from deeplearning.ml4pl.testing import random_networkx_generator
from deeplearning.ml4pl.testing import testing_databases
from labm8.py import decorators
from labm8.py import test
FLAGS = test.FLAGS
###############################################################################
# Fixtures.
###############################################################################
@test.Fixture(
scope="session",
params=testing_databases.GetDatabaseUrls(),
namer=testing_databases.DatabaseUrlNamer("graph_db"),
)
def empty_db(request) -> graph_tuple_database.Database:
"""A test fixture which yields an empty graph proto database."""
yield from testing_databases.YieldDatabase(graph_tuple_database.Database,
request.param)
@test.Fixture(
scope="function",
params=testing_databases.GetDatabaseUrls(),
namer=testing_databases.DatabaseUrlNamer("graph_db"),
)
def db_session(request) -> graph_tuple_database.Database.SessionType:
"""A test fixture which yields an empty graph proto database session."""
with testing_databases.DatabaseContext(graph_tuple_database.Database,
from labm8.py import test
FLAGS = test.FLAGS
def CreateRandomString(min_length: int = 1, max_length: int = 1024) -> str:
"""Generate a random string."""
return "".join(
random.choice(string.ascii_lowercase)
for _ in range(random.randint(min_length, max_length)))
@test.Fixture(
scope="session",
params=testing_databases.GetDatabaseUrls(),
namer=testing_databases.DatabaseUrlNamer("ir_db"),
)
def ir_db(request) -> ir_database.Database:
"""A test fixture which yields an IR database."""
with testing_databases.DatabaseContext(ir_database.Database,
request.param) as db:
rows = []
for i in range(250):
ir = ir_database.IntermediateRepresentation.CreateFromText(
source=random.choice([
"pact17_opencl_devmap",
"poj-104:train",
"poj-104:val",
"poj-104:test",
]),
relpath=str(i),
assert graph_tuple.graph_x_dimensionality == graph_x_dimensionality
assert graph_tuple.graph_y_dimensionality == graph_y_dimensionality
if with_data_flow:
assert graph_tuple.data_flow_steps >= 1
assert graph_tuple.data_flow_root_node >= 0
assert graph_tuple.data_flow_positive_node_count >= 1
else:
assert graph_tuple.data_flow_steps is None
assert graph_tuple.data_flow_root_node is None
assert graph_tuple.data_flow_positive_node_count is None
@test.Fixture(
scope="function",
params=testing_databases.GetDatabaseUrls(),
namer=testing_databases.DatabaseUrlNamer("db"),
)
def db(request) -> graph_tuple_database.Database:
"""A test fixture which yields an empty graph proto database."""
yield from testing_databases.YieldDatabase(
graph_tuple_database.Database, request.param
)
@test.Fixture(scope="function", params=(1, 100))
def graph_count(request) -> int:
"""Test fixture to enumerate graph counts."""
return request.param
@test.Fixture(scope="function", params=(1, 3))
FLAGS = test.FLAGS
# For testing models, always use --strict_graph_segmentation.
FLAGS.strict_graph_segmentation = True
###############################################################################
# Fixtures.
###############################################################################
@test.Fixture(
scope="session",
params=testing_databases.GetDatabaseUrls(),
namer=testing_databases.DatabaseUrlNamer("log_db"),
)
def log_db(request) -> log_database.Database:
"""A test fixture which yields an empty log database."""
yield from testing_databases.YieldDatabase(
log_database.Database, request.param
)
@test.Fixture(scope="session")
def logger(log_db: log_database.Database) -> logging.Logger:
"""A test fixture which yields a logger."""
with logging.Logger(log_db, max_buffer_length=128) as logger:
yield logger
