fractopo

fractopo is a Python library/application that contains tools for validating and analysing lineament and fracture trace maps (fracture networks). It is targeted at structural geologists working on the characterization of bedrock fractures from outcrops and through remote sensing. fractopo is available as a Python library and through a command-line interface. As a Python library, the use of fractopo requires prior (Python) programming knowledge. However, if used through the command-line, using fractopo only requires general knowledge of command-line interfaces in your operating system of choice.

• Full Documentation is hosted on GitHub

Overview of fractopo

Visualisation of fractopo data. fractopo analyses the trace data that can e.g. be digitized from drone orthophotographs (=fractures) or from digital elevation models (=lineaments). The displayed branches and nodes are extracted with fractopo.

Installation

pip and poetry installation only supported for linux -based operating systems. For Windows and MacOS install using (ana)conda.

conda

• Only (supported) installation method for Windows and MacOS!

# Create new environment for fractopo (recommended but optional)
conda env create -n fractopo-env
conda activate fractopo-env
# Available on conda-forge channel
conda install -c conda-forge fractopo

pip

The module is on PyPI.

# Non-development installation
pip install fractopo

poetry

For usage:

poetry add fractopo

For development, only poetry installation of fractopo is supported:

git clone https://github.com/nialov/fractopo
cd fractopo
poetry install

Usage

fractopo has two main use cases:

1. Validation of lineament & fracture trace data
2. Analysis of lineament & fracture trace data

Validation is done to make sure the data is valid for the analysis and is crucial as analysis cannot take into account different kinds of geometric and topological inconsistencies between the traces. Capabilities and associated guides are inexhaustively listed in the table below.

Functionality	Tutorial/Guide/Example
Validation of trace data	Validation 1; Validation 2
Visualize trace map data	Visualizing
Topological branches and nodes	Network; Topological
Trace and branch length distributions	Length-distributions
Orientation rose plots	Orientation 1; Orientation 2
Plot topological ternary node and branch proportions	Proportions
Cross-cutting and abutting relationships	Relationships 1; Relationships 2;
Geometric and topological fracture network parameters	Parameters
Contour grids of fracture network parameters	Contour-grids
Multi-scale length distributions	Multi-scale

For a short tutorial on use of fractopo continue reading:

Input data

Reading and writing spatial filetypes is done in geopandas and you should see geopandas documentation for more advanced read-write use cases:

• https://geopandas.org/

Simple example with trace and area data in GeoPackages:

import geopandas as gpd

# Trace data is in a file `traces.gpkg` in current working directory
# Area data is in a file `areas.gpkg` in current working directory
trace_data = gpd.read_file("traces.gpkg")
area_data = gpd.read_file("areas.gpkg")

Trace data should consists of polyline geometries, i.e., of LineString type. Trace data in MultiLineString format area not supported. Area data should consists of polygon geometries, i.e., of either Polygon or MultiPolygon type.

Trace validation

Trace data must be validated using fractopo validation functionality before analysis. The topological analysis of lineament & fracture traces implemented in fractopo will not tolerate uncertainty related to the topological abutting and snapping relationships between traces. See the documentation for further info on validation error types. Trace validation is recommended before all analysis using Network. Trace and target area data can be validated for further analysis with a Validation object:

from fractopo import Validation

validation = Validation(
    trace_data,
    area_data,
    name="mytraces",
    allow_fix=True,
)

# Validation is done explicitly with `run_validation` method
validated_trace_data = validation.run_validation()

Trace validation is also accessible through the fractopo command-line interface, fractopo tracevalidate which is more straightforward to use than through Python calls. Note that all subcommands of fractopo are available by appending them after fractopo.

tracevalidate always requires the target area that delineates trace data.

# Get full up-to-date command-line interface help
fractopo tracevalidate --help

# Basic usage example:
fractopo tracevalidate /path/to/trace_data.shp /path/to/target_area.shp\
   --output /path/to/validated_trace_data.shp

# Or with automatic saving to validated/ directory
fractopo tracevalidate /path/to/trace_data.shp /path/to/target_area.shp\
   --summary

Geometric and topological trace network analysis

fractopo can be used to extract lineament & fracture size, abundance and topological parameters from two-dimensional lineament and fracture trace, branch and node data.

Trace and target area data (GeoDataFrames) are passed into a Network object which has properties and functions for returning and visualizing different parameters and attributes of trace data.

from fractopo import Network

# Initialize Network object and determine the topological branches and nodes
network = Network(
    trace_data,
    area_data,
    # Give the Network a name!
    name="mynetwork",
    # Specify whether to determine topological branches and nodes
    # (Required for almost all analysis)
    determine_branches_nodes=True,
    # Specify the snapping distance threshold to define when traces are
    # snapped to each other. The unit is the same as the one in the
    # coordinate system the trace and area data are in.
    # In default values, fractopo assumes a metric unit and using metric units
    # is heavily recommended.
    snap_threshold=0.001,
    # If the target area used in digitization is a circle, the knowledge can
    # be used in some analysis
    circular_target_area=True,
    # Analysis on traces can be done for the full inputted dataset or the
    # traces can be cropped to the target area before analysis (cropping
    # recommended)
    truncate_traces=True,
)

# Properties are easily accessible
# e.g.,
network.branch_counts
network.node_counts

# Plotting is done by plot_ -prefixed methods
network.plot_trace_lengths()

Network analysis is also available through the fractopo command-line interface but using the Python interface (e.g. jupyter lab, ipython) is recommended when analysing Networks to have access to all available analysis and plotting methods. The command-line entrypoint is opinionated in what outputs it produces. Brief example of command-line entrypoint:

fractopo network /path/to/trace_data.shp /path/to/area_data.shp\
   --name mynetwork

# Use --help to see all up-to-date arguments and help
fractopo network --help

Data analysis workflow visualisation for fracture trace data (KB11). A. Target area for trace digitisation. B. Digitized traces and target area. C. Orthomosaic used as the base raster from which the traces are digitized from. D. Equal-area length-weighted rose plot of the fracture trace azimuths. E. Length distribution analysis of the trace lengths. F. Determined branches and nodes through topological analysis. G. Cross-cut and abutting relationships between chosen azimuth sets. H. Ternary plot of node (X, Y and I) proportions. I. Ternary plot of branch (C-C, C-I, I-I) proportions.

Citing

To cite this software:

Ovaskainen, N., (2023). fractopo: A Python package for fracture
network analysis. Journal of Open Source Software, 8(85), 5300,
https://doi.org/10.21105/joss.05300

• To cite a specific version of fractopo you can use a zenodo provided DOI. E.g. https://doi.org/10.5281/zenodo.5957206 for version v0.2.6. See the zenodo page of fractopo for the DOI of each version: https://doi.org/10.5281/zenodo.5517485

Support

For issues of any kind: please create a GitHub issue here! Alternatively, you can contact the main developer by email at [email protected].

References

For the scientific background, prior works, definition of traces, branches and nodes along with the explanation of the plots and the plotted parameters, you are referred to multiple sources:

• Sanderson and Nixon, 2015
  • Trace and branch size, abundance and topological parameter definitions.
• Ovaskainen et al, 2022
  • Application of fractopo for subsampling analysis of fracture networks.
• Nyberg et al., 2018
  • A similar package to fractopo with a QGIS GUI.
  • NetworkGT GitHub
• Sanderson and Peacock, 2020
  • Discussion around rose plots and justification for using length-weighted equal-area rose plots.
• Alstott et al. 2014
  • Length distribution modelling using the Python 3 powerlaw package which fractopo uses
  • powerlaw GitHub
• Bonnet et al., 2001
  • Length distribution modelling review.
• My Master’s Thesis, Ovaskainen, 2020
  • Plots used in my Thesis were done with an older version of the same code used for this plugin.

Development

• The package interfaces are nearing stability and breaking changes in code should for the most part be included in the CHANGELOG.md after 25.4.2023. However, this is not guaranteed until the version reaches v1.0.0. The interfaces of Network and Validation can be expected to be the most stable.
• For general contributing guidelines, see CONTRIBUTING.rst

Development dependencies for fractopo include:

• poetry

  • Used to handle Python package dependencies.

  # Use poetry run to execute poetry installed cli tools such as invoke,
  # nox and pytest.
  poetry run '<cmd>'

• copier

  • copier is a project templater. Many Python projects follow a similar framework for testing, creating documentations and overall placement of files and configuration. copier allows creating a template project (e.g., https://github.com/nialov/nialov-py-template) which can be firstly

• copier

  • copier is a project templater. Many Python projects follow a similar framework for testing, creating documentations and overall placement of files and configuration. copier allows creating a template project (e.g., https://github.com/nialov/nialov-py-template) which can be firstly cloned as the framework for your own package and secondly to pull updates from the template to your already started project.

  # To pull copier update from github/nialov/nialov-py-template
  poetry run copier update

• nix

  • fractopo is also packaged with nix. nix provides declarative and immutable packaging which should make fractopo last longer.

  # To run the fractopo command-line using nix
  nix run github:nialov/fractopo#fractopo -- --help

• pytest

  • pytest is a Python test runner. It is used to run defined tests to check that the package executes as expected. The defined tests in ./tests contain many regression tests (done with pytest-regressions) that make it almost impossible to add features to fractopo that changes the results of functions and methods.

  # To run tests implemented in ./tests directory and as doctests
  # within project itself:
  poetry run pytest

• coverage

• sphinx

  • Creates documentation from files in ./docs_src.

Big thanks to all maintainers of the above packages!

License

Copyright © 2020-2023, Nikolas Ovaskainen.

---