falyx/README.md

⚔️ Falyx

Falyx is a battle-ready, introspectable CLI framework for building resilient, asynchronous workflows with:

• ✅ Modular action chaining and rollback
• 🔁 Built-in retry handling
• ⚙️ Full lifecycle hooks (before, after, success, error, teardown)
• 📊 Execution tracing, logging, and introspection
• 🧙‍♂️ Async-first design with Process support
• 🧩 Extensible CLI menus and customizable output

Built for developers who value clarity, resilience, and visibility in their terminal workflows.

---

✨ Why Falyx?

Modern CLI tools deserve the same resilience as production systems. Falyx makes it easy to:

• Compose workflows using Action, ChainedAction, or ActionGroup
• Inject the result of one step into the next (last_result)
• Handle flaky operations with retries and exponential backoff
• Roll back safely on failure with structured undo logic
• Add observability with execution timing, result tracking, and hooks
• Run in both interactive and headless (scriptable) modes
• Customize output with Rich Tables (grouping, theming, etc.)

---

🔧 Installation

pip install falyx

Or install from source:

git clone https://github.com/yourname/falyx.git
cd falyx
poetry install

---

⚡ Quick Example

import asyncio
import random

from falyx import Falyx, Action, ChainedAction

# A flaky async step that fails randomly
async def flaky_step():
    await asyncio.sleep(0.2)
    if random.random() < 0.5:
        raise RuntimeError("Random failure!")
    return "ok"

# Create the actions
step1 = Action(name="step_1", action=flaky_step, retry=True)
step2 = Action(name="step_2", action=flaky_step, retry=True)

# Chain the actions
chain = ChainedAction(name="my_pipeline", actions=[step1, step2])

# Create the CLI menu
falyx = Falyx("🚀 Falyx Demo")
falyx.add_command(
    key="R",
    description="Run My Pipeline",
    action=chain,
    logging_hooks=True,
    preview_before_confirm=True,
    confirm=True,
)

# Entry point
if __name__ == "__main__":
    asyncio.run(falyx.run())

❯ python simple.py
                                🚀 Falyx Demo

  [R] Run My Pipeline
  [Y] History                                         [Q] Exit

>

❯ python simple.py run R
Command: 'R' — Run My Pipeline
└── ⛓ ChainedAction 'my_pipeline'
    ├── ⚙ Action 'step_1'
    │   ↻ Retries: 3x, delay 1.0s, backoff 2.0x
    └── ⚙ Action 'step_2'
        ↻ Retries: 3x, delay 1.0s, backoff 2.0x
Confirm execution of R — Run My Pipeline (calls `my_pipeline`)  [Y/n] y
[2025-04-15 22:03:57] WARNING   ⚠️ Retry attempt 1/3 failed due to 'Random failure!'.
✅ Result: ['ok', 'ok']

---

📦 Core Features

• ✅ Async-native Action, ChainedAction, ActionGroup
• 🔁 Retry policies + exponential backoff
• ⛓ Rollbacks on chained failures
• 🎛️ Headless or interactive CLI with argparse and prompt_toolkit
• 📊 Built-in execution registry, result tracking, and timing
• 🧠 Supports ProcessAction for CPU-bound workloads
• 🧩 Custom Table rendering for CLI menu views
• 🔍 Hook lifecycle: before, on_success, on_error, after, on_teardown

---

🔍 Execution Trace

[2025-04-14 10:33:22] DEBUG    [Step 1] ⚙ flaky_step()
[2025-04-14 10:33:22] INFO     [Step 1] 🔁 Retrying (1/3) in 1.0s...
[2025-04-14 10:33:23] DEBUG    [Step 1] ✅ Success | Result: ok
[2025-04-14 10:33:23] DEBUG    [My Pipeline] ✅ Result: ['ok', 'ok']

---

🧱 Core Building Blocks

Action

A single async unit of work. Painless retry support.

ChainedAction

Run tasks in sequence. Supports rollback on failure and context propagation.

ActionGroup

Run tasks in parallel. Useful for fan-out operations like batch API calls.

ProcessAction

Offload CPU-bound work to another process — no extra code needed.

Falyx

Your CLI controller — powers menus, subcommands, history, bottom bars, and more.

ExecutionContext

Tracks metadata, arguments, timing, and results for each action execution.

HookManager

Registers and triggers lifecycle hooks (before, after, on_error, etc.) for actions and commands.

---

🧠 Design Philosophy

“Like a phalanx: organized, resilient, and reliable.”

Falyx is designed for developers who don’t just want CLI tools to run — they want them to fail meaningfully, recover gracefully, and log clearly.

---