Commit 7922172e authored by Mitar's avatar Mitar
Browse files

Sum primitive.

parent 4f3b7066
Pipeline #15521114 passed with stage
in 3 minutes and 38 seconds
......@@ -31,6 +31,7 @@ setup(
'd3m.primitives': [
'test.MonomialPrimitive = test_primitives.monomial:MonomialPrimitive',
'test.IncrementPrimitive = test_primitives.increment:IncrementPrimitive',
'test.SumPrimitive = test_primitives.sum:SumPrimitive',
import os.path
import pickle
import typing
from http import client
import numpy # type: ignore
from d3m_metadata import container, hyperparams, metadata as metadata_module, params, utils
from primitive_interfaces import base, transformer
from . import __author__, __version__
__all__ = ('SumPrimitive',)
DOCKER_KEY = 'summing'
# It is useful to define these names, so that you can reuse it both
# for class type arguments and method signatures.
# This is just an example of how to define a more complicated input type,
# which is in fact more restrictive than what the primitve can really handle.
# One could probably just use "typing.Container" in this case, if accepting
# a wide range of input types.
Inputs = typing.Union[container.ndarray, container.List[float], container.List[container.List[float]]]
Outputs = container.List[float]
class Hyperparams(hyperparams.Hyperparams):
No hyper-parameters for this primitive.
class SumPrimitive(base.SingletonOutputMixin[Inputs, Outputs, None, Hyperparams], transformer.TransformerPrimitiveBase[Inputs, Outputs, Hyperparams]):
A primitive which sums all the values on input into one number.
# This should contain only metadata which cannot be automatically determined from the code.
metadata = metadata_module.PrimitiveMetadata({
# Simply an UUID generated once and fixed forever. Generated using "uuid.uuid4()".
'id': '9c00d42d-382d-4177-a0e7-082da88a29c8',
'version': __version__,
'name': "Sum Values",
# Keywords do not have a controlled vocabulary. Authors can put here whatever they find suitable.
'keywords': ['test primitive'],
'source': {
'name': __author__,
'uris': [
# Unstructured URIs. Link to file and link to repo in this case.
# A list of dependencies in order. These can be Python packages, system packages, or Docker images.
# Of course Python packages can also have their own dependencies, but sometimes it is necessary to
# install a Python package first to be even able to run of another package. Or you have
# a dependency which is not on PyPi.
'installation': [{
'type': metadata_module.PrimitiveInstallationType.PIP,
'package_uri': 'git+{git_commit}#subdirectory=primitives'.format(
}, {
'type': metadata_module.PrimitiveInstallationType.DOCKER,
# A key under which information about a running container will be provided to the primitive.
'key': DOCKER_KEY,
'image_name': '',
# Instead of a label, an exact hash of the image is required. This assures reproducibility.
# You can see digests using "docker images --digests".
'image_digest': 'sha256:07db5fef262c1172de5c1db5334944b2f58a679e4bb9ea6232234d71239deb64',
# The same path the primitive is registered with entry points in
'python_path': 'd3m.primitives.test.SumPrimitive',
# Choose these from a controlled vocabulary in the schema. If anything is missing which would
# best describe the primitive, make a merge request.
'algorithm_types': [
'primitive_family': metadata_module.PrimitiveFamily.OPERATOR,
# A metafeature about preconditions required for this primitive to operate well.
'preconditions': [
# Instead of strings you can also use available Python enumerations.
def __init__(self, *, hyperparams: Hyperparams, random_seed: int = 0, docker_containers: typing.Dict[str, str] = None) -> None:
super().__init__(hyperparams=hyperparams, random_seed=random_seed, docker_containers=docker_containers)
if DOCKER_KEY not in self.docker_containers:
raise ValueError("Docker key '{docker_key}' missing among provided Docker containers.".format(docker_key=DOCKER_KEY))
def produce(self, *, inputs: Inputs, timeout: float = None, iterations: int = None) -> base.CallResult[Outputs]:
# In the future, we should store here data in Arrow format into
# Plasma store and just pass an ObjectId of data over HTTP.
value = inputs.view(numpy.ndarray)
data = pickle.dumps(value)
# TODO: Retry if connection fails.
# This connection can sometimes fail because the service inside a Docker container
# is not yet ready, despite container itself already running. Primitive should retry
# a few times before aborting.
# Primitive knows the port the container is listening on.
connection = client.HTTPConnection(self.docker_containers[DOCKER_KEY], port=8000)
connection.request('POST', '/', data, {
'Content-Type': 'multipart/form-data',
response = connection.getresponse()
if response.code != 200:
raise ValueError("Invalid HTTP response code: {code}".format(code=response.code))
result = int(
outputs = container.List[float]((result,), {
'schema': metadata_module.CONTAINER_SCHEMA_VERSION,
'structural_type': container.List[float],
'dimension': {
'length': 1,
outputs.metadata = outputs.metadata.update((metadata_module.ALL_ELEMENTS,), {
'structural_type': float,
# Wrap it into default "CallResult" object: we are not doing any iterations.
return base.CallResult(outputs)
# Because numpy arrays do not contain shapes and dtype as part of their structural types,
# we have to manually check those in metadata. In this case, just dtype which is stored as
# "structural_type" on values themselves (and not the container or dimensions).
def can_accept(cls, *, method_name: str, arguments: typing.Dict[str, typing.Union[metadata_module.Metadata, type]]) -> typing.Optional[metadata_module.DataMetadata]:
output_metadata = super().can_accept(method_name=method_name, arguments=arguments)
# If structural types didn't match, don't bother.
if output_metadata is None:
return None
if 'inputs' not in arguments:
return output_metadata
inputs_metadata = arguments['inputs']
dimension_index = 0
while True:
metadata = inputs_metadata.query((metadata_module.ALL_ELEMENTS,) * dimension_index)
if 'dimension' not in metadata:
dimension_index += 1
inputs_value_structural_type = metadata.get('structural_type', None)
if inputs_value_structural_type is None:
return None
# Not a perfect way to check for a numeric type but will do for this example.
# Otherwise check out "pandas.api.types.is_numeric_dtype".
if not issubclass(inputs_value_structural_type, (float, int, numpy.number)):
return None
return output_metadata
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment