stage-tracker 0.1.0

Lightweight complete run and individual stage tracker for modularized pipelines

stage-tracker

A lightweight, zero-dependency Python library for tracking multi-stage pipeline executions.
stage-tracker helps developers log structured run metadata, track stage durations, and debug complex flows — all with minimal effort.

---

📦 What is stage-tracker?

When you're running a pipeline — whether in AI research, cybersecurity defense, data processing, or experimentation — it's critical to:

• Capture what ran
• When it ran
• How long it took
• What went wrong (or right)

stage-tracker simplifies this with structured JSON output that records:

• Pipeline-level metadata (e.g., run ID, args, start/end time)
• Named stages, each with start time, duration, output data, and error info

This makes your system observable, auditable, and debuggable.

---

✨ Features

• 🔹 UUID-based run_id for traceability
• 🕒 ISO-formatted timestamps with timezone
• 📊 Track multiple stages independently or sequentially
• 💥 Record exception/error messages per stage
• 📁 Save output as a structured JSON file
• 🧼 Minimal, clean API with no third-party dependencies

---

🧑‍💻 Installation

pip install -e .
# Or using uv (recommended)
uv pip install -e .

---

🚀 Quick Start

from stage_tracker import Tracker

# Initialize a tracker for a given run
tracker = Tracker(args="Reconnaissance_M1")

# Track stages
tracker.start_stage("Detection")
# ... run detection logic ...
tracker.end_stage("Detection", data={"status": "clean"})

tracker.start_stage("Enrichment")
# ... enrichment logic ...
tracker.end_stage("Enrichment", data={"added_cves": 3})

# Finalize and save
tracker.finalize("run_output.json")

Output (run_output.json):

{
  "run_id": "af3c980c-3d22-46e8-83cd-7ff99abee586",
  "start_time": "2025-05-05T23:14:10.235518+00:00",
  "args": "Reconnaissance_M1",
  "stages": {
    "Detection": {
      "start_time": "...",
      "duration_sec": 3.42,
      "data": {"status": "clean"},
      "error": null
    },
    "Enrichment": {
      "start_time": "...",
      "duration_sec": 1.01,
      "data": {"added_cves": 3},
      "error": null
    }
  },
  "completed_at": "...",
  "total_duration_sec": 9.42
}

---

🧩 Use Cases

• 🔍 AI/ML Experiments: Measure durations of preprocessing, training, evaluation.
• 🛡 Cybersecurity Pipelines: Log detection, enrichment, and mitigation stages.
• ⚙️ ETL/Batch Jobs: Trace and debug stage-level runtime or failures.
• 📄 Paper Reproducibility: Attach structured logs for each experiment run.

---

🛠 How It Works

1. Instantiate Tracker(args=...)
2. For each stage:
   • Call start_stage("StageName")
   • Run logic
   • Call end_stage("StageName", data=..., error=...)
3. When done, call tracker.finalize("output.json")

Internally, it tracks timestamps, computes durations, and writes JSON output.

---

📂 Project Structure

stage-tracker/
├── stage_tracker/
│   ├── __init__.py
│   └── tracker.py
├── pyproject.toml
├── LICENSE
└── README.md

---

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.

---

🤝 Contributing

Pull requests are welcome. Feel free to fork and improve! If you find a bug or want to request a feature, please open an issue and I'll do my best to review and address issues promptly.

---

📬 Contact

Maintained by Karim Ahmed.
Email: [e-mail] GitHub: rexmirak