pySEAFOM 0.1.9

DAS Performance Analysis Tools

pySEAFOM

A Python library for performance analysis and testing of Distributed Acoustic Sensing (DAS) interrogators, developed by SEAFOM's Measuring Sensor Performance group. This package provides standardized tools for testing, benchmarking, and performance evaluation of DAS systems following SEAFOM recommended procedures.

🌐 Purpose

To promote transparency, consistency, and collaboration in the evaluation of DAS interrogator performance by providing open-source tools and standardized workflows.

⚡ Quick Start

Installation

pip install pySEAFOM

Basic Usage

Option 1: Import specific functions directly

from pySEAFOM import calculate_self_noise, plot_combined_self_noise_db

import numpy as np

Option 2: Import modules (recommended when using multiple engines)

import pySEAFOM

import numpy as np

  

# Load your DAS data (channels × time samples)

data = np.load('your_das_data.npy')  # Shape: (n_channels, n_samples)

  

# Define test sections (channel ranges to analyze)

sections = [data[0:50, :], data[100:150, :]]  # Two cable sections

section_names = ['Section A', 'Section B']

  

# Calculate self-noise for each section (using direct import)

results = calculate_self_noise(

    sections,

    interrogation_rate=10000,  # Hz

    gauge_length=10.0,         # meters

    window_function='blackman-harris',

    data_type='pε'             # picostrain

)

  

# OR using module import:

# results = pySEAFOM.self_noise.calculate_self_noise(

    sections,

    interrogation_rate=10000,  # Hz

    gauge_length=10.0,         # meters

    window_function='blackman-harris',

    data_type='pε'             # picostrain

)

  

# Visualize results

plot_combined_self_noise_db(

    results=results,

    test_sections=section_names,

    gauge_length=10.0,

    org_data_unit='pε',

    title='DAS Self-Noise Test Results'

)


# Fidelity (THD) example (single-section call; loop sections externally)
section_ranges = [[0, 49], [100, 149]]
section_names = ['Section A', 'Section B']

for name, (ch0, ch1) in zip(section_names, section_ranges):
  section = data[ch0:ch1 + 1, :]
  fidelity_results = pySEAFOM.fidelity.calculate_fidelity_thd(
    section,
    fs=10000,
    levels_time_steps=[[0, 600000], [660000, 1500000]],
    stimulus_freq=500,
    snr_threshold_db=-40,
    section_name=name,
  )
  pySEAFOM.fidelity.report_fidelity_thd(fidelity_results)

📁 Features & Modules

Current Modules

pySEAFOM.self_noise

Self-noise analysis

pySEAFOM.dynamic_range

Dynamic range analysis

pySEAFOM.fidelity

Fidelity (THD) analysis

pySEAFOM.crosstalk

Crosstalk analysis

pySEAFOM.frequency_response

Frequency response analysis

Future Modules (Planned)

• Spatial Resolution: Gauge length verification

• Noise Floor: System noise characterization

📚 Documentation

Live site (GitHub Pages): https://seafom-fiber-optic-monitoring-group.github.io/pySEAFOM/

Main Functions (self_noise)

calculate_self_noise()

Computes RMS amplitude spectral density across channels.

Parameters:

• sections (list): List of 2D arrays (channels × samples) for each test section
• interrogation_rate (float): Sampling frequency in Hz
• gauge_length (float): Gauge length in meters
• window_function (str): FFT window type ('blackman-harris', 'hann', 'none', etc.)
• data_type (str): Data unit ('pε', 'nε', 'rad', or custom)

Returns:

• List of tuples: [(frequencies, asd), ...] for each section

plot_combined_self_noise_db()

Creates publication-quality self-noise plots.

Parameters:

• results: Output from calculate_self_noise()
• test_sections (list): Section names
• gauge_length (float): Gauge length in meters
• data_unit (str): Display unit
• title (str): Plot title
• sampling_freq (float): Sampling rate (for metadata box)
• n_channels (int): Total channels (for metadata box)
• duration (float): Recording duration (for metadata box)

report_self_noise()

Prints formatted text report.

Main Functions (dynamic_range)

load_dynamic_range_data()

Loads one (or many) .npy files, builds a 2D matrix, and extracts a 1D trace at a chosen spatial position.

Parameters:

• folder_or_file (str): Folder with .npy files or a single .npy file
• fs (float): Sampling / interrogator rate in Hz
• delta_x_m (float): Spatial step between channels [m]
• x1_m (float): Spatial window start [m]
• x2_m (float): Spatial window end [m]
• test_sections_channels (float): Position inside the spatial window [m]
• time_start_s (float): Analysis window start time [s]
• duration (float | None): Analysis window duration [s]
• average_over_cols (int): Number of adjacent channels to average
• matrix_layout (str): 'time_space', 'space_time', or 'auto'

Returns:

• (time_s, trace) where:
  • time_s is a 1D time vector [s]
  • trace is a 1D extracted signal

data_processing()

Optional unit conversion (phase to strain) and optional high-pass filtering for the extracted 1D trace.

Parameters:

• trace (1D array): Input trace (phase [rad] or strain)
• data_is_strain (bool): If False, converts phase [rad] to microstrain [µε]
• gauge_length (float): Gauge length [m] (used for converting)
• highpass_hz (float | None): High-pass cutoff [Hz] (set None to disable)
• fs (float): Sampling rate [Hz] (required when high-pass is enabled)

Returns:

• 1D array: processed signal (microstrain [µε] if conversion is enabled)

`calculate_dynamic_range_hilbert()

` Hilbert envelope dynamic range test. Compares measured envelope vs theoretical envelope and triggers when the relative error exceeds a threshold.

Parameters:

• time_s (1D array): Time vector [s]

• signal_microstrain (1D array): Trace in microstrain [µε]

• max_strain_microstrain (float): Final theoretical envelope amplitude [µε]

• ref_freq_hz (float): Expected sine frequency [Hz]

• smooth_window_s (float): Envelope smoothing window [s]

• error_threshold_frac (float): Relative error threshold (e.g., 0.3 = 30%)

• safezone_s (float): Initial safe zone where triggering is ignored [s]

• save_results (bool): Save figure + append CSV row

• radian_basis (float | None): If provided withgauge_length, reports peak_over_basis as dB re rad/√Hz (computed from the peak of the last cycle converted from µε → rad). Otherwise the CSV field is empty and the metadata box omits it

• results_dir (str): Output directory

Outputs:

• Prints a formatted summary (trigger time, limit strain, etc.)
• Optional figure: dynamic_range_hilbert.png
• Optional CSV: dynamic_range_hilbert.csv

calculate_dynamic_range_thd()

Sliding THD dynamic range test. Computes THD in a moving window and triggers when THD exceeds a threshold for a minimum duration.

Parameters:

• time_s (1D array): Time vector [s]

• signal_microstrain (1D array): Trace in microstrain [µε]

• ref_freq_hz (float): Expected fundamental frequency [Hz]

• window_s (float): Sliding window length [s]

• overlap (float): Window overlap fraction (e.g., 0.75 = 75%)

• thd_threshold_frac (float): THD threshold (e.g., 0.15 = 15%)

• median_window_s (float): Median smoothing window applied to the THD curve

• min_trigger_duration (float): Minimum continuous violation time to trigger [s]

• safezone_s (float): Initial safe zone where triggering is ignored [s]

• save_results (bool): Save figure + append CSV row

• radian_basis (float | None): If provided withgauge_length, reports peak_over_basis as dB re rad/√Hz (computed from the peak of the last cycle converted from µε → rad). Otherwise the CSV field is empty and the metadata box omits it

• results_dir (str): Output directory

Outputs:

• Prints a formatted summary (trigger time, limit strain, etc.)
• Optional figure: dynamic_range_thd.png
• Optional CSV: dynamic_range_thd.csv

Main Functions (fidelity)

calculate_fidelity_thd()

Computes fidelity as THD (%) at a known stimulus frequency for a single pre-sliced spatial section, across one or more time “levels”.

Inputs (typical):

• time_series_data (2D array): section matrix (channels_in_section × samples)
• fs (float): Sampling frequency [Hz]
• levels_time_steps (list[[start,end]] | [start,end]): Sample index range(s) per stimulus level
• stimulus_freq (float): Fundamental frequency [Hz]
• snr_threshold_db (float): SNR gate used to accept FFT blocks
• section_name (str, optional): Label used in the report output

Returns:

• A structured dict with one section containing per-level THD and harmonic levels.

report_fidelity_thd()

Prints a compact text summary of calculate_fidelity_thd() results.

Main Functions (crosstalk)

calculate_crosstalk()

Computes a crosstalk profile and maximum crosstalk for a single spatial section centered on the stimulation point.

Returns:

• A result dict containing:
  • crosstalk_db (1D array): dB relative to reference region
  • max_xt_db (float): max crosstalk in the outer region
  • magnitudes (1D array): linear magnitudes at stimulus frequency
  • reference_level (float): linear reference magnitude

plot_crosstalk()

Plots a crosstalk profile (dB vs distance).

report_crosstalk()

Prints a compact text summary of calculate_crosstalk() results.

Main Functions (frequency_response)

load_frequency_response_data()

Loads one (or many) .npy files, builds a 2D matrix, and extracts a 1D trace at a chosen spatial position.

Parameters:

• folder_or_file (str): Folder with .npy files or a single .npy file
• fs (float): Sampling / interrogator rate in Hz
• delta_x_m (float): Spatial step between channels [m]
• stretcher_start_m (float): Spatial window start [m]
• stretcher_end_m (float): Spatial window end [m]
• span_m (int): Number of adjacent channels to average [m]
• matrix_layout (str): 'time_space', 'space_time', or 'auto'

Returns:

• (time_s, trace_raw, distance_m, local_pos_m) where:
  • time_s is a 1D time vector [s]
  • trace_raw is a 1D extracted signal
  • distance_mis the size of the stretcher in [m]
  • local_pos_m is the central position of the stretcher in [m]

data_processing()

Optional unit conversion (phase to strain) and optional high-pass filtering for the extracted 1D trace.

Parameters:

• trace (1D array): Input trace (phase [rad] or strain)
• data_is_strain (bool): If False, converts phase [rad] to microstrain [µε]
• gauge_length (float): Gauge length [m] (used for converting)
• highpass_hz (float | None): High-pass cutoff [Hz] (set None to disable)
• fs (float): Sampling rate [Hz] (required when high-pass is enabled)

Returns:

• 1D array: processed signal (microstrain [µε] if conversion is enabled)

`calculate_frequency_response()

` Frequency response test. Computes the DAS frequency response (FFT magnitude in dB re 1 µε) and the normalized frequency response over the step frequencies.

Parameters:

• time_s (1D array): Time vector [s]
• signal_microstrain (1D array): Local trace in microstrain [µε]
• interrogation_rate_hz (float): Repetition / sampling rate [Hz]
• n_steps (int): Number of frequency steps
• freq_min_frac_nyq (float): Minimum frequency as a fraction of Nyquist (e.g., 0.02)
• freq_max_frac_nyq (float): Maximum frequency as a fraction of Nyquist (e.g., 0.80)
• window_spectrogram_s (float): Spectrogram window length [s] (local diagnostics)
• overlap_spectrogram_frac (float): Spectrogram overlap fraction (0.5 = 50%) (local diagnostics)
• save_results (bool): If True, saves figures + CSV
• results_dir (str): Output directory

Outputs:

• Returns a dictionary with frequency arrays and dB curves
• Optional figure: frequency_response_local_time_spectrogram_fft.png, frequency_response.pngand frequency_response_normalized.png
• Optional CSV: frequency_response_normalized.csv

🧪 Example Notebook

See self_noise_test.ipynb for a complete example using synthetic data:

• Generates known ASD synthetic signals
• Validates calculation accuracy
• Demonstrates all visualization options

See dynamic_range_test.ipynb for a complete example using synthetic data:

• Extract and process data from a npy DAS matrix
• Calculates dynamic range limit using Hilbert (delta_t_from_window_start [s], peak_last_cycle [µε], peak_over_basis [dB re rad/√Hz])
• Calculates dynamic range limit using THD (delta_t_from_window_start [s], peak_last_cycle [µε], peak_over_basis [dB re rad/√Hz])

See fidelity_test.ipynb for a complete example using synthetic data:

• Builds two time “levels” with different harmonic content
• Runs per-section THD using calculate_fidelity_thd()
• Prints a simple report via report_fidelity_thd()

See crosstalk_test.ipynb for a complete example using synthetic data:

• Generates synthetic stimulated data centered on a stimulation point
• Computes crosstalk using calculate_crosstalk()
• Plots the profile using plot_crosstalk() and prints a report via report_crosstalk()

See frequency_response_test.ipynb for a complete example using synthetic data:

• Extract and process data from a npy DAS matrix
• Calculates DAS Frequency Response in [Strain (dB re 1 µε)])
• Calculates Normalized Frequency Response in [dB])

📊 Typical Workflow

Self-Noise Workflow

1. Prepare Data: Load DAS measurements (channels × samples)
2. Define Sections: Select channel ranges for analysis
3. Calculate Self-Noise: Use calculate_self_noise() with appropriate parameters
4. Visualize: Create plots with plot_combined_self_noise_db()
5. Report: Generate text summaries with report_self_noise()

Dynamic Range Workflow

1. Prepare Data: Load DAS measurements (time × channels) from .npy
2. Extract Trace: Use load_dynamic_range_data() to pick x1_m/x2_m, select POS, and average channels
3. Pre-process: Use data_processing() for phase to strain (if needed) and high-pass (optional)
4. Hilbert Test: Run calculate_dynamic_range_hilbert() to detect envelope-error trigger
5. THD Test: Run calculate_dynamic_range_thd() to detect harmonic-distortion trigger
6. Report / Save: Store plots + CSV summaries for traceability

Fidelity (THD) Workflow

1. Prepare Data: Load DAS measurements (channels × samples)
2. Define Sections: Select channel ranges for analysis
3. Define Levels: Select time windows (sample ranges) for each stimulus level
4. Compute THD: For each section, slice channels and run calculate_fidelity_thd() with stimulus_freq + snr_threshold_db
5. Report: Print summaries using report_fidelity_thd()

Crosstalk Workflow

1. Prepare Data: Load one spatial section centered on the stimulation point (SSL × samples)
2. Compute Crosstalk: Run calculate_crosstalk() with stimulus_freq, fs, gauge_length, and channel_spacing
3. Visualize: Plot profiles with plot_crosstalk()
4. Report: Print summaries using report_crosstalk()

Frequency Response Workflow

1. Prepare Data: Load DAS measurements (time × channels) from .npy
2. Extract Trace: Use load_frequency_response_data() to pick STRETCHER_START_M/STRETCHER_END_M center position and average channels
3. Pre-process: Use data_processing() for phase to strain (if needed) and high-pass (optional)
4. Frequency Response Test: Run calculate_frequency_response() to calculate frequency response
5. Report / Save: Store plots + CSV summaries for traceability

🔧 Development Setup

# Clone the repository

git clone https://github.com/SEAFOM-Fiber-Optic-Monitoring-Group/pySEAFOM.git

cd pySEAFOM

  

# Install in development mode

pip install -e .

  

# Install development dependencies

pip install -e ".[dev]"

  

# Run tests (if available)

pytest tests/

📦 Package Structure

pySEAFOM/

├── source/

│   └── simulation_dynamic_range.py      # generate data for dynamic_range

│   └── simulation_frequency_response.py   # generate data for frequency_response

│   └── pySEAFOM/

│       ├── __init__.py            # package exports

│       └── self_noise.py          # self-noise analysis engine

│       └── dynamic_range.py          # dynamic_range analysis engine

│       └── fidelity.py             # fidelity / THD analysis engine

│       └── crosstalk.py             # crosstalk analysis engine

│       └── frequency_response.py      # frequency_response analysis engine

├── testing_notebooks/

│   └── self_noise_test.ipynb      # synthetic validation notebook

│   └── dynamic_range_test.ipynb      # synthetic validation notebook

│   └── fidelity_test.ipynb         # synthetic validation notebook

│   └── crosstalk_test.ipynb         # crosstalk validation notebook

│   └── frequency_response_test.ipynb  # frequency_response validation notebook

├── workflows/

│   └── SELF_NOISE_WORKFLOW.md     # step-by-step processing summary

│   └── DYNAMIC_RANGE_WORKFLOW.md     # step-by-step processing summary

│   └── FIDELITY_WORKFLOW.md        # step-by-step processing summary

│   └── CROSSTALK_WORKFLOW.md     # step-by-step processing summary

│   └── FREQUENCY_RESPONSE_WORKFLOW.md        # step-by-step processing summary

├── README.md

├── pyproject.toml

├── setup.py

├── MANIFEST.in

└── dist/                         # build artifacts (created on release)

🔌 Adding New Modules

To add a new analysis module:

1. Create source/pySEAFOM/your_module.py with your functions

2. Update source/pySEAFOM/__init__.py:

   ```python

   from . import self_noise, your_module

   ```

3. Add documentation to this README (and module docstrings)

4. Add or update an example notebook under testing_notebooks/

See the existing self_noise.py module as a template.

🤝 Contributing

We welcome contributions from researchers, engineers, and developers working in the fiber optic sensing space. Please see our contribution guidelines to get started.

📜 License

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

This repository follows the SEAFOM Governance Policy.