cgmpy 0.6.0

Modular Python library for Continuous Glucose Monitoring (CGM) data analysis

CGMPy — Continuous Glucose Monitoring Analysis

A modular Python library for analyzing Continuous Glucose Monitoring (CGM) data, with a strong focus on clinical metrics, pregnancy-specific workflows, and reproducible research.

---

What is CGMPy?

CGMPy provides a clean, type-hinted Python API for:

• Loading CGM data from CSV, Parquet, or in-memory pandas.DataFrame with automatic device detection (Dexcom Clarity, FreeStyle Libre / Libreview, Tandem, Medtronic CareLink).
• Processing & validating glucose time series (glucometric ranges, gap detection, interval normalization).
• Computing clinical metrics following international consensus:
  • Basic statistics: mean, median, GMI, glucose CV.
  • Time in range (TIR): TIR, TAR, TBR with customizable targets (diabetes vs. pregnancy).
  • Glycemic variability: SD, CV, MAGE, MODD, CONGA, J-Index, LBGI, HBGI, GRI.
  • Pregnancy-specific metrics for gestational diabetes.
• Visualizing ambulatory glucose profiles (AGP), daily trends, and statistical summaries.
• Comparing results against the AGATA reference implementation for validation.

The library is modular: every component (loader, processor, analyzer, exporter, plotter, metrics) can be used independently or composed through the high-level GlucoseAnalysis facade.

---

Installation

# Stable release
pip install cgmpy

# With AGATA integration (optional, requires agata)
pip install cgmpy[agata]

# With documentation / development tools
pip install cgmpy[dev,docs]

CGMPy requires Python ≥ 3.10.

From source

git clone https://github.com/javierpa95/cgmpy.git
cd cgmpy
pip install -e .[dev]

---

Quickstart

from cgmpy import GlucoseAnalysis

# One-line analysis
analysis = GlucoseAnalysis("path/to/glucose.csv")
report = analysis.get_comprehensive_report()
print(f"Time in Range: {report['time_in_range']['tir']}%")

# Generate the standard Ambulatory Glucose Profile
analysis.plot_comprehensive_dashboard()

Lower-level modular API

from cgmpy.data import ModularGlucoseData
from cgmpy.metrics import ModularGlucoseMetrics
from cgmpy.metrics.targets import get_targets

data = ModularGlucoseData("cgm_data.csv")
data.filter_by_date_range("2024-01-01", "2024-01-31")

# Compute metrics with pregnancy-specific cutoffs
targets = get_targets("pregnancy")
metrics = ModularGlucoseMetrics(data, targets=targets)
print(metrics.time_in_range())
print(metrics.variability().mage())

---

Features

Domain	Capabilities
Data ingestion	CSV, Parquet, DataFrame. Auto delimiter, header detection, device-specific loaders.
Validation	Glucose range check, time interval regularity, gap analysis, data quality scoring.
Basic metrics	Mean, median, GMI, SD, IQR, % time in any custom range.
Time in range	TIR, TAR (level 1/2), TBR (level 1/2) with international consensus cutoffs.
Variability	CV, MAGE, MODD, CONGA, J-Index, LBGI, HBGI, GRI, MAG, ADRR.
Pregnancy	Dedicated GestationalDiabetes metrics, pregnancy glucose targets (63–140 mg/dL).
Visualization	AGP (Ambulatory Glucose Profile), daily traces, statistical dashboards.
AGATA parity	AgataAnalysis wrapper for side-by-side comparison with the AGATA reference.
Export	To Parquet (optimized) / CSV / Excel, with deduplication and type coercion.

---

Documentation

• Getting Started
• User Guide
• API Reference
• Development Guide
• Roadmap
• Changelog

---

Clinical & Research Use

CGMPy is built to support clinical research workflows in diabetes and gestational diabetes. If you use CGMPy in a publication, please cite the version you used (see CITATION.cff).

Important — Not a medical device. CGMPy is a research and analysis tool. It is not a medical device and must not be used as a substitute for professional medical advice. Always validate clinical interpretations with a qualified healthcare provider.

---

Contributing

We welcome contributions! Please read CONTRIBUTING.md for guidelines on reporting bugs, proposing features, and submitting pull requests. By participating, you agree to abide by the Code of Conduct.

Quick links:

• Good first issues
• Help wanted
• Discussion forum

---

Security & Privacy

CGMPy can process sensitive health data. Please review SECURITY.md for:

• How to report vulnerabilities.
• The strict policy on committing Protected Health Information (PHI).
• How to anonymize datasets before sharing examples or bug reports.

Never commit real patient data, real CGM exports with personal identifiers, or any HIPAA/GDPR-protected data. Examples and tests must use synthetic or fully anonymized data only.

---

License

CGMPy is released under the MIT License. See LICENSE for details.

Copyright © 2024–2026 Javier Peñate Arrieta.

---

Acknowledgments

• The clinical metric definitions follow international consensus guidelines (Battelino et al., Diabetes Care 2019; Bergenstal et al., J. Diabetes Sci. Technol. 2013).
• The AGATA library is the reference implementation used for cross-validation.
• Inspired by R packages like iglu and Python packages in the diabetes research ecosystem.