Library hsa been internalized into Metsi. Librarized development will continue later with better defined library responsibilities and interface boundaries between components.
This project is a collection of Python classes for a structured object model of forest stands with contained tree strata and reference trees. We also serve a library implementation for converting forest inventory data into these classes. This project's primary purpose is to serve as the data model and data conversion library for the Metsi forest development simulator.
Original implementation of the data model and data conversion by Arto Haara, Natural Resources Institute Finland.
Backwards compatibility with semantic versioning is guaranteed for users of this library.
This is a Python library developed under Python 3.10. We offer support for this library via GitHub issues and welcome pull requests for improvement.
This project uses the pyproject.toml for configuration and build. For setting up development environment, run
pip install .[tests]. Development can be readily done utilizing the unit test suites in tests. The project is
deployed under the namespace lukefi.metsi.data.
We expect
- semantic commits constraining changes into categories in the spirit of https://sparkbox.com/foundry/semantic_commit_messages
- all functionality to be sufficiently unit tested to prove operation and address regressions
The data structures for forest stands and reference trees were originally created from the RSD data format specification of the MELA forest calculation utility. They have been extended for tree strata. The core data model is gradually being changed from RSD categorizations to be a unified representation for Finnish national forest survey (VMI) and Finnish Forest Center (SMK) data formats. This is achieved by mapping out enumerations and utilizing them for conversion between the source data formats, the Metsi data model and the RSD output model.
We provide an input converter from VMI12 and VMI13 forest inventory data[^1] and Finnish Forest Center forest inventory data[^2]. In addition, output data conversion is provided for creating RSD-like files according to the MELA forest calculation utility 2016 specification. The conversion functionality is under consideration to be moved into a separate code project.
The code for this project is structured into namespace packages as follows.
| path | comment |
|---|---|
| l.m.d.model | Main data structures module |
| l.m.d.conversion | Utility package for converting enumerations between data formats |
| l.m.d.enums | Package for category variable enumerations |
| l.m.d.formats.ForestBuilder | Builder pattern style classes for populating a collection of forest stands with reference tree and stratum data |
| l.m.d.formats.io_utils | Utilities for formatting data for various output formats |
| l.m.d.formats.rsd_const | support structures for RSD data indices |
| l.m.d.formats.smk_util | Forest Centre XML data related parsing logic |
| l.m.d.formats.util | general utility functions |
| l.m.d.formats.vmi_const | support structures for VMI data indices |
| l.m.d.formats.vmi_util | support functionality for VMI data parsing and conversion |
| tests | Test suites |
Main data structure classes for the internal representation are as follows.
| class | description |
|---|---|
| ForestStand | Forest stand is a measurement area of interest for forest statistics and metadata |
| ReferenceTree | Reference tree is a representation of trees of a certain species and size within a forest stand |
| TreeStratum | Tree stratum is a statistical representation of trees of a certain species within a forest stand |
Enumerated properties for above classes are as follows, with equivalent enumerations from source and target data formats
prefixed with Vmi*, ForestCentre* and Mela*:
| class | enumerated property | description |
|---|---|---|
| ForestStand | LandUseCategory | Type of land within this area |
| ForestStand | OwnerCategory | Ownership of this area |
| ForestStand | SoilPeatlandCategory | Soil type of this area |
| ForestStand | SiteType | Site richness of this area |
| ForestStand | DrainageCategory | Drainage status of this area |
| ForestStand | ... | Work in progress |
| ReferenceTree | TreeSpecies | Species of these trees |
| ReferenceTree | ... | Work in progress |
| TreeStratum | TreeSpecies | Species of trees of this stratum |
Geopandas dependency resolution in Windows using pip
For using the Forestry Centre (SMK) XML source files, geopandas library is needed for geocoordinate handling. When
using pip for dependency management, geopandas depends on the Fiona library, which depends on GDAL. Fiona and GDAL are
not properly installable without a C++ build environment.
Workaround:
Use Anaconda[^3] for Python virtual environment and library dependency management.
OR
Obtain WHL prepackaged Fiona and GDAL libraries from https://www.lfd.uci.edu/~gohlke/pythonlibs where you will find a file such
- Fiona‑1.8.21‑cp310‑cp310‑win_amd64.whl
- GDAL‑3.4.2‑cp310‑cp310‑win_amd64.whl
Depending on the python version you are using, download the correct file by replacing the cp310 in the filename above:
- Python 3.10 -> cp310
- Python 3.9 -> cp39
- etc...
Install the package using pip install GDAL‑3.4.2‑cp310‑cp310‑win_amd64.whl
and pip install Fiona‑1.8.21‑cp310‑cp310‑win_amd64.whl before proceeding with pip install -r requirements.txt.
[^2] Finnish Forest Center MV1.9 inventory data
[^3] Anaconda