def test_skip_errors_basic(self):
''' test the basic usage of skip_errors '''
self.assertRaises(ValueError,
lambda: G(range(5)).map(_double_only_evens).to(list))
self.assertEqual(
G(range(5)).map(_double_only_evens).skip_errors().to(list),
[0, 4, 8])
def test_switch_with_default_1(self):
self.assertEqual(
G(
range(10)
).switch(
lambda i: i%5,
{
0: lambda i: i+1,
4: lambda i: float(i)
},
lambda i: None
).to(list),
[
1,
None,
None,
None,
4.0,
6,
None,
None,
None,
9.0
]
)
def test_switch_with_default_2(self):
self.assertEqual(
G(
range(10)
).switch(
lambda i: i%5,
{
0: lambda i: i+1,
4: lambda i: float(i)
},
range
).to(list),
[
1,
range(0, 1),
range(0, 2),
range(0, 3),
4.0,
6,
range(0, 6),
range(0, 7),
range(0, 8),
9.0
]
)
def tail(file_path: str) -> Iterable[str]:
assert os.path.isfile(file_path)
notifier = inotify.adapters.Inotify()
notifier.add_watch(file_path)
with open(file_path, 'r') as f:
notifier.add_watch(file_path)
yield from G( # create a Generator fed by the notifier
notifier.event_gen(yield_nones=False)
).filter( # filter for IN_MODIFY events (mask equals 2)
lambda i: i[0].mask == 2
).map( # when the file is modified, get the new size
lambda i: os.path.getsize(i[2])
).uniq( # filter duplicates, just incase nothing was added to the file
).window( # window the (previous_size, current_size)
2
).side_task( # seek the file descriptor and pass the input since f.seek returns None
lambda i: f.seek(i[0])
).map( # read in the newly added data
lambda i: f.read(i[1] - i[0])
).chain( # chain the incoming chunks since they might not be single lines
).groupby( # seperate groups by lines
lambda i: i == '\n'
).filter( # exclude groups that are just '\n', since they are the delimiters
lambda i: i[0] == False
).map( # join the characters to construct each line as a string
lambda i: ''.join(i[1])
#).print('-', use_repr=True # uncomment this line to see the constructed lines
)
def test_insert_relation(self):
db = self.db
#for i in range(20000):
# db.store_item(i)
report(
'relation insertion',
rps(
G(count()).map(
lambda i: db.store_relation(i, 'less_than', i + 1))))
def show_generator_output(sess, n_img, inp_z, out_channel, img_mode='RGB'):
# Show the output from the generator
cmap = None if img_mode == 'RGB' else 'gray'
z_dim = inp_z.get_shape().as_list()[-1]
example_z = np.random.uniform(-1, 1, size=[n_img, z_dim])
samples = sess.run(G(inp_z, out_channel, False),
feed_dict={inp_z: example_z})
images_grid = helper.images_square_grid(samples, image_mode)
return images_grid
def test_switch_basic_usage_3(self):
self.assertEqual(
G(
'hello world or something'.split(' ')
).switch(
lambda i: i[:2],
{
'wo': lambda i: i+i,
'or': lambda i: i[0],
'he': lambda i: i.upper()
}
).to(list),
['HELLO', 'worldworld', 'o', 'something']
)
def test_switch_benchmark(self):
'''ensure switch can process 1,000,000 items a second minimum'''
from itertools import count
speed = self.assertGreater(
G(
count()
).switch(
lambda i: i%3,
{
0: lambda i: [i],
2: lambda i: float(i)
}
).benchmark(),
100000
)
def scrape_urls(target):
return G( # create a generator that iterates over the target url's content
get(target).iter_content()
).filter( # clean the incoming data to just printables
lambda char: char in printable_bytes
).groupby( # create groups by seperating the content by its quotes
lambda char: char in {b'"', b"'"}
).map( # join the bytes between quotes and encode them to strings
lambda flag_group: b''.join(flag_group[1]).decode()
).filter( # filter for relative and non-relative urls
lambda s: (s.startswith('/') or '://' in s) and ' ' not in s
).map( # add target to the beginning of relative urls
lambda url: '/'.join(target.split('/')[:3]) + url if url.startswith('/') else url
# ).print('found:' # uncomment this line to see all urls the script is finding before deduplication
).to(set)
def model_loss(inp_real, inp_z, out_channel, alpha=0.2, smooth_factor=0.1):
# Loss from the real image for G and D
d_model_real, d_logits_real = D(inp_real, alpha=alpha)
d_loss_real = tf.reduce_mean(
tf.nn.sigmoid_cross_entropy_with_logits(
logits=d_logits_real,
labels=tf.ones_like(d_model_real) * (1 - smooth_factor)))
# Loss from the fake image for G and D
inp_fake = G(inp_z, out_channel, alpha=alpha)
d_model_fake, d_logits_fake = D(inp_fake, reuse=True, alpha=alpha)
d_loss_fake = tf.reduce_mean(
tf.nn.sigmoid_cross_entropy_with_logits(
logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
g_loss = tf.reduce_mean(
tf.nn.sigmoid_cross_entropy_with_logits(
logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))
return d_loss_real + d_loss_fake, g_loss
def test_switch_basic_usage_2(self):
self.assertEqual(
G(
range(10)
).switch(
lambda i: i%5,
{
0: lambda i: i+1,
4: lambda i: float(i)
}
).to(list),
[
1,
1,
2,
3,
4.0,
6,
6,
7,
8,
9.0
]
)
def test_switch_basic_usage_1(self):
self.assertEqual(
G(
range(10)
).switch(
lambda i: i%3,
{
0: lambda i: [i],
2: lambda i: float(i)
}
).to(list),
[
[0],
1,
2.0,
[3],
4,
5.0,
[6],
7,
8.0,
[9]
]
)
def test_map_parallel_basic(self):
''' test the basic usage of map_parallel '''
self.assertEqual(
G(range(5)).map_parallel(float, 2).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 3).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 4).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 5).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 6).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 7).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
self.assertEqual(
G(range(5)).map_parallel(float, 8).to(list),
[0.0, 1.0, 2.0, 3.0, 4.0])
def test_chunks_3_dimensional_g_usage(self):
self.assertEqual(
G(range(16)).chunk(2, 2, 2).to(list), [
(((0, 1), (2, 3)), ((4, 5), (6, 7))),
(((8, 9), (10, 11)), ((12, 13), (14, 15))),
])
def test_chunks_2_dimensional_g_usage(self):
self.assertEqual(
G(range(18)).chunk(3, 3).to(list),
[((0, 1, 2), (3, 4, 5), (6, 7, 8)),
((9, 10, 11), (12, 13, 14), (15, 16, 17))])
def test_chunk_on_speed(self):
self.assertGreater(
G(count()).chunk_on(lambda i: i % 10 == 0).benchmark(), 100_000)
def test_chunk_on_g_usage(self):
self.assertEqual(
G(range(32)).chunk_on(lambda i: str(i).endswith('0')).to(list),
[(0, 1, 2, 3, 4, 5, 6, 7, 8, 9),
(10, 11, 12, 13, 14, 15, 16, 17, 18, 19),
(20, 21, 22, 23, 24, 25, 26, 27, 28, 29), (30, 31)])
def test_ifmap_basic_g_usage_1(self):
''' test the basic usage of ifmap with G invocation '''
self.assertEqual(
G(range(5)).ifmap(bool, str).to(list), [0, "1", "2", "3", "4"])
def test_ifmap_basic_g_usage_2(self):
''' test the basic usage of ifmap with G invocation '''
self.assertEqual(
G(range(5)).ifmap(lambda i: i % 2, double).to(list),
[0, "1", "2", "3", "4"])
def test_ifmap_benchmark_1(self):
''' test the basic usage of ifmap with G invocation '''
self.assertGreater(G(count).ifmap(lambda i: i % 2, double), 1_000_000)
def test_serialization_deserialization(self):
db = self.db
report('serialization/deserialization',
rps(G(count()).map(lambda i: db.deserialize(db.serialize(i)))))
def test_skip_errors_with_log_false(self):
''' test skip_errors with log specifically enabled '''
self.assertEqual(
G(range(5)).map(_double_only_evens).skip_errors(
log=False).to(list), [0, 4, 8])
def valid_ip(s: str) -> bool:
try:
ip_address(s)
return True
except:
return False
host_dns = G( # iterate over each line on /etc/hosts
open('/etc/hosts', 'r')
).map( # strip trailing whitespace
str.strip
).filter( # remove empty lines
bool
).map( # replace tabs with spaces, split remaining lines by spaces and filter empty strings
lambda line:
[field for field in line.replace('\t', ' ').split(' ') if field]
).filter( # only lines that start with a valid ip address and have at least one following hostname
lambda fields: valid_ip(fields[0]) and len(fields) > 1
).map( # map every hostname on the line to the ip address
lambda fields: [[hostname, fields[0]] for hostname in fields[1:]]
).chain( # chain the lists so you have one long k,v stream
#).print( # uncomment this line to see the k,v pairs be materialized
).to(dict) # feed the k,v stream to dict to materialize the dictionary
if __name__ == '__main__':
print(host_dns)
def test_ifmap_benchmark_0(self):
''' test the basic usage of ifmap with G invocation '''
self.assertGreater(G(count).ifmap(bool, str), 1_000_000)
def test_insert_item(self):
db = self.db
report('object insertion',
rps(G(count()).map(lambda i: db.store_item(i))))
