def main():
mf = "./docs/examples/geometry/geometry.map"
mapfile = mappyfile.open(mf)
mapfile["size"] = [600, 600]
output_folder = os.path.join(os.getcwd(), "docs/images")
dilated = dilation(mapfile)
create_image("dilated", mapfile, output_folder=output_folder)
erosion(mapfile, dilated)
create_image("erosion", mapfile, output_folder=output_folder)
def create_frame(mapfile, line, dist):
# get the polygon layer
pl = mappyfile.find(mapfile["layers"], "name", "polygon")
# buffer the line
dilated = line.buffer(dist, cap_style=3)
# now set the FEATURES in the Mapfile to be the WKT from Shapely
pl["features"][0]["wkt"] = "'%s'" % dilated.wkt
# create an image from this Mapfile
return create_image("animation_%s" % str(dist), mapfile, format="gif")
def create_frame(mapfile, line, dist):
# get the polygon layer
pl = mappyfile.find(mapfile["layers"], "name", "polygon")
# buffer the line
dilated = line.buffer(dist, cap_style=3)
# now set the FEATURES in the Mapfile to be the WKT from Shapely
pl["features"][0]["wkt"] = "'%s'" % dilated.wkt
# create an image from this Mapfile
return create_image("animation_%s" % str(dist), mapfile, format="gif")
def train(self, test_data=None, epochs=30):
captions = []
with open("dataset_captions/captions.txt") as fp:
for line in fp:
caption_text = [
char_to_index(character) for character in line
] #Convert to indices
#Convert to encoding
caption_text = to_categorical(caption_text,
num_classes=VOCAB_SIZE)
captions.append(caption_text)
images = []
#Prepare the test data and train on batch size = 1
for i in range(0, self.N):
rgb_img = image_array("dataset/test_{}.jpg".format(i))
images.append(rgb_img)
images = np.array(images)
captions = np.array(captions)
captions = pad_sequences(captions, maxlen=42)
#Train here
self.model.fit(captions, images, epochs=epochs, batch_size=4)
test_data = to_categorical([
char_to_index(character)
for character in "Red lights in a faraway city."
],
num_classes=VOCAB_SIZE)
test_data = np.array([test_data])
test_data = pad_sequences(test_data, maxlen=42)
#Predict a result
prediction = self.model.predict(test_data, batch_size=1)
create_image(prediction[0])