fractopo 0.0.4

Brittle Geology Analysis Toolkit

fractopo is a Python module that contains tools for validating and analysing lineament and fracture trace maps (fracture networks).

Overview of fractopo

Development status

• In constant development, will have breaking changes.

• Critical issues:

  • Installation on Windows is currently not supported due to problems with installation of gdal-based packages like geopandas.

  • Contour grid sampling is sometimes slow and requires refactor at some point.

  • snap_traces in branch and node determination is not completely stable. Some edge cases cause artifacts which only sometimes are recognized as error branches. (Mostly solved as of 1.3.2021).

    • Reinforces that some amount of responsibility is always in the hands of the digitizer.

    • Issue mostly avoided with a snap_threshold of 0.001

Full documentation

• Documentation hosted on Read the Docs:

  • Documentation

Installation

Currently installation is supported only for linux-based operating systems. There’s a known problem with installing any gdal-based Python package onto a Windows machine. (Could be circumvented at some point by using conda.)

Omit --dev or [dev] for regular installation. Keep if you want to test/develop or otherwise install all development python dependencies.

Pip

The module is on PyPI.

# Non-development installation
pip install fractopo

Or locally for development:

git clone https://github.com/nialov/fractopo --depth 1
cd fractopo
# Omit [dev] from end if you do not want installation for development
pip install --editable .[dev]

Pipenv

git clone https://github.com/nialov/fractopo --depth 1
cd fractopo
# Omit --dev from end if you do not want installation for development
pipenv sync --dev

Usage

See Notebooks with examples for more advanced usage guidance and examples.

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 validation

Trace and target area data can be validated for further analysis with a Validation object.

from fractopo.tval.trace_validation 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 as a command-line script, tracevalidate which is more straightforward to use than through Python calls.

tracevalidate always requires the target area that delineates trace data.

# Get full up-to-date script help

tracevalidate --help

# Basic usage:
# --fix is recommended due to automatic fixing being very minor in effect
# currently
# --output can be omitted. By default the same spatial filetype
# as the input is used and the output is saved as e.g.
# /path/to/validated/trace_data_validated.shp
# i.e. a new folder is created (or used) for validated data
# --summary can be given to print out summary data of validation
# i.e. error types and error counts

tracevalidate /path/to/trace_data.shp /path/to/target_area.shp --fix --output /path/to/output_data.shp

# Or with automatic saving to validated/ directory

tracevalidate /path/to/trace_data.shp /path/to/target_area.shp --fix --summary

Geometric and topological trace network analysis

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.analysis.network import Network
network = Network(
    trace_data, area_data, name="mynetwork", determine_branches_nodes=True,
)

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

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