def test_get_specific_drawing():
qdg = QuickDrawDataGroup("anvil")
# get the first anvil drawing and test the values
d = qdg.get_drawing(0)
assert d.name == "anvil"
assert d.key_id == 5355190515400704
assert d.recognized == True
assert d.countrycode == "PL"
assert d.timestamp == 1488368345
# 1 stroke, 2 x,y coords, 33 points
assert len(d.image_data) == 1
assert len(d.image_data[0]) == 2
assert len(d.image_data[0][0]) == 33
assert len(d.image_data[0][1]) == 33
assert d.no_of_strokes == 1
assert len(d.strokes) == 1
assert len(d.strokes[0]) == 33
assert len(d.strokes[0][0]) == 2
assert isinstance(d.image, Image)
assert isinstance(
d.get_image(stroke_color=(10, 10, 10),
stroke_width=4,
bg_color=(200, 200, 200)), Image)
def __init__(self,
name: str,
max_drawings: int = 1000,
recognized: Optional[bool] = None,
transform: Callable[[QuickDrawing], torch.Tensor] = None):
self.ds = QuickDrawDataGroup(name,
max_drawings=max_drawings,
recognized=recognized)
if transform is None:
self.transform = lambda x: x
else:
self.transform = transform
def test_get_random_drawing():
qdg = QuickDrawDataGroup("anvil")
d = qdg.get_drawing(0)
assert d.name == "anvil"
assert isinstance(d.key_id, int)
assert isinstance(d.recognized, bool)
assert isinstance(d.timestamp, int)
assert isinstance(d.countrycode, str)
assert isinstance(d.image_data, list)
assert len(d.image_data) == d.no_of_strokes
assert isinstance(d.strokes, list)
assert len(d.strokes) == d.no_of_strokes
for stroke in d.strokes:
for point in stroke:
assert len(point) == 2
assert isinstance(d.image, Image)
assert isinstance(
d.get_image(stroke_color=(10, 10, 10),
stroke_width=4,
bg_color=(200, 200, 200)), Image)
def test_unrecognized_data():
qdg = QuickDrawDataGroup("anvil", recognized=False)
assert qdg.drawing_count == 1000
rec = 0
unrec = 0
for drawing in qdg.drawings:
if drawing.recognized:
rec += 1
else:
unrec += 1
assert rec == 0
assert unrec == qdg.drawing_count
def test_search_drawings():
qdg = QuickDrawDataGroup("anvil")
# test a search with no criteria returns 1000 results
r = qdg.search_drawings()
assert len(r) == 1000
# test a recognized search
r = qdg.search_drawings(recognized=True)
for d in r:
assert d.recognized
r = qdg.search_drawings(recognized=False)
for d in r:
assert not d.recognized
# test a country search
r = qdg.search_drawings(countrycode="US")
for d in r:
assert d.countrycode == "US"
# pull first drawing
key_id = r[0].key_id
timestamp = r[0].timestamp
# test key_id search
r = qdg.search_drawings(key_id=key_id)
for d in r:
assert d.key_id == key_id
# test timestamp search
r = qdg.search_drawings(timestamp=timestamp)
for d in r:
assert d.timestamp == timestamp
# test a compound search of recognized and country code
r = qdg.search_drawings(recognized=True, countrycode="US")
for d in r:
assert d.recognized
assert d.countrycode == "US"
def __init__(self,
recognized: Optional[bool] = None,
transform: Callable[[QuickDrawing], torch.Tensor] = None):
self.qd = QuickDrawData()
self.qd_class_names = self.qd.drawing_names
# dictionary of QuickDrawDataGroups based on all possible names, loads 1000 examples from each class, but can
# be changed by specifying max_drawings
self.qd_DataGroups = {
name: QuickDrawDataGroup(name, recognized=recognized)
for name in self.qd_class_names
}
if transform is None:
self.transform = lambda x: x
else:
self.transform = transform
class QuickDrawDataGroupDataset(Dataset):
def __init__(self,
name: str,
max_drawings: int = 1000,
recognized: Optional[bool] = None,
transform: Callable[[QuickDrawing], torch.Tensor] = None):
self.ds = QuickDrawDataGroup(name,
max_drawings=max_drawings,
recognized=recognized)
if transform is None:
self.transform = lambda x: x
else:
self.transform = transform
def __getitem__(self, index):
return self.transform(self.ds.get_drawing(index))
def __len__(self):
return self.ds.drawing_count
def main(cache_dir, save_img_dir, img_format, num_workers, categories,
drawing_recognized):
work = Queue()
# Start worker threads
for _ in range(num_workers):
w = Thread(target=worker_thread,
args=(work, save_img_dir, img_format),
daemon=True)
w.start()
# Start feeding in work
for name in categories:
group = QuickDrawDataGroup(name=name,
recognized=drawing_recognized,
cache_dir=cache_dir,
max_drawings=float('inf'))
for i, drawing in enumerate(group.drawings):
# Wait until some work gets done before continuing
while work.qsize() > 10000:
sleep(0.1)
work.put((name, i, drawing))
# Wait for work to finish
work.join()
def test_get_data_group():
qdg = QuickDrawDataGroup("anvil")
assert qdg.drawing_count == 1000
qdg = QuickDrawDataGroup("anvil", max_drawings=2000)
assert qdg.drawing_count == 2000
from quickdraw import QuickDrawData
from quickdraw import QuickDrawDataGroup
from scipy.spatial.distance import directed_hausdorff
import numpy as np
from itertools import chain
qd = QuickDrawData()
anvil = qd.get_drawing("anvil")
anvils = QuickDrawDataGroup("anvil")
#anvil = anvils.get_drawing()
#iterate through drawings
#get one drawing group
#multiple drawings within that group (for one name, ie airplane)
#get coordinates of first drawing and second drawing
#hausdorff distance to get min distance across all drawings in that group
#save distances comparing one drawing per group to the next drawing in that group
#find the minimum distance
#VERSION1
def get_drawing(qd):
for drawing in qd.drawing_names: #list of names of drawings
name = qd.get_drawing(qd.drawing_names[drawing]) #gets drawing name
for i in len(name.image_data): #loops through coords of all drawings
for j in range(1):
drawing_1 = np.array(name.image_data[i][0])
drawing_2 = np.array(name.image_data[i][1])
return (xcoords, ycoords)
def load_data_by_label(label_name, num_samples):
# print(f"Loading data for label: {label_name}")
qdg = QuickDrawDataGroup(name=label_name, recognized=True, max_drawings=num_samples)
data = torch.stack([transforms.ToTensor()(qdg.get_drawing(i).image.convert("L")) for i in range(num_samples)])
return data
from config import *
import os
from quickdraw import QuickDrawDataGroup
with open(os.path.join(DATASET_ROOT, 'categories.txt')) as f:
for name in f.readlines():
name = name.strip()
data = QuickDrawDataGroup(name,
max_drawings=SAMPLES_IN_CLASS,
recognized=True)
i = 0
for sample in data.drawings:
if i < train_samples_num:
if not os.path.isdir(
os.path.join(DATASET_ROOT, 'train/' + name)):
os.mkdir(os.path.join(DATASET_ROOT, 'train/' + name))
sample.image.save(DATASET_ROOT + '/train/' + name + '/' +
str(sample.key_id) + '.png')
i += 1
else:
if not os.path.isdir(os.path.join(DATASET_ROOT,
'test/' + name)):
os.mkdir(os.path.join(DATASET_ROOT, 'test/' + name))
sample.image.save(DATASET_ROOT + '/test/' + name + '/' +
str(sample.key_id) + '.png')
def test_drawings():
qdg = QuickDrawDataGroup("anvil")
count = 0
for drawing in qdg.drawings:
count += 1
assert count == 1000
from quickdraw import QuickDrawDataGroup
anvils = QuickDrawDataGroup("anvil")
print(anvils.drawing_count)
print(anvils.get_drawing())
from quickdraw import QuickDrawDataGroup
anvils = QuickDrawDataGroup("anvil")
results = anvils.search_drawings(countrycode="PL")
from quickdraw import QuickDrawDataGroup
from PIL import Image, ImageDraw, ImageOps
import cv2
import numpy as np
from color_it import ColorIt
qdg = QuickDrawDataGroup("circle")
results = qdg.search_drawings(key_id=int(4865367048454144))
anvil = results[0]
#anvil_image = Image.new("RGB", (255,255), color=(255,255,255))
#anvil_drawing = ImageDraw.Draw(anvil_image)
# for stroke in anvil.strokes:
# # anvil_drawing.line(stroke, fill=(0,0,0), width=2)
# for coordinate in range(len(stroke)-1):
# x1 = stroke[coordinate][0]
# y1 = stroke[coordinate][1]
# x2 = stroke[coordinate+1][0]
# y2 = stroke[coordinate+1][1]
# anvil_drawing.line((x1,y1,x2,y2), fill=(0,0,0), width=4)
#old_size = anvil_image.size # old_size[0] is in (width, height) format
#clearnew_im = ImageOps.expand(anvil_image,border = 10, fill = (0,0,0))
#new_im.show()
c = ColorIt(anvil)
c.fill(3)
from quickdraw import QuickDrawDataGroup
qdg = QuickDrawDataGroup("anvil", cache_dir="C:\\path\\to\\cache")
print(qdg.drawing_count)
print(qdg.get_drawing())
from quickdraw import QuickDrawDataGroup
anvils = QuickDrawDataGroup("anvil")
for anvil in anvils.drawings:
print(anvil)
from quickdraw import QuickDrawData
from quickdraw import QuickDrawDataGroup
import json
from trained_qd_model import *
qd = QuickDrawData()
#####
# load trained model, get arbitrary image, pass image into trained model, get weights, save weights to a file
# load model
# model = load_model('trained_quickdraw.model')
# model.summary()
arms = QuickDrawDataGroup("arm")
arm = arms.get_drawing()
# extract key id from arm
# print(arm)
#x = arms.search_drawings(key_id= int(5778229946220544))
# print the output?
# x[0]
arm.image.save('arm2.png')
im = Image.open('arm2.png')
im = im.resize((92, 92), Image.ANTIALIAS)
im.save('armcopy.png')
# print('width: %d - height: %d' % im.size)
img_width = 28
img_height = 28
from quickdraw import QuickDrawDataGroup
from quickdraw import QuickDrawData
import sys
import os
qd = QuickDrawData()
##for drawing_num, drawing_name in enumerate(qd.drawing_names):
## if drawing_num < 86:
## continue
## print("Saving " + drawing_name + " images")
## qdg = QuickDrawDataGroup(drawing_name, max_drawings=20)
## directory_name = 'images/' + drawing_name
## if not os.path.isdir(os.path.join(os.getcwd(), directory_name)):
## os.mkdir(directory_name)
## for drawing_count, drawing in enumerate(qdg.drawings, 1):
## drawing.image.save(directory_name + "/" + drawing_name + "{:0>4d}.png".format(drawing_count))
drawing_name = "face"
print("Saving " + drawing_name + " images")
qdg = QuickDrawDataGroup(drawing_name, max_drawings=300)
directory_name = 'images/' + drawing_name
if not os.path.isdir(os.path.join(os.getcwd(), directory_name)):
os.mkdir(directory_name)
for drawing_count, drawing in enumerate(qdg.drawings, 1):
drawing.image.save(directory_name + "/" + drawing_name +
"{:0>4d}.png".format(drawing_count))
print(args)
qd = QuickDrawData(recognized=None, max_drawings=1000, refresh_data=False, jit_loading=True, print_messages=False, cache_dir='./.quickdrawcache')
print(f'Total categories count {len(qd.drawing_names)}')
currentIteration = 0
def rgb2gray(rgb):
return (np.dot(rgb[...,:3], [0.298, 0.586, 0.143])/ 255).clip(0, 1)
all_drawings = []
for name in qd.drawing_names:
group = QuickDrawDataGroup(name)
for i in range(0, args.count):
data_point = group.get_drawing()
img = data_point.image.resize((32,32))
all_drawings.append(rgb2gray(np.array(img)))
#img.save(f'{args.output}/{name}_{data_point.key_id}.png')
if args.verbose:
currentIteration += 1
os.system("cls")
print(f'Images loaded [{name}]: {currentIteration}/{len(qd.drawing_names)}')
all_drawings = np.array(all_drawings)
np.save(f'{args.output}/image_set_d{len(qd.drawing_names)}_c{args.count}', all_drawings)