def test_XYextraction(self):
X = processing.extract_ROI(raster,vector)
self.assertRaises(ValueError,processing.extract_ROI,raster,vector,'Type')
assert(X.ndim == 2)
X,y = processing.extract_ROI(raster,vector,'Class')
assert(X.shape[0] == y.shape[0])
X,y,g = processing.extract_ROI(raster,vector,'Class','uniquefid')
assert(X.shape[0] == y.shape[0] == g.shape[0])
self.assertRaises(ValueError,processing.extract_ROI,'wrong/path','wrong/path/too')
assert(processing.extract_ROI(raster,vector).shape[1] == gdal.Open(raster).RasterCount)
self.assertRaises(ValueError,processing.extract_ROI,raster,vector,'kodidk')
def zonal_stats(in_image, in_vector, unique_id, stats=[
'mean', 'median', 'std'], verbose=False):
"""
Extract zonal stats according to an predifined id.
Parameters
-----------
in_image : str.
Path of the raster file where the vector file will be rasterize.
in_vector : str.
Path of the vector file to rasterize.
unique_id : str or False.
If False, MuseoToolBox will create a field called 'uniquefid' using thefunction :func:`_add_vector_unique_fid`.
stats : list, optional (default=['mean','median','std']).
str in list must be a function available from numpy.
For example ['var'] will output the variance per band and per unique id.
verbose : bool or int, optional (default=True).
The higher is the int verbose, the more it will returns informations.
Returns
--------
stats : np.ndarray
Returns as many np.ndarray as number of stats asked.
Stats ordered by bands.
For example (each line correspond to the unique_id ordered asc) :
+-------------+---------------+-------------+
| mean band_1 + mean band_2 | mean band_3 |
+-------------+---------------+-------------+
| mean band_1 + mean band_2 | mean band_3 |
+-------------+---------------+-------------+
Examples
---------
>>> raster,vector = mtb.datasets.load_historical_data()
>>> mean,var = mtb.stats.zonal_stats(raster,vector,'uniquefid',stats=['mean','var'])
>>> mean.shape
(17, 3)
>>> mean[0,:] # mean of the first unique_id
array([117.75219446, 109.80958812, 79.64213369])
>>> var[0,:]
array([1302.29983482, 1250.59980003, 1015.76659747])
"""
if unique_id is False:
_add_vector_unique_fid(in_vector, 'unique_id', verbose=verbose)
unique_id = 'unique_id'
X, y = extract_ROI(in_image, in_vector, unique_id)
n_unique_id = len(np.unique(y))
n_bands = X.shape[1]
out_stats = [np.zeros([n_unique_id, n_bands]) for n in range(len(stats))]
for idx_stat, stat in enumerate(stats):
stat_function = getattr(__import__('numpy'), stat)
for pos, label in enumerate(np.unique(y)):
res = stat_function(X[np.where(y == label)], axis=0)
out_stats[idx_stat][pos, :] = res
return out_stats
n_repeats=2,
random_state=12,
verbose=False)
for tr, vl in SKF.split(X=None, y=y):
print(tr, vl)
###############################################################################
# .. note::
# Split is made to generate each fold
# Show label
for tr, vl in SKF.split(X=None, y=y):
print(y[tr], y[vl])
##############################################################################
# .. note::
# The first one is made with polygon only.
# When learning/predicting, all pixels will be taken in account
# TO generate a full X and y labels, extract samples from ROI
X, y = processing.extract_ROI(raster, vector, field)
for tr, vl in SKF.split(X, y):
print(tr, vl)
print(tr.shape, vl.shape)
##########################
# Plot example
from __drawCVmethods import plotMethod
plotMethod('SKF-pixel')
############################################################################## # Import librairies # ------------------------------------------- from museotoolbox.cross_validation import LeaveOneSubGroupOut from museotoolbox.processing import extract_ROI from museotoolbox import datasets ############################################################################## # Load HistoricalMap dataset # ------------------------------------------- raster, vector = datasets.load_historical_data() field = 'Class' group = 'uniquefid' X, y, s = extract_ROI(raster, vector, field, group) ############################################################################## # Create CV # ------------------------------------------- # if n_splits is False (default), the number of splits will be the smallest # number of subgroup of all labels. valid_size = 0.5 # Means 50% LOSGO = LeaveOneSubGroupOut(verbose=False, random_state=12) # ############################################################################### # .. note:: # Split is made to generate each fold LOSGO.get_n_splits(X, y, s)
def load_pottoks(return_X_y=True, return_target=True, return_only_path=False):
"""
Load two images of pottoks.
Parameters
-----------
return_X_y : bool, optional (default=True)
If True, will return the array in 2d where there are labels (X) and the labels (y)
If False, will return only X in 3d (the array of the image)
return_target : bool, optional (default=True)
If True, will return two arrays in a list
If False, will return only the source (a brown pottok)
Examples
--------
>>> brown_pottok_arr, brown_pottok_label = load_pottoks(return_target=False)
>>> brown_pottok_arr.shape
(4610, 3)
>>> brown_pottok_label.shape
(4610,)
"""
brown_pottok_uri = os.path.join(
__pathFile,
'brownpottok')
black_pottok_uri = os.path.join(
__pathFile,
'blackpottok')
to_return = []
if return_only_path :
to_return.extend([brown_pottok_uri + '.tif', brown_pottok_uri + '.gpkg'])
if return_target :
to_return.extend([black_pottok_uri + '.tif', black_pottok_uri + '.gpkg'])
elif return_X_y:
Xs, ys = extract_ROI(brown_pottok_uri + '.tif',
brown_pottok_uri + '.gpkg', 'level')
to_return.extend([Xs, ys])
if return_target:
Xt, yt = extract_ROI(
black_pottok_uri + '.tif', black_pottok_uri + '.gpkg', 'level')
to_return.extend([Xt, yt])
elif return_X_y is False:
brown_pottok_arr = RasterMath(
brown_pottok_uri + '.tif',
return_3d=True,
verbose=False).get_image_as_array()
to_return.append(brown_pottok_arr)
if return_target:
black_pottok_arr = RasterMath(
black_pottok_uri + '.tif',
return_3d=True,
verbose=False).get_image_as_array()
to_return.append(black_pottok_arr)
return to_return
from museotoolbox.cross_validation import SpatialLeaveOneSubGroupOut
from museotoolbox import datasets, processing
##############################################################################
# Load HistoricalMap dataset
# -------------------------------------------
_, centroid = datasets.load_historical_data(low_res=True, centroid=True)
raster, vector = datasets.load_historical_data(low_res=True)
field = 'Class'
##############################################################################
# Extract label ('Class' field) and groups ('uniquefid' field)
# Compute distanceMatrix with centroid (one point per group)
X, y, groups = processing.extract_ROI(raster, vector, field, 'uniquefid')
distance_matrix, distance_label = processing.get_distance_matrix(
raster, centroid, 'uniquefid')
##############################################################################
# Create CV
# -------------------------------------------
# n_splits will be the number of the least populated class
SLOSGO = SpatialLeaveOneSubGroupOut(distance_thresold=100,
distance_matrix=distance_matrix,
distance_label=distance_label,
random_state=12)
###############################################################################
# .. note::
# -------------------------------------------
from museotoolbox.ai import SuperLearner
from museotoolbox.cross_validation import RandomStratifiedKFold
from museotoolbox.processing import extract_ROI
from museotoolbox import datasets
from sklearn.ensemble import RandomForestClassifier
from sklearn import metrics
##############################################################################
# Load HistoricalMap dataset
# -------------------------------------------
raster,vector = datasets.load_historical_data(low_res=True)
field = 'Class'
X,y = extract_ROI(raster,vector,field)
##############################################################################
# Create CV
# -------------------------------------------
SKF = RandomStratifiedKFold(n_splits=2,
random_state=12,verbose=False)
##############################################################################
# Initialize Random-Forest and metrics
# --------------------------------------
classifier = RandomForestClassifier(random_state=12,n_jobs=1)
#
kappa = metrics.make_scorer(metrics.cohen_kappa_score)
def test_extract_position(self):
X,pixel_position=processing.extract_ROI(raster,vector,get_pixel_position=True,prefer_memory=False)
assert(pixel_position.shape[0] == X.shape[0])
raster, vector = datasets.load_historical_data(low_res=True)
field = 'Class'
group = 'uniquefid'
##############################################################################
# Create CV
# -------------------------------------------
valid_size = 0.5 # Means 50%
LPSGO = LeavePSubGroupOut(valid_size=0.5, random_state=12, verbose=False)
###############################################################################
# Extract X,y and group.
# -------------------------------------------
X, y, g = processing.extract_ROI(raster, vector, field, group)
###############################################################################
# .. note::
# Split is made to generate each fold
for tr, vl in LPSGO.split(X, y, g):
print(tr.shape, vl.shape)
print('y label with number of samples')
print(np.unique(y[tr], return_counts=True))
##############################################################################
# Differences with scikit-learn
# -------------------------------------------
from sklearn.model_selection import LeavePGroupsOut
# You need to specify the number of groups