Create test_scikeo.py

tests/test_scikeo.py

@@ -0,0 +1,161 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# +
+""" Tests for the scikit-eo package. """
+
+import numpy as np
+import pytest
+import rasterio
+import pandas as pd
+from scikeo.process import confintervalML
+from scikeo.sma import sma
+from scikeo.pca import PCA
+from scikeo.tassCap import tassCap
+from scikeo.fusionrs import fusionrs
+from scikeo.process import crop
+from scikeo.rkmeans import rkmeans
+from scikeo.mla import MLA
+from scikeo.process import extract
+from dbfread import DBF
+from scikeo.calmla import calmla
+from scikeo.deeplearning import DL
+
+
+def test_confintervalML():
+
+    """Confusion Matrix by Estimated Proportions of area an uncertainty is tested. 
+    
+    To carry out it, this function was tested using ground-truth values obtained by 
+    Olofsson et al. (2014).
+    """
+    conf_error = pd.read_csv("tests/data/confusion_matrix.csv", index_col= 0, sep = ';')
+    # only confusion matrix values
+    values = conf_error.iloc[0:4,0:4].to_numpy()
+    # number of pixels for each class
+    img = np.array([200000, 150000, 3200000, 6450000])
+    res = confintervalML(matrix = values, image_pred = img, pixel_size = 30, nodata = -9999)
+    assert round(res.get('Overall_accuracy'),3) == 0.947
+
+def test_sma():
+    """Spectral mixture analysis is tested."""
+    # image to be processed
+    img = rasterio.open('tests/data/LC08_232066_20190727.tif')
+    # endmembers
+    endm =[[8980,8508,8704,13636,16579,11420], # soil
+            [8207,7545,6548,16463,9725,6673], # forest
+            [9276,9570,10089,6743,5220,5143], # water
+           ]
+    endm = np.array(endm)
+    # applying the sma function
+    frac = sma(image = img, endmembers = endm)
+    assert frac.shape[2] == 3
+
+def test_pca():
+    """Principal Component Analysis is tested."""
+    # image to be processed
+    img = rasterio.open('tests/data/LC08_232066_20190727.tif')
+    # Applying the PCA function:
+    arr_pca = PCA(image = img, stand_varb = True)
+    img_pca = arr_pca.get('PCA_image')
+    assert img_pca.shape[2] == 6
+
+def test_tassCap():
+    """Tasseled-Cap is tested."""
+    # image to be processed
+    img = rasterio.open('tests/data/LC08_232066_20190727.tif')
+    # Applying the tassCap function:
+    arr_tct = tassCap(image = img, sat = 'Landsat8OLI')
+    assert arr_tct.shape[2] == 3
+
+def test_fusionrs():
+    """Fusion optical and radar satelite images is tested."""
+    # image to be processed
+    # optical
+    optical = rasterio.open('tests/data/LC08_003069_20180906_clip.tif')
+    # radar
+    radar = rasterio.open('tests/data/S1_2018_VV_VH_clip.tif')
+    # Applying the fusionrs function:
+    fusion = fusionrs(optical = optical, radar = radar)
+    # Cumulative variance (%)
+    cum_var = fusion.get('Cumulative_variance')*100
+    assert round(cum_var[8],0) == 100
+
+def test_crop():
+    """cliping a satellite images is tested."""
+    # raster to be clipped
+    path_raster = "tests/data/LC08_232066_20190727.tif"
+    # area of Interes -> shapefile
+    path_shp = "tests/data/clip.shp"
+    # Path where the image will be saved
+    output_path_raster = "tests/data"
+    # The raster name
+    output_name = 'LC08_232066_20190727_clip'
+    # Applying the crop() function:
+    crop(image = path_raster, shp = path_shp, 
+         filename = output_name, 
+         filepath = output_path_raster)
+    clip_image = rasterio.open(output_path_raster + '/' + output_name+ '.tif')
+    assert type(clip_image) == rasterio.io.DatasetReader
+
+def test_kmeans():
+    """k-means classification is tested."""
+    # image to be processed
+    img = rasterio.open('tests/data/LC08_232066_20190727.tif')
+    # Applying rkmeans() algorithm with four classes:
+    arr_rkmeans = rkmeans(image = img, k = 4, max_iter = 300)
+    assert type(arr_rkmeans) == np.ndarray
+
+def test_machineLearning():
+    """Machine learning algorithms are tested."""
+    # image to be processed
+    path_raster = "tests/data/LC08_232066_20190727.tif"
+    img = rasterio.open(path_raster)
+    # endmembers
+    path_endm = "tests/data/endmembers.dbf"
+    endm = DBF(path_endm)
+    endm = pd.DataFrame(endm)
+    # Instance of mla():
+    inst = MLA(image = img, endmembers = endm)
+    # Applying Random Forest with 60% of data to train
+    rf_class = inst.RF(training_split = 0.6)
+    assert rf_class.get('Overall_Accuracy') >= 0.7
+
+def test_CalibratingMachineLearning():
+    """Calibrating machine learning algorithms are tested."""
+    # endmembers
+    path_endm = "tests/data/endmembers.dbf"
+    endm = DBF(path_endm)
+    endm = pd.DataFrame(endm)
+    # Instance of Calmla():
+    inst = calmla(endmembers = endm)
+    # Instance of splitData():
+    data = inst.splitData()
+    # Calibrating with Monte Carlo Cross-Validation Calibration (MCCV)
+    error_mccv = inst.MCCV(split_data = data, models = ('svm', 'dt', 'rf', 'nb'), n_iter = 10)
+    # error of Random Forest
+    error_rf = np.mean(np.array(error_mccv.get('rf')))
+    # error of Decision Tree
+    error_dt = np.mean(np.array(error_mccv.get('dt')))
+    assert error_dt > error_rf
+
+def test_DeepLearning():
+    """Deep Learning is tested."""
+    # image to be processed
+    path_raster = "tests/data/LC08_232066_20190727.tif"
+    img = rasterio.open(path_raster)
+    # endmembers
+    path_endm = "tests/data/endmembers.dbf"
+    endm = DBF(path_endm)
+    endm = pd.DataFrame(endm)
+    # Instance of DL():
+    inst = DL(image = img, endmembers = endm)
+    # Applying the FullyConnected() function of Deep Learning:
+    fc = inst.FullyConnected(hidden_layers = 4, 
+                         hidden_units = [64,16,8,8], 
+                         output_units = 4,
+                         input_shape = (6,), 
+                         epochs = 100, 
+                         batch_size = 32, 
+                         training_split = 0.8)
+
+    assert fc.get('Overall_Accuracy') >= 0.7