// Create a geometry representing an export region.
var HtiPolygon = ee.Geometry.Polygon([[[-74.300537109375, 19.202241064923054],
          [-74.6795654296875, 18.19282519773319],
          [-73.465576171875, 17.86374708440522],
          [-73.3941650390625, 18.104087015773956],
          [-71.5704345703125, 17.97873309555617],
          [-71.685791015625, 18.531700307384043],
          [-71.6583251953125, 18.890695349102117],
          [-71.5704345703125, 19.254108316440032],
          [-71.60888671875, 19.694314241825747],
          [-71.9384765625, 19.89588839936505],
          [-72.6416015625, 20.086888505561017],
          [-73.1195068359375, 20.15394153657739],
          [-73.65234375, 19.880391767822505]]]); 
//Map.addLayer(geometry);
// Start with an image collection for a 1 year period.
var collection =ee.ImageCollection('LANDSAT/LE07/C01/T1_SR')
          .filterBounds(Haiti)
          .filterDate('1999-12-30', '2001-01-28')
          .set('SENSOR_ID', 'OLI_TIRS');

//function to remove cloud          
var maskClear = function(image) {
         // Bits 3 and 5 are cloud shadow and cloud, respectively.
         var cloudShadowBitMask = ee.Number(2).pow(3).int();
         var cloudsBitMask = ee.Number(2).pow(5).int();
         // Get the pixel QA band.
         var qa = image.select('pixel_qa');
         // Both flags should be set to zero, indicating clear conditions.
         var mask = qa.bitwiseAnd(cloudShadowBitMask).eq(0)
         .and(qa.bitwiseAnd(cloudsBitMask).eq(0));
         
         // Return the masked image, scaled to [0, 1].
         return image.updateMask(mask).divide(10000);
         
};

// Map the cloud masking function over the collection.
var collection1 = collection.map(maskClear);
print('collection1',collection1);

var compositeClip = collection1.median()
.select(ee.List([0,1,2,3,4,5,6,7]),
ee.List(['B1', 'B2', 'B3', 'B4', 'B5', 'B6' , 'B7','B8']));


//print(compositeClip)
Map.addLayer(compositeClip,{min:0,max: 0.25, bands: ['B4','B3','B2']});  
          
  

// subsample training polygons with random points
// this ensures all classes have same sample size
// also EE can't handle more than ~5000 cells in training areas

var trainingLayers =[TreeCover, Shrubs, Agriculture, Plantation, Pasture, BuildUp, BareGround, Water];


var n = 500;
var trainingData;
var testingData;
// loop over training layers
for (var i = 0; i < trainingLayers.length; i++) { 
  // sample points within training polygons
  var pts = ee.FeatureCollection
    .randomPoints(trainingLayers[i].geometry(), n);
  
  // add class
  var thisclass = trainingLayers[i].get('class');
  pts = pts.map(function(f) {
    return f.set({class: thisclass});
});

  // extract raster cell values
  var trainTestData = compositeClip.sampleRegions(pts, ['class'], 30);

  //split the samples into training and testing
  var srsRandomized = trainTestData.randomColumn('random');
  var CUTOFF = 0.7;
  // Filter Random Sample for Visualization
  var testing = srsRandomized.filter(ee.Filter.gte('random', CUTOFF));
  var training = srsRandomized.filter(ee.Filter.lt('random', CUTOFF));

  //combine trainging regions together
if (i === 0) {
    trainingData = training;
    testingData = testing;
} else {
    trainingData = trainingData.merge(training);
    testingData = testingData.merge(testing);
}
}

//// classify with random forests
// use bands 1-8
var bands = ['B1', 'B2', 'B3', 'B4', 'B5', 'B6','B7', 'B8'];
// fit a random forests model
var classifierModel = ee.Classifier.randomForest(30)
.train(trainingData, 'class', bands);


//Classify the validation data.
var validated = testingData.classify(classifierModel);

print(validated);
//Get a confusion matrix representing expected accuracy.
var testAccuracy = validated.errorMatrix('class', 'classification');
print('Validation error matrix: ', testAccuracy);
print('Validation overall accuracy: ', testAccuracy.accuracy());
print('Validation Producer  accuracy: ', testAccuracy.producersAccuracy());
print('Validation Consumer  accuracy: ', testAccuracy.consumersAccuracy());
print('Kappa:', testAccuracy.kappa());

var table = ee.FeatureCollection([
    ee.Feature(null, {
         'error matrix': testAccuracy.array(),
         'overall accuracy': testAccuracy.accuracy(),
         'producer  accuracy': testAccuracy.producersAccuracy(),
         'consumer  accuracy':  testAccuracy.consumersAccuracy(),
         'kappa'             :  testAccuracy.kappa()
    })
]);
Export.table(table, 'exportTabletestAccuracy', {fileFormat: 'CSV'});

// produce the forest change map
var classified = compositeClip.classify(classifierModel);
var p = ['107113', '5cff08', '5d0511','fff704','f28311','777575', 'e618ff', '0446ff'];
print(classified);


// display
Map.addLayer(classified, {palette: p, min: 0, max: 7}, 'classification');
Export.image(classified,'classified',{
  scale: 100,
  maxPixles: 1e10
});