ecotrace 1.3.0

EcoTrace: High-precision carbon tracking engine for production Python. Function-level CPU/GPU emission measurement with 50ms continuous sampling, Boavizta TDP database, and audit-ready PDF reporting.

EcoTrace

High-Precision Energy and Emissions Instrumentation

---

v1.3.0 — Feature Release. Multi-run history, programmatic get_summary() API, Apple Silicon energy monitoring, per-epoch ML callbacks, and a live browser dashboard — all with zero new mandatory dependencies.

EcoTrace is a lightweight library for granular carbon footprint measurement of Python applications. No configuration files, no background services—just real-time hardware-level transparency.

Real-time monitoring | 50+ Global Zones | AI-powered insights | Zero-configuration

[!TIP] VS Code Extension: Monitor application carbon footprint in real-time during development. Download here.

Function-level carbon measurement with real-time monitoring

---

Core Features in v1.3.0

Feature Release. v1.3.0 adds multi-run observability, a programmatic summary API, Apple Silicon hardware energy reading, per-epoch ML callbacks, and a live browser dashboard.

• Multi-Run History — Every session gets a unique RunID. Use ecotrace history to compare runs and ecotrace trends for an ASCII carbon trend chart.
• get_summary() API — Programmatic access to all session metrics as a typed dict. Ideal for notebooks, dashboards, and CI assertions.
• Apple Silicon Support — Exact CPU energy readings via powermetrics on M-series Macs (falls back to Boavizta silently).
• Per-Epoch ML Callbacks — EcoTracePyTorchCallback and EcoTraceKerasCallback log per-epoch carbon to CSV. PyTorch and TensorFlow are optional extras.
• Live Dashboard — ecotrace dashboard serves a real-time browser UI on localhost:8585. Zero extra dependencies (stdlib only).
• EcoTraceML Tracking Engine — A dedicated context manager and decorator for measuring carbon and energy footprint of ML model training sessions.
• Windows NVML Auto-Discovery — Dynamically resolves and injects NVML DLL paths to prevent load errors on Windows platforms. See CHANGELOG.md.

---

Quick Install

pip install ecotrace

Optional extras:

pip install ecotrace[gpu]   # NVIDIA GPU support
pip install ecotrace[ai]    # Gemini AI insights
pip install ecotrace[all]   # Everything

---

Quick Start

Option 1: Zero-Code Profiling (CLI)

Measure any script without changing a single line of code:

ecotrace run my_script.py

Option 2: Programmatic Tracking (Library)

Decorate functions for granular instrumentation:

from ecotrace import EcoTrace

eco = EcoTrace(region_code="US")

@eco.track
def my_function():
    # Your heavy processing here
    pass

my_function()

# Export audit-ready reports or check cumulative totals
eco.generate_pdf_report("carbon_audit.pdf")
print(f"Total Carbon Emitted: {eco.total_carbon} gCO2")

Option 3: Carbon Budget Mode

Set a limit and let EcoTrace enforce it:

eco = EcoTrace(
    region_code="TR",
    carbon_limit=5.0,                   # 5 gCO2 budget
    on_budget_exceeded=lambda t, l: print(f"Budget exceeded: {t:.4f}/{l:.4f} gCO2")
)

@eco.track
def training_pipeline():
    ...

training_pipeline()
print(f"Remaining budget: {eco.remaining_budget} gCO2")

Expected Output

When initialized, EcoTrace performs automated hardware detection:

[EcoTrace] INFO: [INFO] EcoTrace instrumentation session initialized (STATIC).
[EcoTrace] INFO: -----------------------------------------------------
[EcoTrace] INFO: Region        : TR (475 gCO2/kWh)
[EcoTrace] INFO: Hardware Logic: 13th Gen Intel Core i7-13700H
[EcoTrace] INFO: Specifications: 20 Cores | 45.0W TDP
[EcoTrace] INFO: Energy Sensor : Boavizta Advanced Estimation
[EcoTrace] INFO: Memory Config : 15.6 GB DDR4
[EcoTrace] INFO: GPU Accelerator: Intel Iris Xe Graphics (15.0W TDP)
[EcoTrace] INFO: -----------------------------------------------------

At process exit, a session summary is printed automatically:

=======================================================
  EcoTrace — Session Summary
=======================================================
  Duration       : 12.34s
  Functions      : 5 tracked
  Total Carbon   : 0.00312000 gCO2
  Region         : TR (475 gCO2/kWh)
  Budget         : 0.003120 / 5.000000 gCO2 (0.1%) [OK]
  Equivalent     : 0.4 min of LED bulb (10W)
=======================================================

---

CI/CD Integration

Official GitHub Action

Enforce carbon budgets in your pipeline with our official GitHub Action. Add this to your .github/workflows/ci.yml:

- name: EcoTrace Carbon Gate
  uses: Zwony/ecotrace@v1.3.0
  with:
    budget: '10.0'
    region: 'US'

Manual CLI Integration

You can also run the gate manually:

ecotrace gate --budget 10.0

If total emissions exceed the budget, the gate fails with exit code 1 — preventing carbon-heavy code from being merged.

---

Why EcoTrace?

Feature	EcoTrace v1.0	CodeCarbon	CarbonTracker
Sampling Interval	50ms	15s	Per Epoch
Isolation	Process-scoped	System-wide	System-wide
Budget Enforcement	Built-in	No	No
CI/CD Gate	Built-in	No	No
Idle Noise Subtraction	Automatic	No	No
Async Support	Native	Limited	No

• Deep Transparency: Derived from verified manufacturer TDP specifications rather than category averages.
• Fail-Safe Architecture: Guaranteed application continuity even if hardware drivers or API keys are missing.
• Actionable AI: Integrates with Google Gemini to provide specific code optimization advice (optional).

---

Documentation

Full documentation is available at ecotrace.readthedocs.io.

• Architecture and Science — How the energy model and process isolation work.
• Advanced Usage — GPU tracking, AI insights, benchmarks, and comparison tables.
• API Reference — Technical documentation for core classes and functions.
• Support and Reference — Troubleshooting, region codes, and hardware compatibility.

---

Contributing

We welcome contributions! Please see our CONTRIBUTING.MD for guidelines on reporting bugs, suggesting features, or contributing hardware data.

---

Community

CHANGELOG.md · SECURITY.MD

---

Author and License

Emre Ozkal — GitHub · [e-mail]

MIT License — Use it however you like.

Developed for sustainable software development practices.