#https://numpy.org/doc/stable/user/absolute_beginners.html
#https://docs.python.org/3/tutorial/introduction.html#lists
from sklearn.datasets import fetch_olivetti_faces

from sklearn.tree import DecisionTreeClassifier
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import ExtraTreesRegressor
from sklearn.ensemble import VotingRegressor
from sklearn.ensemble import GradientBoostingRegressor
import sklearn
from sklearn.tree import export_text
from sklearn import tree
from matplotlib import image
from matplotlib import pyplot
from PIL import Image
import numpy
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import RidgeCV



import glob
 #The glob module finds all the pathnames matching a specified pattern according to the rules used by the Unix shell,
 # # although results are returned in arbitrary order.



filelist = glob.glob('coil-100/*.png')
#list comprehension for each filename in filelist, open, convert to grayscale L , flatten to 128x128=16384
data = numpy.array([numpy.array(Image.open(fname).convert('L')).flatten() for fname in filelist])
print(data.shape, "shape db")
#(7200, 128, 128, 3) shape db
#(7200, 16384) shape db

mask = numpy.random.choice([False, True], len(data), p=[0.90, 0.10])
data = data[mask] #only first 500




# load image to predice as pixel array
img = Image.open('door3.jpg').convert('L')
# summarize shape of the pixel array
imgarr = numpy.array(img) 
print(imgarr.shape)
#imgarrf = imgarr / 255 #normalize
imgarrf = imgarr
tophalflady =  imgarrf.flatten()[0: 8192]  #16384 / 2

pyplot.imshow( tophalflady.reshape((64,128)) ) #reshape to 32  rows and 64 columns
pyplot.show()


print(len(data)) #400 rows
print(len(data[0])) #4096 columns = 64 x 64 = 4096
n_pixels = data.shape[1] # (400,4096) [1] = 4096
print(n_pixels) #should be 4096 flattened
#slice the list data and target using index : meens start
# Upper half of the faces from 0  numpy arrays [rom start to middle]
X_train = data[:, : (n_pixels + 1) // 2]
# Lower half of the faces [for all from end to middle] from end to 
y_target = data[:, n_pixels // 2 :]
#clf = ExtraTreesRegressor(n_estimators=3, max_depth=48, max_features=96, random_state=0).fit(X_train,y_target)
#clf = RandomForestRegressor(n_estimators=10, max_depth=None).fit(X_train,y_target)
clf = ExtraTreesRegressor().fit(X_train,y_target)

#clf = sklearn.neighbors.KNeighborsRegressor(n_neighbors=200,weights='distance',algorithm='brute').fit(X_train,y_target)
print('to fit')
#clf = LinearRegression().fit(X_train, y_target)
#clf = RidgeCV().fit(X_train, y_target)
#clf = RandomForestRegressor(n_estimators=10).fit(X_train, y_target)
#clf = DecisionTreeRegressor(max_depth=30).fit(X_train, y_target)
print('to predict')
arrpredict= clf.predict([ tophalflady ])

# load and display an image with Matplotlib


#print(tophalflady.shape)#2048,1
tophalflady = numpy.expand_dims(tophalflady, axis=0)
#print(tophalflady.shape)#2048,1
#print(arrpredict.shape) #1,2048 linear flatten array, reshape to 2d
#arrpredict = numpy.transpose(arrpredict)
print(arrpredict.shape) #1,2048 linear flatten array, reshape to 2d
print(tophalflady.shape)#2048,1

#merge array top and predicted
topbottom = numpy.concatenate((tophalflady, arrpredict), axis=1)
print(topbottom.shape,'final shape')
# display the array of pixels as an image
#The image is quantized to 256 grey levels and stored as unsigned 8-bit integers; 
# #the loader will convert these to floating point values on the interval [0, 1],
#  which are easier to work with for many algorithms.
from numpy import asarray
#pyplot.imshow(oneimg.reshape( (64,64))) #reshape to 64 by 64
pyplot.imshow( topbottom.reshape((128,128)) ) #reshape to 32  rows and 64 columns
pyplot.show()