tranq 0.2.2

Modern, decorator-based error handling for Python – calm your code.

tranq

Calm error handling for Python – decorator-based, zero boilerplate.

---

Why tranq?

Writing repetitive try/except blocks clutters your code and hides the business logic. tranq gives you declarative error handling with decorators, context managers, and a rich set of retry strategies – so you can focus on what your code does, not how it recovers from failures.

• 🧘 Tranquil – clean, readable, and maintainable.
• 🔁 Smart retries – exponential, linear, Fibonacci backoff, jitter, and max delay.
• 🚦 Circuit Breaker – prevent cascading failures (sync & async).
• 🧪 Conditional retry – on specific exceptions or result values.
• 📦 Retry groups – all-or-nothing execution for multiple functions.
• 📊 Built‑in metrics & profiling – monitor performance and error rates.
• 📝 Pluggable reporters – send errors to files, Sentry, Slack, or custom destinations.
• 🧩 Context manager API – use with tranq.retry(...): when decorators aren't ideal.
• 🔧 Stateful retry – persist attempt count across calls.
• 🎭 Mock error injection – test your error handling with ease.

---

Installation

pip install tranq

Requires Python 3.9 or later.

---

Quick Start

Decorator (@handle)

import tranq

@tranq.handle(on=ValueError, retry=3, delay=0.5, backoff=2.0)
def risky():
    # This will be retried up to 3 times with exponential backoff
    ...

Async (@handle_async)

@tranq.handle_async(on=ConnectionError, retry=2, fallback=lambda: "offline")
async def fetch_data():
    ...

Circuit Breaker

cb = tranq.CircuitBreaker(failure_threshold=5, timeout=60)
@tranq.handle(circuit_breaker=cb)
def call_unstable_service():
    ...

Context Manager

with tranq.retry(on=ValueError, retry=2) as ctx:
    result = ctx.run(my_function, arg1, arg2)

Retry Group (all‑or‑nothing)

group = tranq.retry_group(step1, step2, step3, on=Exception, retry=1)
results = group.run()  # if any step fails, all are retried together

---

Features in Depth

1. Retry with Backoff

Choose from exponential, linear, or Fibonacci backoff. Add jitter to avoid thundering herds.

@tranq.handle(
    on=TimeoutError,
    retry=5,
    delay=0.1,
    backoff=2.0,
    backoff_strategy="exponential",  # "linear", "fibonacci", or custom callable
    max_delay=10.0,
    jitter=True,
)
def fetch():
    ...

2. Conditional Retry

· retry_if – retry only when the exception matches a condition. · retry_on_result – retry if the result is unacceptable (e.g., None).

@tranq.handle(
    on=requests.RequestException,
    retry_if=lambda e: e.response.status_code == 429,  # rate‑limit
    retry=3,
)
def call_api():
    ...

@tranq.handle(
    retry_on_result=lambda result: result is None,
    retry=2,
)
def get_data():
    ...

3. Error Handlers (on_error)

Run different callbacks for different exception types.

def log_warning(e):
    print(f"Warning: {e}")

def alert_admin(e):
    send_alert(e)

@tranq.handle(
    on=(ValueError, ConnectionError),
    on_error={ValueError: log_warning, ConnectionError: alert_admin},
)
def process():
    ...

4. Circuit Breaker (Sync & Async)

Prevent repeated calls to a failing service. Available as CircuitBreaker (sync) and AsyncCircuitBreaker (async).

from tranq import CircuitBreaker, AsyncCircuitBreaker

# Sync
cb = CircuitBreaker(failure_threshold=3, timeout=30, half_open_requests=1)
@tranq.handle(circuit_breaker=cb)
def sync_call():
    ...

# Async
acb = AsyncCircuitBreaker(failure_threshold=3, timeout=30)
@tranq.handle_async(circuit_breaker=acb)
async def async_call():
    ...

5. Stateful Retry

Persist the attempt counter across multiple invocations – useful for batch processing.

@tranq.handle(on=ValueError, retry=3, stateful=True)
def process_item(item):
    # If it fails, the next call continues from the same attempt number
    ...

6. Reporters

Send error details to files, Sentry, Slack, or your own reporter.

from tranq import FileReporter, SentryReporter, SlackReporter

reporters = [
    FileReporter("/var/log/tranq_errors.log"),
    SentryReporter(dsn="..."),
    SlackReporter(webhook_url="..."),
]

@tranq.handle(on=Exception, reporters=reporters)
def critical_task():
    ...

Implement your own by subclassing Reporter and defining report(exception, context).

7. Metrics & Profiling

· Metrics: track call count, error count, and total duration (enable with metrics=True). · Profiling: use the @profile decorator to measure function runtime.

@tranq.handle(metrics=True, metric_prefix="myapp")
def expensive_op():
    ...

from tranq import get_metrics, profile, get_profile

@profile
def heavy_computation():
    ...

print(get_metrics())          # all metric data
print(get_profile("heavy_computation"))  # calls, total_duration

8. Mock Error Injection (Testing)

Inject errors with a given probability to test your error‑handling logic.

from tranq import mock_errors

with mock_errors(ValueError, probability=0.8):
    # 80% of the time, ValueError is raised inside this block
    result = my_function()

9. Dependency Injection

Pass runtime dependencies directly into your decorated function.

@tranq.handle(inject={"logger": logging.getLogger("app")})
def do_work(logger=None):
    logger.info("Working...")

---

Advanced Examples

Combining Features

cb = CircuitBreaker(failure_threshold=3, timeout=60)

@tranq.handle(
    on=requests.RequestException,
    retry=5,
    backoff_strategy="fibonacci",
    max_delay=30,
    jitter=True,
    retry_if=lambda e: e.response.status_code in (429, 503),
    circuit_breaker=cb,
    metrics=True,
    metric_prefix="api",
    reporters=[FileReporter("api_errors.log")],
    fallback=lambda: {"status": "fallback"},
)
def fetch_from_external_api():
    ...

Retry Group with Mixed Sync/Async

from tranq import retry_group, async_retry_group

def step1(): ...
def step2(): ...
async def step3(): ...

# Sync group (all functions must be sync)
group = retry_group(step1, step2, on=ValueError, retry=2)
results = group.run()

# Async group – mix sync and async functions
async_group = async_retry_group(step1, step3, on=Exception, retry=1)
results = await async_group.run()

---

API Reference

Decorators

· handle(...) · handle_async(...)

Context Manager

· retry(...)

Retry Groups

· retry_group(*funcs, **kwargs) · async_retry_group(*funcs, **kwargs)

Circuit Breakers

· CircuitBreaker(failure_threshold, timeout, half_open_requests) · AsyncCircuitBreaker(...)

Policies

· Policy – dataclass with all configurable parameters. · set_global_policy(policy) – set a default policy for all decorators. · get_global_policy()

Reporters

· Reporter (abstract base class) · FileReporter(file_path) · SentryReporter(dsn) · SlackReporter(webhook_url)

Utilities

· get_metrics(), reset_metrics() · profile(func), get_profile(name=None) · mock_errors(exception, probability)

Exceptions

· TranqError – base exception. · RetryExhaustedError – raised when retries are exhausted and reraise=True. · CircuitBreakerError – raised when circuit is open. · ResultNotAcceptedError – raised when retry_on_result condition fails. · RetryGroupError – raised by retry groups on failure.

---

Contributing

Contributions are welcome! Please open an issue or submit a pull request on GitHub.

1. Fork the repository.
2. Create a feature branch.
3. Install development dependencies: pip install -e '.[dev]'
4. Run tests: pytest
5. Submit a PR.

---

License

MIT © RaptorVampire

---

Acknowledgements

Inspired by libraries like tenacity and backoff, but built with a focus on simplicity, modern Python features, and a consistent API for both sync and async code.

---

Happy error handling! 🧘