Main features

Here you will find basic usage of API. For more informations check the API documentation.

0. Setup¶

0.1 Init QGIS¶

In [1]:

from qgis.core import QgsApplication
from niva.niva_provider import NivaProvider


qgs = QgsApplication([], False)
qgs.setPrefixPath("/path/to/qgis/install", True)
qgs.initQgis()

provider = NivaProvider()
qgs.processingRegistry().addProvider(provider)


from processing.core.Processing import Processing
from niva.niva_provider import NivaProvider
Processing.initialize()

# Shut down warnings
import warnings
warnings.filterwarnings('ignore')

from qgis.core import QgsApplication from niva.niva_provider import NivaProvider qgs = QgsApplication([], False) qgs.setPrefixPath("/path/to/qgis/install", True) qgs.initQgis() provider = NivaProvider() qgs.processingRegistry().addProvider(provider) from processing.core.Processing import Processing from niva.niva_provider import NivaProvider Processing.initialize() # Shut down warnings import warnings warnings.filterwarnings('ignore')

Application path not initialized
qt.qpa.plugin: Could not find the Qt platform plugin "wayland" in ""

0.2 Data¶

Enter paths to your layers to run the tutorial.

In [ ]:

from qgis.core import QgsVectorLayer

roi_file_path = (
    "/path/to/roi.shp"  # Region of interest
)

artificial_file_path = (
    "/path/to/data/artif.shp"  # Artificial layer
)
iacs_file_path = "/path/to/data/iacs.shp"  # IACS layer
naa_file_path = (
    "/path/to/data/naa.shp"  # NAA layer optionnal
)

roi = QgsVectorLayer(roi_file_path)
artificial = QgsVectorLayer(artificial_file_path)
iacs = QgsVectorLayer(iacs_file_path)
naa = QgsVectorLayer(naa_file_path)

print(roi.isValid(), artificial.isValid(), iacs.isValid(), naa.isValid())

from qgis.core import QgsVectorLayer roi_file_path = ( "/path/to/roi.shp" # Region of interest ) artificial_file_path = ( "/path/to/data/artif.shp" # Artificial layer ) iacs_file_path = "/path/to/data/iacs.shp" # IACS layer naa_file_path = ( "/path/to/data/naa.shp" # NAA layer optionnal ) roi = QgsVectorLayer(roi_file_path) artificial = QgsVectorLayer(artificial_file_path) iacs = QgsVectorLayer(iacs_file_path) naa = QgsVectorLayer(naa_file_path) print(roi.isValid(), artificial.isValid(), iacs.isValid(), naa.isValid())

True True True True

In [3]:

list(iacs.fields())

list(iacs.fields())

Out[3]:

[<QgsField: source (String)>,
 <QgsField: CODE_CULTU (String)>,
 <QgsField: ID_PARCEL (String)>,
 <QgsField: layer (String)>,
 <QgsField: path (String)>,
 <QgsField: ID.NIVA (Integer)>]

1. Grid¶

Create a grid that fits your ROI (Region Of Interest).

In [ ]:

from niva.core import GridBuilder

grid_builder = GridBuilder(roi)

# Comment to keep only one
# Create
grid = grid_builder.create()
print(grid.featureCount())
# Extract
grid = grid_builder.extract(QgsVectorLayer("/path/to/data/laea_grid.shp"))
print(grid.featureCount())

from niva.core import GridBuilder grid_builder = GridBuilder(roi) # Comment to keep only one # Create grid = grid_builder.create() print(grid.featureCount()) # Extract grid = grid_builder.extract(QgsVectorLayer("/path/to/data/laea_grid.shp")) print(grid.featureCount())

1239
767

2. Map Niva ID to IACS¶

In [ ]:

from niva.core.mapper.niva_mapper import RPGMapper, LucasMapper, CustomCsvMapper

# Choose the mapper you need according to the nomenclature you use to map Niva field.
print(list(iacs.fields()))
mapper = RPGMapper()
iacs_niva = mapper.map(iacs, field="CODE_CULTU")

from niva.core.mapper.niva_mapper import RPGMapper, LucasMapper, CustomCsvMapper # Choose the mapper you need according to the nomenclature you use to map Niva field. print(list(iacs.fields())) mapper = RPGMapper() iacs_niva = mapper.map(iacs, field="CODE_CULTU")

[<QgsField: source (String)>, <QgsField: CODE_CULTU (String)>, <QgsField: ID_PARCEL (String)>, <QgsField: layer (String)>, <QgsField: path (String)>, <QgsField: ID.NIVA (Integer)>]

3. DataManager¶

DataManager is a class that provide logic to extract landscape features at lower level. T1GridDataManager is a subclass to compute landscape features for a cell for Tier 1. It returns a Sample object withou niva index, branch & novelty index.

In [6]:

from niva.core.prediction import T1GridDataManager

data_manager = T1GridDataManager(iacs=iacs, artificial=artificial, naa=naa)
cell = grid.getFeature(100).geometry() # extract target cell geometry
cell_prediction = data_manager.compute(cell)

print(cell_prediction, cell_prediction.to_dict(), cell_prediction.to_feature())

from niva.core.prediction import T1GridDataManager data_manager = T1GridDataManager(iacs=iacs, artificial=artificial, naa=naa) cell = grid.getFeature(100).geometry() # extract target cell geometry cell_prediction = data_manager.compute(cell) print(cell_prediction, cell_prediction.to_dict(), cell_prediction.to_feature())

T1Sample(mdi=0.0, branch='', novelty=<class 'float'>, richness=7, simpson=4.513337042783619, snc=24.109810250065202, mfs=2.4794393381967956) {'mdi': 0.0, 'branch': '', 'novelty': <class 'float'>, 'richness': 7, 'simpson': 4.513337042783619, 'snc': 24.109810250065202, 'mfs': 2.4794393381967956} <qgis._core.QgsFeature object at 0x7ba3ffe879b0>

4. Predict¶

Predict with ContraTree. If 'mdi' is None it's because the novelty index is too high so the prediction isn't valable.

In [7]:

from niva.core.contra import ContraTreeFactory

tree = ContraTreeFactory.build("T1") # build contra tree for Tier 1
cell_prediction.mdi, cell_prediction.novelty, cell_prediction.branch = tree.predict(cell_prediction)
print(cell_prediction.to_dict())

from niva.core.contra import ContraTreeFactory tree = ContraTreeFactory.build("T1") # build contra tree for Tier 1 cell_prediction.mdi, cell_prediction.novelty, cell_prediction.branch = tree.predict(cell_prediction) print(cell_prediction.to_dict())

{'mdi': 55.587498256657234, 'branch': 'FFFU', 'novelty': np.float64(0.3365562464779253), 'richness': 7, 'simpson': 4.513337042783619, 'snc': 24.109810250065202, 'mfs': 2.4794393381967956}

5. Full prediction¶

You can run a full prediction with NivaMdi subclasses. It will return a new layer with Niva index prediction.

In [8]:

from niva.core.prediction import NivaMdiGridT1


niva_mdi = NivaMdiGridT1(grid, iacs=iacs, artificial=artificial, naa=naa)
niva_prediction = niva_mdi.predict_mdi()

from niva.core.prediction import NivaMdiGridT1 niva_mdi = NivaMdiGridT1(grid, iacs=iacs, artificial=artificial, naa=naa) niva_prediction = niva_mdi.predict_mdi()

In [9]:

niva_prediction.isValid()

values = niva_prediction.getFeature(100).attributes()
fields = niva_prediction.getFeature(100).fields()
print(dict(zip([f.name() for f in fields], values)))

niva_prediction.isValid() values = niva_prediction.getFeature(100).attributes() fields = niva_prediction.getFeature(100).fields() print(dict(zip([f.name() for f in fields], values)))

{'mdi': 55.587498256657234, 'branch': 'FFFU', 'novelty': 0.3365562464779253, 'richness': 7.0, 'simpson': 4.513337042783619, 'snc': 24.109810250065202, 'mfs': 2.4794393381967956}