def __iter__(self):
filepath = download(self._url, self._cache_dir)
with tarfile.open(filepath) as archive:
for filename in archive.getnames():
if filename.startswith('aclImdb/train/pos/'):
yield self._read(archive, filename), True
elif filename.startswith('aclImdb/train/neg/'):
yield self._read(archive, filename), False
def test_reference_file_readonly(filename, md5, ref_sum, ref_num_cells):
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
tol = 1.0e-2
s = mesh.points.sum()
assert abs(s - ref_sum) < tol * ref_sum
assert {k: len(v) for k, v in mesh.cells.items()} == ref_num_cells
assert {
k: len(v["gmsh:physical"]) for k, v in mesh.cell_data.items()
} == ref_num_cells
def test_reference_file(filename, md5, ref_sum, ref_num_cells, write_binary):
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
tol = 1.0e-2
s = numpy.sum(mesh.points)
assert abs(s - ref_sum) < tol * ref_sum
assert len(mesh.cells["triangle"]) == ref_num_cells
writer = partial(meshio.vtk_io.write, write_binary=write_binary)
helpers.write_read(writer, meshio.vtk_io.read, mesh, 1.0e-15)
return
def test_reference_file(filename, md5, ref_sum, ref_num_cells, write_binary):
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
tol = 1.0e-2
s = mesh.points.sum()
assert abs(s - ref_sum) < tol * ref_sum
assert {k: len(v) for k, v in mesh.cells.items()} == ref_num_cells
assert {
k: len(v["gmsh:physical"]) for k, v in mesh.cell_data.items()
} == ref_num_cells
writer = partial(meshio.msh_io.write, fmt_version="2", write_binary=write_binary)
helpers.write_read(writer, meshio.msh_io.read, mesh, 1.0e-15)
def _read_pages(self, url):
"""
Extract plain words from a Wikipedia dump and store them to the pages
file. Each page will be a line with words separated by spaces.
"""
wikipedia_path = download(url, self._cache_dir)
with bz2.open(wikipedia_path) as wikipedia, \
bz2.open(self._pages_path, 'wt') as pages:
for _, element in etree.iterparse(wikipedia, tag='{*}page'):
if element.find('./{*}redirect') is not None:
continue
page = element.findtext('./{*}revision/{*}text')
words = self._tokenize(page)
pages.write(' '.join(words) + '\n')
element.clear()
def test_reference_file_with_mixed_cells():
filename = "med/cylinder.med"
md5 = "e36b365542c72ef470b83fc21f4dad58"
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
# Points
assert numpy.isclose(mesh.points.sum(), 16.53169892762988)
# Cells
ref_num_cells = {"pyramid": 18, "quad": 18, "line": 17, "tetra": 63, "triangle": 4}
assert {k: len(v) for k, v in mesh.cells.items()} == ref_num_cells
# Point tags
assert mesh.point_data["point_tags"].sum() == 52
ref_point_tags_info = {2: ["Side"], 3: ["Side", "Top"], 4: ["Top"]}
assert mesh.point_tags == ref_point_tags_info
# Cell tags
ref_sum_cell_tags = {
"pyramid": -116,
"quad": -75,
"line": -48,
"tetra": -24,
"triangle": -30,
}
assert {
k: v["cell_tags"].sum() for k, v in mesh.cell_data.items()
} == ref_sum_cell_tags
ref_cell_tags_info = {
-6: ["Top circle"],
-7: ["Top", "Top and down"],
-8: ["Top and down"],
-9: ["A", "B"],
-10: ["B"],
-11: ["B", "C"],
-12: ["C"],
}
assert mesh.cell_tags == ref_cell_tags_info
helpers.write_read(meshio.med_io.write, meshio.med_io.read, mesh, 1.0e-15)
def test_reference_file_with_mixed_cells():
filename = "med/cylinder.med"
md5 = "e36b365542c72ef470b83fc21f4dad58"
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
# Points
assert numpy.isclose(mesh.points.sum(), 16.53169892762988)
# Cells
ref_num_cells = {"pyramid": 18, "quad": 18, "line": 17, "tetra": 63, "triangle": 4}
assert {k: len(v) for k, v in mesh.cells.items()} == ref_num_cells
# Point tags
assert mesh.point_data["point_tags"].sum() == 52
ref_point_tags_info = {2: ["Side"], 3: ["Side", "Top"], 4: ["Top"]}
assert mesh.point_tags == ref_point_tags_info
# Cell tags
ref_sum_cell_tags = {
"pyramid": -116,
"quad": -75,
"line": -48,
"tetra": -24,
"triangle": -30,
}
assert {
k: v["cell_tags"].sum() for k, v in mesh.cell_data.items()
} == ref_sum_cell_tags
ref_cell_tags_info = {
-6: ["Top circle"],
-7: ["Top", "Top and down"],
-8: ["Top and down"],
-9: ["A", "B"],
-10: ["B"],
-11: ["B", "C"],
-12: ["C"],
}
assert mesh.cell_tags == ref_cell_tags_info
helpers.write_read(meshio.med_io.write, meshio.med_io.read, mesh, 1.0e-15)
def test_reference_file_with_point_cell_data():
filename = "med/box.med"
md5 = "0867fb11bd14b83ad11ab20e2b1fd57d"
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
# Points
assert numpy.isclose(mesh.points.sum(), 12)
# Cells
assert {k: len(v) for k, v in mesh.cells.items()} == {"hexahedron": 1}
# Point data
data_u = mesh.point_data["resu____DEPL"]
assert data_u.shape == (8, 3)
assert numpy.isclose(data_u.sum(), 12)
# Cell data
# ELNO (1 data point for every node of each element)
data_eps = mesh.cell_data["hexahedron"]["resu____EPSI_ELNO"]
assert data_eps.shape == (1, 8, 6) # (n_cells, n_nodes_per_element, n_components)
data_eps_mean = numpy.mean(data_eps, axis=1)[0]
eps_ref = numpy.array([1, 0, 0, 0.5, 0.5, 0])
assert numpy.allclose(data_eps_mean, eps_ref)
data_sig = mesh.cell_data["hexahedron"]["resu____SIEF_ELNO"]
assert data_sig.shape == (1, 8, 6) # (n_cells, n_nodes_per_element, n_components)
data_sig_mean = numpy.mean(data_sig, axis=1)[0]
sig_ref = numpy.array(
[7328.44611253, 2645.87030114, 2034.06063679, 1202.6, 569.752, 0]
)
assert numpy.allclose(data_sig_mean, sig_ref)
data_psi = mesh.cell_data["hexahedron"]["resu____ENEL_ELNO"]
assert data_psi.shape == (1, 8, 1) # (n_cells, n_nodes_per_element, n_components)
# ELEM (1 data point for each element)
data_psi_elem = mesh.cell_data["hexahedron"]["resu____ENEL_ELEM"]
assert numpy.isclose(numpy.mean(data_psi, axis=1)[0, 0], data_psi_elem[0])
helpers.write_read(meshio.med_io.write, meshio.med_io.read, mesh, 1.0e-15)
def test_reference_file_with_point_cell_data():
filename = "med/box.med"
md5 = "0867fb11bd14b83ad11ab20e2b1fd57d"
filename = helpers.download(filename, md5)
mesh = meshio.read(filename)
# Points
assert numpy.isclose(mesh.points.sum(), 12)
# Cells
assert {k: len(v) for k, v in mesh.cells.items()} == {"hexahedron": 1}
# Point data
data_u = mesh.point_data["resu____DEPL"]
assert data_u.shape == (8, 3)
assert numpy.isclose(data_u.sum(), 12)
# Cell data
# ELNO (1 data point for every node of each element)
data_eps = mesh.cell_data["hexahedron"]["resu____EPSI_ELNO"]
assert data_eps.shape == (1, 8, 6) # (n_cells, n_nodes_per_element, n_components)
data_eps_mean = numpy.mean(data_eps, axis=1)[0]
eps_ref = numpy.array([1, 0, 0, 0.5, 0.5, 0])
assert numpy.allclose(data_eps_mean, eps_ref)
data_sig = mesh.cell_data["hexahedron"]["resu____SIEF_ELNO"]
assert data_sig.shape == (1, 8, 6) # (n_cells, n_nodes_per_element, n_components)
data_sig_mean = numpy.mean(data_sig, axis=1)[0]
sig_ref = numpy.array(
[7328.44611253, 2645.87030114, 2034.06063679, 1202.6, 569.752, 0]
)
assert numpy.allclose(data_sig_mean, sig_ref)
data_psi = mesh.cell_data["hexahedron"]["resu____ENEL_ELNO"]
assert data_psi.shape == (1, 8) # (n_cells, n_nodes_per_element, ) with 1 cut off
# ELEM (1 data point for each element)
data_psi_elem = mesh.cell_data["hexahedron"]["resu____ENEL_ELEM"]
assert numpy.isclose(numpy.mean(data_psi, axis=1)[0], data_psi_elem[0])
helpers.write_read(meshio.med_io.write, meshio.med_io.read, mesh, 1.0e-15)
def getData():
r = download('https://adventofcode.com/2018/day/3/input')
data = []
regex = re.compile(r'^#(\d+) @ (\d+),(\d+): (\d+)x(\d+)$')
# The wording of the problem says the size is "at least 1000x1000"
# My input is exactly 1000x1000, but let's play it safe and work it out anyway
max_x = 0
max_y = 0
for line in [x.decode() for x in r.iter_lines()]:
linematch = regex.match(line)
_, left, top, width, height = map(int, linematch.groups())
# Process each line into a list of (x,y) tuples that its rectangle covers
coords = []
for x in range(left, left + width):
for y in range(top, top + height):
coords.append((x, y))
max_x = max(max_x, left + width)
max_y = max(max_y, top + height)
data.append(coords)
print('Grid size is {}x{}'.format(max_x, max_y))
return data, (max_x, max_y)
def getData():
r = download('https://adventofcode.com/2018/day/12/input')
iterlines = r.iter_lines()
# Prepend and append some padding to allow the 5-wide window tests
initialState = padding + [x == b'#'[0]
for x in iterlines.__next__()[15:]] + padding
# Skip the empty line in the input
_ = iterlines.__next__()
rules = []
for line in iterlines:
# Only store the rules which grow plants
if line[9] == b'#'[0]:
rule = 0
for x in line[0:5]:
rule = rule << 1
rule += x == b'#'[0]
rules.append(rule)
return np.array(initialState, dtype=np.bool), np.array(rules,
dtype=np.uint8)
def getData():
r = download('https://adventofcode.com/2018/day/7/input')
regex = re.compile(
r'^Step ([A-Z]) must be finished before step ([A-Z]) can begin\.$')
data = {}
available = []
for line in r.iter_lines():
blocker, blockee = regex.match(line.decode()).groups()
if blocker not in data:
data[blocker] = task(blocker)
available.append(data[blocker])
if blockee not in data:
data[blockee] = task(blockee)
data[blocker].blocks.add(data[blockee])
data[blockee].blockedby.add(data[blocker])
if data[blockee] in available:
available.remove(data[blockee])
# Sort the available tasks, by their name, backwards so that .pop() gets the next one
available.sort(reverse=True)
return data, available
def getData():
r = download('https://adventofcode.com/2018/day/13/input')
track = []
carts = []
for y, line in enumerate(r.iter_lines()):
track.append([])
for x, char in enumerate(line.decode()):
if char in ['^', 'v']:
track[-1].append('|')
carts.append((x, y, 0, 1 if char == '^' else -1, 0))
elif char in ['>', '<']:
track[-1].append('-')
carts.append((x, y, 1 if char == '>' else -1, 0, 0))
else:
track[-1].append(char)
carts = np.array(carts,
dtype={
'names': ['x', 'y', 'dirx', 'diry', 'rot'],
'formats': [np.int16 for _ in range(5)]
})
return np.array(track), carts
def getData():
r = download('https://adventofcode.com/2018/day/23/input')
regex = re.compile(r'^pos=<(-?\d+),(-?\d+),(-?\d+)>, r=(\d+)$')
data = np.array(list(map(lambda x: tuple(map(int, regex.match(x.decode()).groups())), r.iter_lines())), dtype={'names': ['x', 'y', 'z', 'r'], 'formats': [np.int32, np.int32, np.int32, np.uint32]})
return data
def getData():
r = download('https://adventofcode.com/2018/day/9/input')
regex = re.compile(r'^(\d+) players; last marble is worth (\d+) points$')
return tuple(map(int, regex.match(r.text.strip()).groups()))
def add_from(request, prefix_lower, prefix_normal, multi=False):
url = request.form.get(f'{prefix_lower}-art-url')
if url == None and request.get_json() != None:
url = request.get_json()[f'{prefix_lower}-art-url']
if url != '' and url != None:
if prefix_lower == 'deviantart':
art = scrapers.deviant_art(url)
elif prefix_lower == 'artstation':
if multi:
art = scrapers.art_station(url, True)
else:
art = scrapers.art_station(url)
elif prefix_lower == 'pixiv':
art = scrapers.pixiv(url, load_pickle().get('pixiv_username'), load_pickle().get('pixiv_password'))
elif prefix_lower == 'tumblr':
art = scrapers.tumblr(url)
elif prefix_lower == 'instagram':
art = scrapers.instagram(url)
elif prefix_lower == 'reddit':
art = scrapers.reddit(url)
elif prefix_lower == 'twitter':
art = scrapers.twitter(url)
title = art['title']
if multi:
images = []
for image_url in art['image_url']:
images.append(helpers.download(image_url, UPLOAD_FOLDER))
images = ','.join(images)
image = images
else:
if prefix_lower == 'pixiv':
image = helpers.download(art['image_url'], UPLOAD_FOLDER, art['source'])
else:
image = helpers.download(art['image_url'], UPLOAD_FOLDER)
source = art['source']
artist_name = art['artist_name']
artist_website = art['artist_website']
g.db = connect_db()
if request.form.get('existing-artist'):
artist_id = request.form.get('artist-id')
else:
artist = g.db.execute('SELECT id FROM artist WHERE website=?', [artist_website]).fetchone()
if artist != None:
artist_id = artist[0]
else:
cursor = g.db.execute('INSERT into artist(name, website) VALUES(?,?)', (artist_name, artist_website))
artist_id = cursor.lastrowid
cursor = g.db.execute('INSERT into art(title, image_url, artist_id, source) VALUES(?,?,?,?)', (title, image, artist_id, source))
inserted_row_id = cursor.lastrowid
g.db.commit()
g.db.close()
if helpers.request_wants_json():
return jsonify(status='success', message='Art added', id=inserted_row_id)
else:
flash('Art added', 'success')
else:
if helpers.request_wants_json():
return jsonify(status='error', message=f'{prefix_normal} Image url was empty')
else:
flash(f'{prefix_normal} Image url was empty', 'error')
return redirect('/add')
if not helpers.request_wants_json():
return redirect('/art/' + str(inserted_row_id))
def edit_art():
id = request.form.get('id')
title = request.form.get('title')
images_from_files = request.files.getlist('image-from-file')
images_from_urls = request.form.getlist('image-from-url')
existing_images = request.form.getlist('existing-image')
images = []
for local_image in request.form.getlist('local-image'):
if local_image == 'true':
file = images_from_files[0]
images_from_files.pop(0)
if file.filename == '':
flash('No file selected', 'error')
return redirect('/')
if file and helpers.allowed_file(file.filename, ALLOWED_EXTENSIONS):
filename = secure_filename(file.filename)
filename = helpers.prepend_date_time_to_string(filename)
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
image = filename
images.append(image)
else:
flash('File extension not allowed', 'error')
return redirect('/')
elif local_image == 'false':
image = images_from_urls[0]
images_from_urls.pop(0)
if image != '':
image = helpers.download(image, UPLOAD_FOLDER)
images.append(image)
else:
flash('Image url was empty', 'error')
return redirect('/')
elif local_image == 'existing':
image = existing_images[0]
existing_images.pop(0)
images.append(image)
images_string = ','.join(images)
g.db = connect_db()
if request.form.get('existing-artist') == 'true':
artist_id = request.form.get('artist-id')
else:
artist_name = request.form.get('artist-name')
artist_website = request.form.get('artist-website')
cursor = g.db.execute('INSERT into artist(name, website) VALUES(?,?)', (artist_name, artist_website))
artist_id = cursor.lastrowid
source = request.form.get('source')
images_before_edit = g.db.execute('SELECT image_url FROM art WHERE id=?', [id]).fetchone()[0]
images_before_edit = images_before_edit.split(',')
images_that_are_no_longer_in_use = list(set(images_before_edit) - set(images))
for image_that_is_no_longer_in_use in images_that_are_no_longer_in_use:
try:
os.remove(os.path.join(UPLOAD_FOLDER, image_that_is_no_longer_in_use))
except Exception as e:
# print(e)
pass
g.db.execute('UPDATE art SET title=?, image_url=?, artist_id=?, source=?, updated_at=CURRENT_TIMESTAMP WHERE id=?', (title, images_string, artist_id, source, id))
tags = request.form.getlist('tags')
if tags != []:
g.db.execute('DELETE FROM art_tag WHERE art_id=?', [id])
for tag in tags:
g.db.execute('INSERT into art_tag(art_id, tag_id) VALUES(?, ?)', (id, tag))
g.db.commit()
g.db.close()
flash('Art updated', 'success')
return redirect('/art/' + id)
def getData():
r = download('https://adventofcode.com/2018/day/18/input')
return np.array([list(x.decode()) for x in r.iter_lines()])
def getData():
r = download('https://adventofcode.com/2018/day/21/input')
iterator = r.iter_lines()
iprLine = iterator.__next__().decode().split(' ')
instructions = [[int(y) if i > 0 else y for i, y in enumerate(x.decode().split(' '))] for x in iterator]
return int(iprLine[1]), instructions
def getData():
return download('https://adventofcode.com/2018/day/15/input')
def create_alexnet():
helpers.mkdir('data')
numpy_data_path = os.path.join('data', 'bvlc_alexnet.npy')
download_url = 'https://www.dropbox.com/s/gl5wa3uzru555nd/bvlc_alexnet.npy?dl=1'
print('Downloading pre-trained AlexNet weights.')
helpers.download(download_url, numpy_data_path)
print('Weights downloaded.')
variable_data = np.load(numpy_data_path, encoding='bytes').item()
conv1_preW = variable_data["conv1"][0]
conv1_preb = variable_data["conv1"][1]
conv2_preW = variable_data["conv2"][0]
conv2_preb = variable_data["conv2"][1]
conv3_preW = variable_data["conv3"][0]
conv3_preb = variable_data["conv3"][1]
conv4_preW = variable_data["conv4"][0]
conv4_preb = variable_data["conv4"][1]
conv5_preW = variable_data["conv5"][0]
conv5_preb = variable_data["conv5"][1]
fc6_preW = variable_data["fc6"][0]
fc6_preb = variable_data["fc6"][1]
fc7_preW = variable_data["fc7"][0]
fc7_preb = variable_data["fc7"][1]
fc8_preW = variable_data["fc8"][0]
fc8_preb = variable_data["fc8"][1]
pixel_depth = 255.0
resized_height = 227
resized_width = 227
num_channels = 3
print('Creating AlexNet model.')
graph = tf.Graph()
with graph.as_default():
x = tf.placeholder(tf.uint8, [None, None, None, num_channels],
name='input')
to_float = tf.cast(x, tf.float32)
resized = tf.image.resize_images(to_float,
[resized_height, resized_width])
# Convolution 1
with tf.name_scope('conv1') as scope:
kernel = tf.Variable(conv1_preW, name='weights')
biases = tf.Variable(conv1_preb, name='biases')
conv = tf.nn.conv2d(resized, kernel, [1, 4, 4, 1], padding="SAME")
bias = tf.nn.bias_add(conv, biases)
conv1 = tf.nn.relu(bias, name=scope)
# Local response normalization 2
radius = 2
alpha = 2e-05
beta = 0.75
bias = 1.0
lrn1 = tf.nn.local_response_normalization(conv1,
depth_radius=radius,
alpha=alpha,
beta=beta,
bias=bias)
# Maxpool 1
pool1 = tf.nn.max_pool(lrn1,
ksize=[1, 3, 3, 1],
strides=[1, 2, 2, 1],
padding='VALID',
name='pool1')
# Convolution 2
with tf.name_scope('conv2') as scope:
kernel = tf.Variable(conv2_preW, name='weights')
biases = tf.Variable(conv2_preb, name='biases')
input_a, input_b = tf.split(pool1, 2, 3)
kernel_a, kernel_b = tf.split(kernel, 2, 3)
with tf.name_scope('A'):
conv_a = tf.nn.conv2d(input_a,
kernel_a, [1, 1, 1, 1],
padding="SAME")
with tf.name_scope('B'):
conv_b = tf.nn.conv2d(input_b,
kernel_b, [1, 1, 1, 1],
padding="SAME")
conv = tf.concat([conv_a, conv_b], 3)
bias = tf.nn.bias_add(conv, biases)
conv2 = tf.nn.relu(bias, name=scope)
# Local response normalization 2
radius = 2
alpha = 2e-05
beta = 0.75
bias = 1.0
lrn2 = tf.nn.local_response_normalization(conv2,
depth_radius=radius,
alpha=alpha,
beta=beta,
bias=bias)
# Maxpool 2
pool2 = tf.nn.max_pool(lrn2,
ksize=[1, 3, 3, 1],
strides=[1, 2, 2, 1],
padding='VALID',
name='pool2')
with tf.name_scope('conv3') as scope:
kernel = tf.Variable(conv3_preW, name='weights')
biases = tf.Variable(conv3_preb, name='biases')
conv = tf.nn.conv2d(pool2, kernel, [1, 1, 1, 1], padding="SAME")
bias = tf.nn.bias_add(conv, biases)
conv3 = tf.nn.relu(bias, name=scope)
with tf.name_scope('conv4') as scope:
kernel = tf.Variable(conv4_preW, name='weights')
biases = tf.Variable(conv4_preb, name='biases')
input_a, input_b = tf.split(conv3, 2, 3)
kernel_a, kernel_b = tf.split(kernel, 2, 3)
with tf.name_scope('A'):
conv_a = tf.nn.conv2d(input_a,
kernel_a, [1, 1, 1, 1],
padding="SAME")
with tf.name_scope('B'):
conv_b = tf.nn.conv2d(input_b,
kernel_b, [1, 1, 1, 1],
padding="SAME")
conv = tf.concat([conv_a, conv_b], 3)
bias = tf.nn.bias_add(conv, biases)
conv4 = tf.nn.relu(bias, name=scope)
with tf.name_scope('conv5') as scope:
kernel = tf.Variable(conv5_preW, name='weights')
biases = tf.Variable(conv5_preb, name='biases')
input_a, input_b = tf.split(conv4, 2, 3)
kernel_a, kernel_b = tf.split(kernel, 2, 3)
with tf.name_scope('A'):
conv_a = tf.nn.conv2d(input_a,
kernel_a, [1, 1, 1, 1],
padding="SAME")
with tf.name_scope('B'):
conv_b = tf.nn.conv2d(input_b,
kernel_b, [1, 1, 1, 1],
padding="SAME")
conv = tf.concat([conv_a, conv_b], 3)
bias = tf.nn.bias_add(conv, biases)
conv5 = tf.nn.relu(bias, name=scope)
# Maxpool 2
pool5 = tf.nn.max_pool(conv5,
ksize=[1, 3, 3, 1],
strides=[1, 2, 2, 1],
padding='VALID',
name='pool5')
# Fully connected 6
with tf.name_scope('fc6'):
weights = tf.Variable(fc6_preW, name='fc6_weights')
bias = tf.Variable(fc6_preb, name='fc6_bias')
shape = tf.shape(pool5)
size = shape[1] * shape[2] * shape[3]
z = tf.matmul(tf.reshape(pool5, [-1, size]), weights) + bias
fc6 = tf.nn.relu(z, name='relu')
# Fully connected 7
with tf.name_scope('fc7'):
weights = tf.Variable(fc7_preW, name='weights')
bias = tf.Variable(fc7_preb, name='bias')
z = tf.matmul(fc6, weights) + bias
fc7 = tf.nn.relu(z, name='relu')
# Fully connected 8
with tf.name_scope('fc8'):
weights = tf.Variable(fc8_preW, name='weights')
bias = tf.Variable(fc8_preb, name='bias')
fc8 = tf.matmul(fc7, weights) + bias
softmax = tf.nn.softmax(fc8)
init = tf.global_variables_initializer()
print('Model created.')
sess = tf.Session(graph=graph)
sess.run(init)
print('Exporting TensorBoard graph to tbout/alexnet')
writer = tf.summary.FileWriter('tbout/alexnet', graph=graph)
writer.close()
print('Exporting TensorFlow model to data/alexnet')
with graph.as_default():
saver = tf.train.Saver()
save_path = saver.save(sess, 'data/alexnet')
########################################################################################
### Subreddit loop and reply function
### Scan through subreddits and classifies food images
########################################################################################
for submission in subreddit.hot(limit=5):
if submission.id not in reply_log.Post_ID.values: # Check that the bot hasn't already replied to this post
if submission.stickied: # Ignore stickied posts e.g. rules
continue
# Download and save picture and thumbnail
if submission_is_image:
filename = submission.id + ".jpg"
download(submission.url, filename, pic_dest)
download(submission.thumbnail, filename, thumbnail_dest)
# Call CNN to classify and estimate calories
#classification, calories = food_CNN(pic_dest, filename)
classification, calories = (None, None)
# Reply to post and generate new log entry
text = "Hi" # TODO: learn to format
new_reply = send_reply(text, submission, classification, calories)
# Save reply in dataframe and csv file
save_reply(new_reply, reply_log, csv_file)
else:
print("Already replied to post")
def parse_detail(self, response):
category = ''
name = ''
address = ''
city = ''
phone = ''
fax = ''
email = ''
website = ''
description = ''
url = response.url or ''
image_url = ''
# check type
lis = response.css('.comp-body li')
trs = response.css('table.table.description tr')
if len(lis) > 0:
# type 1
for li in lis:
k = li.css('::text').get().strip().split(':')[0].strip()
v = li.css('::text').get().strip().split(':')[-1].strip()
if len(k) == 0:
continue
if 'Company Name' in k:
name = v
elif 'Address' in k:
address = v
elif 'Telephone' in k:
phone = li.css('a::text').get()
elif 'Fax' in k:
fax = v
elif 'Email' in k:
email = li.css('a::text').get()
# description
description = []
for p in response.css('.comp-row > p::text'):
txt = p.get().strip()
if len(txt) == 0 or 'Description' in txt:
continue
description.append(txt)
description = ' '.join(description)
# website
website = response.css('.comp-row > p > a::attr(href)').get() or ''
if self.name in website:
website = ''
# category
category = response.css('.title-comp .col-sm-10::text')[-1].get()
# image_url
image_url = response.css('.img-container img::attr(src)').get() or ''
elif len(trs) > 0:
# type 2
for tr in trs:
k = tr.css('td::text')[0].get()
v = tr.css('td::text')[-1].get()
if len(k) == 0:
continue
if 'Nama Perusahaan' in k:
name = v
elif 'Alamat' in k:
address = tr.css('td')[-1].css('p::text').get()
elif 'Kategori' in k:
category = v
elif 'Telepon' in k:
phone = tr.css('td')[-1].css('a::text').get()
elif 'Fax' in k:
fax = tr.css('td')[-1].css('a::text').get()
elif 'Email' in k:
email = tr.css('td')[-1].css('a::text').get()
# description
description = []
for p in response.css('.container > p::text'):
txt = p.get().strip()
if len(txt) == 0:
continue
description.append(txt)
description = ' '.join(description)
# website
website = response.css('a.btn.btn-contactus.btn-go-to::attr(href)').get() or ''
if self.name in website:
website = ''
# image_url
image_url = response.css('img.center-img::attr(src)').get() or ''
if email is None or len(email) == 0:
self.logger.info('{} : EMPTY EMAIL'.format(url))
email = ''
if phone is None or len(phone) == 0:
self.logger.info('{} : EMPTY PHONE'.format(url))
phone = ''
# if len(email) > 0 and len(phone) > 0:
name = helpers.fix_title(name)
slug = helpers.get_slug(name)
if image_url is not None and len(image_url) > 0:
image_url = image_url.strip()
ext = image_url.split('.')[-1]
image_name = slug
target_dir = 'images/{}/{}.{}'.format(self.name, image_name, ext)
self.logger.info('downloading image: {} => {}'.format(image_url, target_dir))
r = helpers.download(image_url, target_dir)
if not r:
self.logger.info('Failed download {} => {}'.format(image_url, target_dir))
yield {
'category': category.strip(),
'name': name.strip(),
'slug': slug.strip(),
'address': address.strip(),
'city': city.strip(),
'phone': phone.strip(),
'email': email.strip(),
'website': website.strip(),
'description': description.strip(),
'url': url.strip(),
}
def getData():
r = download('https://adventofcode.com/2018/day/25/input')
return np.array([[int(y) for y in x.decode().split(',')] for x in r.iter_lines()])
def __init__(self, cache_dir):
path = download(type(self).URL, cache_dir)
lines = self._read(path)
data, target = self._parse(lines)
self.data, self.target = self._pad(data, target)
def getData():
r = download('https://adventofcode.com/2018/day/14/input')
return int(r.text.strip())
def getData():
r = download('https://adventofcode.com/2018/day/1/input')
return list(map((lambda x: int(x)), r.iter_lines()))
def getData():
r = download('https://adventofcode.com/2018/day/5/input')
# Deliberately return the byte encoding
return np.array([x[0] for x in r.iter_content() if x is not b'\n'])
def link_crawler(start_url,
link_regex,
robots_url=None,
user_agent='statista',
max_depth=-1,
delay=3,
proxies=None,
num_retries=2,
cache=None,
scraper_callback=None):
#: Initialze a crawl queue with a seed url to start the crawl from
crawl_queue = [start_url]
#: keep track of seen urls
seen = {}
robots = {}
throttle = Throttle(delay)
#: start the crawl
while crawl_queue:
url = crawl_queue.pop()
#: robots.txt
robots_file_present = False
if 'http' not in url:
continue
#: Get the domain
domain = '{}://{}'.format(urlparse(url).scheme, urlparse(url).netloc)
#: Get the robot parser for this domain from the robots dictionary
robot_parser = robots.get(domain)
#: set a default robots url and a parser for it if there isn't one
if not robot_parser and domain not in robots:
robots_url = '{}/robots.txt'.format(domain)
robot_parser = get_robots_parser(robots_url)
if not robot_parser:
#: continue to crawl even if there are problems finding robots.txt
#: file
robots_file_present = True
# associate each domain with a corresponding parser, whether
# present or not
robots[domain] = robot_parser
elif domain in robots:
robots_file_present = True
#: crawl only when url passes robots.txt restrictions
if robots_file_present or robot_parser.can_fetch(user_agent, url):
depth = seen.get(url, 0)
if depth == max_depth:
#: Skip link if you have crawled it more than max depth
print('Skipping %s due to depth' % url)
continue
throttle.wait(url)
html = download(url, num_retries=num_retries)
if not html:
continue
if scraper_callback:
scraper_callback(url, html)
#: Get all links from page and filter only those matching given pattern
for link in get_links(html):
if re.search(link_regex, link):
if 'http' not in link:
# check if link is well formed and correct
if link.startswith('//'):
link = '{}:{}'.format(urlparse(url).scheme, link)
elif link.startswith('://'):
link = '{}{}'.format(urlparse(url).scheme, link)
else:
link = urljoin(domain, link)
if link not in seen:
seen[link] = depth + 1
crawl_queue.append(link)
else:
print('Blocked by robots.txt:', url)
def getData():
r = download('https://adventofcode.com/2018/day/2/input')
return list(r.iter_lines())
def parse_detail(self, response):
category = response.css('ol.breadcrumb.pull-left > li > a')[-1].css('::text').get() or ''
name = response.css('h1.business-title span::text').get() or ''
address = []
city = response.css('span[itemprop=addressLocality]::text').get() or ''
phone = response.css('span[itemprop=telephone]::text').get() or ''
email = ''
website = response.css('ul.dropdown-menu > li > a[itemprop=url]::attr(href)').get() or ''
description = []
url = response.url or ''
# email
try:
cfemail = response.css('span.__cf_email__::attr(data-cfemail)').get() or ''
if len(cfemail) > 0:
email = helpers.cfDecodeEmail(cfemail)
except:
email = ''
# address
address_1 = response.css('h4 > span > span::text')
address_2 = response.css('h4 > span::text')
for index, a1 in enumerate(address_1):
a1 = a1.get().strip()
a2 = address_2[index].get().strip()
address.append(a1)
address.append(a2)
address = ' '.join(address)
address = address.replace(' ,', ',')
# description
for txt in response.css('.col-sm-12 > p p'):
d = txt.css('::text').get() or ''
description.append(d.strip())
description = '. '.join(description)
description = description.replace('..', '.')
description = description.replace('. . ', '. ')
description = description.replace('. . ', '. ')
if len(email) == 0:
self.logger.info('{} : EMPTY EMAIL'.format(url))
if len(phone) == 0:
self.logger.info('{} : EMPTY PHONE'.format(url))
if len(email) > 0 and len(phone) > 0:
image_url = response.css('.detail-listing-img > img::attr(src)').get()
if image_url is not None and image_url[-1] != '/':
image_url = image_url.strip()
ext = image_url.split('.')[-1]
image_name = helpers.get_slug(helpers.fix_title(name))
target_dir = 'images/{}/{}'.format(self.name, image_name)
self.logger.info('downloading image: {} => {}'.format(image_url, target_dir))
helpers.download(image_url, target_dir)
yield {
'category': category.strip(),
'name': name.strip(),
'address': address.strip(),
'city': city.strip(),
'phone': phone.strip(),
'email': email.strip(),
'website': website.strip(),
'description': description.strip(),
'url': url.strip(),
}
def getData():
r = download('https://adventofcode.com/2018/day/4/input')
# Make an array for the year 1518 (which is not a leap year)
initialData = [None] * 12
for i, v in enumerate([31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]):
initialData[i] = [None] * v
regex = re.compile(r'^\[(\d+)-(\d+)-(\d+) (\d+):(\d+)\] (?:(w)akes up|(f)alls asleep|Guard #(\d+) begins shift)$')
guardSeen = []
guardMap = []
guardMaxDays = 0
for line in [x.decode() for x in r.iter_lines()]:
linematch = regex.match(line)
# Two of the last 3 groups (w,f,g) will return None since they weren't matched
_, m, d, H, M, w, f, g = [(x if not x or x in ['w', 'f'] else int(x)) for x in linematch.groups()]
# Handle a special case where guards start just before midnight
if H == 23:
# Roll over a month if we're on the last day, else add a day
if d > len(initialData[m]):
m += 1
d = 1
else:
d += 1
# Days and months are 1-indexed, so decrement one now to get the index
m -= 1
d -= 1
# Initialise this entry in the array if it hasn't been seen yet
if not initialData[m][d]:
initialData[m][d] = {'g': 0, 'w': [], 'f': []}
if w or f: # This line is a wake or fall asleep action
action = w if w else f
# Store the minutes at which this action happened on this day in reverse order -
initialData[m][d][action] = sorted(initialData[m][d][action][:] + [M], reverse=True)
else: # Then this is a guard shift start
# If we haven't seen this guard before, add them to the guardMap
# Maintaining a map like this rather than indexing on the raw id compacts the dataset a lot
# For the input I had, this reduces it from (3468, 19, 60) to only (22, 19, 60)
if g not in guardMap:
guardMapIndex = len(guardMap)
guardMap.append(g)
else:
guardMapIndex = guardMap.index(g)
# Set the guardMap index for this date
initialData[m][d]['g'] = guardMapIndex
# Keep track of the number of times we've seen each guard
if guardMapIndex < len(guardSeen):
guardSeen[guardMapIndex] += 1
else:
guardSeen.append(1)
# Keep track of the most days we've seen any particular guard
guardMaxDays = max(guardMaxDays, guardSeen[guardMapIndex])
# Build the template for the numpy array, in shape (guard count, guardMaxDays, 60)
# Add one to guardMaxId to avoid 0-index issues
template = [None] * len(guardMap)
for i in range(len(guardMap)):
template[i] = [None] * guardMaxDays
for j in range(guardMaxDays):
template[i][j] = [False] * 60
data = np.array(template)
# Loop over the intial data again to convert it into something a little more useful
for month in initialData:
for day in month:
# We don't have data for some days
if day:
g = day['g']
# We use the seen counter to track where we insert this day's data
# Decrement comes first because of 0-indexing
guardSeen[g] -= 1
while len(day['f']) > 0:
# Pop the earliest pair of falling/waking off the end of the lists
# This works because we sorted in reverse order before
f = day['f'].pop()
w = day['w'].pop()
# Set each minute in the interval to True for this (guard id, seen count) pair
data[g][guardSeen[g]][f:w] = [True for _ in range(f, w)]
return data, guardMap