError	Ensure file exists

with debug_ai():
    x = 1 / 0

[PyHeal DEBUG AI]
Error: division by zero
Suggestion: Avoid division by zero (add validation)

---

Decorators

@safe(default=None)

Wraps a function safely.

If the function crashes, the error is caught and a default value is returned instead.

Parameter	Description
default	Value returned when exception occurs

@safe(default=0)
def divide(a, b):
    return a / b

print(divide(10, 0))

[PyHeal SAFE] Caught: division by zero
0

---

Patch System

apply_patch(error_type, fix)

Registers a runtime patch for an error type.

Parameter	Description
error_type	Exception class or identifier
fix	Patch function or description

apply_patch(
    ZeroDivisionError,
    "Add denominator validation"
)

[PyHeal PATCH] <class 'ZeroDivisionError'> -> Add denominator validation

---

watch_and_restart(file)

Runs a Python script and automatically restarts it if it crashes.

Parameter	Description
file	Python file to monitor

watch_and_restart("app.py")

• Runs the script
• Detects non-zero exit codes
• Waits 2 seconds
• Restarts automatically

[PyHeal] Watching: app.py
[PyHeal] Crash detected. Restarting in 2s...

---

Internal Classes

Config

Stores global runtime configuration.

Attribute	Default
autofix	False
debug_ai	False
retries	1

---

PATCHES

Global dictionary storing registered patches.

PATCHES = {
    ZeroDivisionError: "Add validation"
}

---

Notes

• heal() currently contains placeholder healing logic.
• skip() and strict() are implemented twice in the source; the later definitions override the earlier class-based versions.
• Exceptions inside engine() are always suppressed.
• debug_ai() provides simple rule-based suggestions rather than real AI debugging.

---