def _create_header(self, columns: FieldLike) -> None:
# Check types of fieldlike column descriptors and convert them to field
# descriptors, that are accepted by dataclasses.make_dataclass()
fields: list = []
for each in columns:
if isinstance(each, str):
fields.append(each)
continue
check.has_type(f"field {each}", each, tuple)
check.has_size(f"field {each}", each, min_size=2, max_size=3)
check.has_type("first arg", each[0], str)
check.has_type("second arg", each[1], type)
if len(each) == 2:
fields.append(each)
continue
check.has_type("third arg", each[2], (Field, dict))
if isinstance(each[2], Field):
fields.append(each)
continue
field = dataclasses.field(**each[2])
fields.append(each[:2] + (field,))
# Create record namespace with table hooks
namespace = {
'_create_row_id': self._create_row_id,
'_delete_hook': self._remove_row_id,
'_restore_hook': self._append_row_id,
'_update_hook': self._update_row_diff,
'_revoke_hook': self._remove_row_diff}
# Create Record dataclass and constructor
self._Record = dataclasses.make_dataclass(
'Row', fields, bases=(Record, ), namespace=namespace)
# Create slots
self._Record.__slots__ = ['id', 'state'] + [
field.name for field in dataclasses.fields(self._Record)]
# Reset store, diff and index
self._store = []
self._diff = []
self._index = []
class PaginationAdvisory(ABC):
message: str = field(init=False)
class DataTrainingArguments:
"""
Arguments pertaining to what data we are going to input our model for training and eval.
"""
dataset_name: Optional[str] = field(
default=None,
metadata={
"help":
"The name of the dataset to use (via the datasets library)."
})
dataset_config_name: Optional[str] = field(
default=None,
metadata={
"help":
"The configuration name of the dataset to use (via the datasets library)."
})
train_file: Optional[str] = field(
default=None,
metadata={"help": "The input training data file (a text file)."})
validation_file: Optional[str] = field(
default=None,
metadata={
"help":
"An optional input evaluation data file to evaluate the perplexity on (a text file)."
},
)
overwrite_cache: bool = field(
default=False,
metadata={"help": "Overwrite the cached training and evaluation sets"})
validation_split_percentage: Optional[int] = field(
default=5,
metadata={
"help":
"The percentage of the train set used as validation set in case there's no validation split"
},
)
max_seq_length: Optional[int] = field(
default=None,
metadata={
"help":
("The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated.")
},
)
preprocessing_num_workers: Optional[int] = field(
default=None,
metadata={
"help": "The number of processes to use for the preprocessing."
},
)
mlm_probability: float = field(
default=0.15,
metadata={
"help": "Ratio of tokens to mask for masked language modeling loss"
})
line_by_line: bool = field(
default=False,
metadata={
"help":
"Whether distinct lines of text in the dataset are to be handled as distinct sequences."
},
)
pad_to_max_length: bool = field(
default=False,
metadata={
"help":
("Whether to pad all samples to `max_seq_length`. "
"If False, will pad the samples dynamically when batching to the maximum length in the batch."
)
},
)
max_train_samples: Optional[int] = field(
default=None,
metadata={
"help":
("For debugging purposes or quicker training, truncate the number of training examples to this "
"value if set.")
},
)
max_eval_samples: Optional[int] = field(
default=None,
metadata={
"help":
("For debugging purposes or quicker training, truncate the number of evaluation examples to this "
"value if set.")
},
)
def __post_init__(self):
if self.dataset_name is None and self.train_file is None and self.validation_file is None:
raise ValueError(
"Need either a dataset name or a training/validation file.")
else:
if self.train_file is not None:
extension = self.train_file.split(".")[-1]
assert extension in [
"csv", "json", "txt"
], "`train_file` should be a csv, a json or a txt file."
if self.validation_file is not None:
extension = self.validation_file.split(".")[-1]
assert extension in [
"csv", "json", "txt"
], "`validation_file` should be a csv, a json or a txt file."
class DataTrainingArguments:
"""
Arguments pertaining to what data we are going to input our model for training and eval.
"""
dataset_name: Optional[str] = field(
default=None, metadata={"help": "The name of the dataset to use (via the datasets library)."}
)
dataset_config_name: Optional[str] = field(
default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
)
train_file: Optional[str] = field(default=None, metadata={"help": "The input training data file (a text file)."})
validation_file: Optional[str] = field(
default=None,
metadata={"help": "An optional input evaluation data file to evaluate the perplexity on (a text file)."},
)
test_file: Optional[str] = field(
default=None,
metadata={"help": "An optional input test data file to test the perplexity on (a text file)."},
)
overwrite_cache: bool = field(
default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
)
preprocessing_num_workers: Optional[int] = field(
default=None,
metadata={"help": "The number of processes to use for the preprocessing."},
)
max_seq_length: int = field(
default=384,
metadata={
"help": "The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
},
)
pad_to_max_length: bool = field(
default=True,
metadata={
"help": "Whether to pad all samples to `max_seq_length`. "
"If False, will pad the samples dynamically when batching to the maximum length in the batch (which can "
"be faster on GPU but will be slower on TPU)."
},
)
max_train_samples: Optional[int] = field(
default=None,
metadata={
"help": "For debugging purposes or quicker training, truncate the number of training examples to this "
"value if set."
},
)
max_val_samples: Optional[int] = field(
default=None,
metadata={
"help": "For debugging purposes or quicker training, truncate the number of validation examples to this "
"value if set."
},
)
max_test_samples: Optional[int] = field(
default=None,
metadata={
"help": "For debugging purposes or quicker training, truncate the number of test examples to this "
"value if set."
},
)
version_2_with_negative: bool = field(
default=False, metadata={"help": "If true, some of the examples do not have an answer."}
)
null_score_diff_threshold: float = field(
default=0.0,
metadata={
"help": "The threshold used to select the null answer: if the best answer has a score that is less than "
"the score of the null answer minus this threshold, the null answer is selected for this example. "
"Only useful when `version_2_with_negative=True`."
},
)
doc_stride: int = field(
default=128,
metadata={"help": "When splitting up a long document into chunks, how much stride to take between chunks."},
)
n_best_size: int = field(
default=20,
metadata={"help": "The total number of n-best predictions to generate when looking for an answer."},
)
max_answer_length: int = field(
default=30,
metadata={
"help": "The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another."
},
)
def __post_init__(self):
if (
self.dataset_name is None
and self.train_file is None
and self.validation_file is None
and self.test_file is None
):
raise ValueError("Need either a dataset name or a training/validation/test file.")
else:
if self.train_file is not None:
extension = self.train_file.split(".")[-1]
assert extension in ["csv", "json"], "`train_file` should be a csv or a json file."
if self.validation_file is not None:
extension = self.validation_file.split(".")[-1]
assert extension in ["csv", "json"], "`validation_file` should be a csv or a json file."
if self.test_file is not None:
extension = self.test_file.split(".")[-1]
assert extension in ["csv", "json"], "`test_file` should be a csv or a json file."
class Parameter:
"""
{begin_markdown Parameter}
{spell_markdown param init}
# `curvefit.core.parameter.Parameter`
## A parameter for the functional form of the curve
A `Parameter` is a parameter of the functional form for the curve. For example, if the parametric curve you want
to fit has three parameters then you need 3 `Parameter` objects. A `Parameter` is made up of one or more
[`Variable`](Variable.md) objects, which represent the fixed and random effects. Parameters may have
link functions that transform the parameter into some other
space to enforce, for example, positivity of the parameter (e.g. the parameter representing the scale of a
Gaussian distribution can't be negative).
## Arguments
- `param_name (str)`: name of parameter, e.g. 'alpha'
- `link_fun (Callable)`: the link function for the parameter
- `variables (List[curvefit.core.parameter.Variable])`: a list of `Variable` instances
## Attributes
All attributes from the `Variable`s in the list in the `variables` argument are carried over to
`Parameter` but they are put into a list. For example, the `fe_init` attribute for `Parameter` is a list of
`fe_init` attributes for each `Variable` in the order that they were passed in `variables` list.
*Additional* attributes that are not lists of the individual `Variable` attributes are listed below.
- `self.num_fe (int)`: total number of effects for the parameter (number of variables)
## Usage
```python
from curvefit.core.parameter import Parameter, Variable
var = Variable(covariate='ones', var_link_fun=lambda x: x, fe_init=0., re_init=0.)
param = Parameter(param_name='alpha', link_fun=lambda x: x, variables=[var])
```
{end_markdown Parameter}
"""
param_name: str
link_fun: Callable
variables: InitVar[List[Variable]]
num_fe: int = field(init=False)
covariate: List[str] = field(init=False)
var_link_fun: List[Callable] = field(init=False)
fe_init: List[float] = field(init=False)
re_init: List[float] = field(init=False)
re_zero_sum_std: List[float] = field(init=False)
fe_gprior: List[List[float]] = field(init=False)
re_gprior: List[List[float]] = field(init=False)
fe_bounds: List[List[float]] = field(init=False)
re_bounds: List[List[float]] = field(init=False)
def __post_init__(self, variables):
assert isinstance(variables, list)
assert len(variables) > 0
assert isinstance(variables[0], Variable)
self.num_fe = len(variables)
for k, v in consolidate(Variable, variables).items():
self.__setattr__(k, v)
class CoctMt090003Uv01Device:
class Meta:
name = "COCT_MT090003UV01.Device"
realm_code: List[Cs] = field(
default_factory=list,
metadata={
"name": "realmCode",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
type_id: Optional[Ii] = field(
default=None,
metadata={
"name": "typeId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
template_id: List[Ii] = field(
default_factory=list,
metadata={
"name": "templateId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
id: List[Ii] = field(
default_factory=list,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
manufacturer_model_name: Optional[Sc] = field(
default=None,
metadata={
"name": "manufacturerModelName",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
software_name: Optional[Sc] = field(
default=None,
metadata={
"name": "softwareName",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
null_flavor: Optional[NullFlavor] = field(
default=None,
metadata={
"name": "nullFlavor",
"type": "Attribute",
}
)
class_code: Optional[EntityClassDevice] = field(
default=None,
metadata={
"name": "classCode",
"type": "Attribute",
"required": True,
}
)
determiner_code: EntityDeterminer = field(
init=False,
default=EntityDeterminer.INSTANCE,
metadata={
"name": "determinerCode",
"type": "Attribute",
"required": True,
}
)
class BaseConfig(JsonSchemaMixin, allow_additional_props=False):
"""Taskcat configuration file"""
general: GeneralConfig = field(default_factory=GeneralConfig)
project: ProjectConfig = field(default_factory=ProjectConfig)
tests: Dict[TestName, TestConfig] = field(default_factory=dict)
# pylint doesn't like instance variables being added in post_init
# pylint: disable=attribute-defined-outside-init
def __post_init__(self):
self._source: Dict[str, Any] = {}
self._propogate()
self.set_source("UNKNOWN")
self._propogate_source()
@staticmethod
def _merge(source, dest):
for section_key, section_value in source.items():
if section_key in PROPAGATE_KEYS + PROPOGATE_ITEMS:
if section_key not in dest:
dest[section_key] = section_value
continue
if section_key in PROPAGATE_KEYS:
for key, value in section_value.items():
dest[section_key][key] = value
return dest
def _propogate(self):
project_dict = self._merge(self.general.to_dict(),
self.project.to_dict())
self.project = ProjectConfig.from_dict(project_dict)
for test_key, test in self.tests.items():
test_dict = self._merge(self.project.to_dict(), test.to_dict())
self.tests[test_key] = TestConfig.from_dict(test_dict)
def _propogate_source(self):
self._source["project"] = self._merge(self._source["general"],
self._source["project"])
for test_key in self._source["tests"]:
test = self._merge(self._source["project"],
self._source["tests"][test_key])
self._source["tests"][test_key] = test
def set_source(self,
source_name: str,
dest: Optional[Any] = None) -> Optional[Union[str, dict]]:
base_case = False
if dest is None:
base_case = True
self._source = self.to_dict()
dest = self._source
if not isinstance(dest, dict):
return source_name
if isinstance(dest, dict):
for item in dest:
dest[item] = self.set_source(source_name, dest[item])
if not base_case:
return dest
return None
@classmethod
def merge(cls, base_config: "BaseConfig",
merge_config: "BaseConfig") -> "BaseConfig":
merged = base_config.to_dict()
merge_nested_dict(merged, merge_config.to_dict())
merged_source = base_config._source.copy()
merge_nested_dict(merged_source, merge_config._source)
config = cls.from_dict(merged)
config._source = merged_source
config._propogate_source() # pylint: disable=protected-access
return config
class BswOsTaskExecutionEvent:
"""This BswEvent is supposed to execute BswSchedulableEntitys which have to
react on the execution of specific OsTasks.
Therefore, this event is unconditionally raised whenever the OsTask
on which it is mapped is executed. The main use case for this event
is scheduling of Runnables of Complex Drivers which have to react on
task executions.
:ivar short_name: This specifies an identifying shortName for the
object. It needs to be unique within its context and is intended
for humans but even more for technical reference.
:ivar short_name_fragments: This specifies how the
Referrable.shortName is composed of several shortNameFragments.
:ivar long_name: This specifies the long name of the object. Long
name is targeted to human readers and acts like a headline.
:ivar desc: This represents a general but brief (one paragraph)
description what the object in question is about. It is only one
paragraph! Desc is intended to be collected into overview
tables. This property helps a human reader to identify the
object in question. More elaborate documentation, (in particular
how the object is built or used) should go to "introduction".
:ivar category: The category is a keyword that specializes the
semantics of the Identifiable. It affects the expected existence
of attributes and the applicability of constraints.
:ivar admin_data: This represents the administrative data for the
identifiable object.
:ivar introduction: This represents more information about how the
object in question is built or is used. Therefore it is a
DocumentationBlock.
:ivar annotations: Possibility to provide additional notes while
defining a model element (e.g. the ECU Configuration Parameter
Values). These are not intended as documentation but are mere
design notes.
:ivar activation_reason_representation_ref: If the
activationReasonRepresentation is referenced from the enclosing
AbstractEvent this shall be taken as an indication that the
latter contributes to the activating vector of this
ExecutableEntity that owns the referenced
ExecutableEntityActivationReason.
:ivar context_limitation_refs: The existence of this reference
indicates that the usage of the event is limited to the context
of the referred BswDistinguishedPartitions.
:ivar disabled_in_mode_irefs: The modes, in which this event is
disabled.
:ivar starts_on_event_ref: The entity which is started by the event.
:ivar variation_point: This element was generated/modified due to an
atpVariation stereotype.
:ivar s: Checksum calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine if
an ArObject has changed. The checksum has no semantic meaning
for an AUTOSAR model and there is no requirement for AUTOSAR
tools to manage the checksum.
:ivar t: Timestamp calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine
the last change of an ArObject. The timestamp has no semantic
meaning for an AUTOSAR model and there is no requirement for
AUTOSAR tools to manage the timestamp.
:ivar uuid: The purpose of this attribute is to provide a globally
unique identifier for an instance of a meta-class. The values of
this attribute should be globally unique strings prefixed by the
type of identifier. For example, to include a DCE UUID as
defined by The Open Group, the UUID would be preceded by "DCE:".
The values of this attribute may be used to support merging of
different AUTOSAR models. The form of the UUID (Universally
Unique Identifier) is taken from a standard defined by the Open
Group (was Open Software Foundation). This standard is widely
used, including by Microsoft for COM (GUIDs) and by many
companies for DCE, which is based on CORBA. The method for
generating these 128-bit IDs is published in the standard and
the effectiveness and uniqueness of the IDs is not in practice
disputed. If the id namespace is omitted, DCE is assumed. An
example is "DCE:2fac1234-31f8-11b4-a222-08002b34c003". The uuid
attribute has no semantic meaning for an AUTOSAR model and there
is no requirement for AUTOSAR tools to manage the timestamp.
"""
class Meta:
name = "BSW-OS-TASK-EXECUTION-EVENT"
short_name: Optional[Identifier] = field(
default=None,
metadata={
"name": "SHORT-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
"required": True,
}
)
short_name_fragments: Optional["BswOsTaskExecutionEvent.ShortNameFragments"] = field(
default=None,
metadata={
"name": "SHORT-NAME-FRAGMENTS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
long_name: Optional[MultilanguageLongName] = field(
default=None,
metadata={
"name": "LONG-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
desc: Optional[MultiLanguageOverviewParagraph] = field(
default=None,
metadata={
"name": "DESC",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
category: Optional[CategoryString] = field(
default=None,
metadata={
"name": "CATEGORY",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
admin_data: Optional[AdminData] = field(
default=None,
metadata={
"name": "ADMIN-DATA",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
introduction: Optional[DocumentationBlock] = field(
default=None,
metadata={
"name": "INTRODUCTION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
annotations: Optional["BswOsTaskExecutionEvent.Annotations"] = field(
default=None,
metadata={
"name": "ANNOTATIONS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
activation_reason_representation_ref: Optional["BswOsTaskExecutionEvent.ActivationReasonRepresentationRef"] = field(
default=None,
metadata={
"name": "ACTIVATION-REASON-REPRESENTATION-REF",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
context_limitation_refs: Optional["BswOsTaskExecutionEvent.ContextLimitationRefs"] = field(
default=None,
metadata={
"name": "CONTEXT-LIMITATION-REFS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
disabled_in_mode_irefs: Optional["BswOsTaskExecutionEvent.DisabledInModeIrefs"] = field(
default=None,
metadata={
"name": "DISABLED-IN-MODE-IREFS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
starts_on_event_ref: Optional["BswOsTaskExecutionEvent.StartsOnEventRef"] = field(
default=None,
metadata={
"name": "STARTS-ON-EVENT-REF",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
variation_point: Optional[VariationPoint] = field(
default=None,
metadata={
"name": "VARIATION-POINT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
s: Optional[str] = field(
default=None,
metadata={
"name": "S",
"type": "Attribute",
}
)
t: Optional[str] = field(
default=None,
metadata={
"name": "T",
"type": "Attribute",
"pattern": r"([0-9]{4}-[0-9]{2}-[0-9]{2})(T[0-9]{2}:[0-9]{2}:[0-9]{2}(Z|([+\-][0-9]{2}:[0-9]{2})))?",
}
)
uuid: Optional[str] = field(
default=None,
metadata={
"name": "UUID",
"type": "Attribute",
}
)
@dataclass
class ShortNameFragments:
short_name_fragment: List[ShortNameFragment] = field(
default_factory=list,
metadata={
"name": "SHORT-NAME-FRAGMENT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class Annotations:
annotation: List[Annotation] = field(
default_factory=list,
metadata={
"name": "ANNOTATION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class ActivationReasonRepresentationRef(Ref):
dest: Optional[ExecutableEntityActivationReasonSubtypesEnum] = field(
default=None,
metadata={
"name": "DEST",
"type": "Attribute",
"required": True,
}
)
@dataclass
class ContextLimitationRefs:
context_limitation_ref: List["BswOsTaskExecutionEvent.ContextLimitationRefs.ContextLimitationRef"] = field(
default_factory=list,
metadata={
"name": "CONTEXT-LIMITATION-REF",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class ContextLimitationRef(Ref):
dest: Optional[BswDistinguishedPartitionSubtypesEnum] = field(
default=None,
metadata={
"name": "DEST",
"type": "Attribute",
"required": True,
}
)
@dataclass
class DisabledInModeIrefs:
disabled_in_mode_iref: List[ModeInBswModuleDescriptionInstanceRef] = field(
default_factory=list,
metadata={
"name": "DISABLED-IN-MODE-IREF",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class StartsOnEventRef(Ref):
dest: Optional[BswModuleEntitySubtypesEnum] = field(
default=None,
metadata={
"name": "DEST",
"type": "Attribute",
"required": True,
}
)
class ContextLimitationRefs:
context_limitation_ref: List["BswOsTaskExecutionEvent.ContextLimitationRefs.ContextLimitationRef"] = field(
default_factory=list,
metadata={
"name": "CONTEXT-LIMITATION-REF",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class ContextLimitationRef(Ref):
dest: Optional[BswDistinguishedPartitionSubtypesEnum] = field(
default=None,
metadata={
"name": "DEST",
"type": "Attribute",
"required": True,
}
)
class SwAddrMethod:
"""Used to assign a common addressing method, e.g. common memory section,
to data or code objects.
These objects could actually live in different modules or
components.
:ivar short_name: This specifies an identifying shortName for the
object. It needs to be unique within its context and is intended
for humans but even more for technical reference.
:ivar short_name_fragments: This specifies how the
Referrable.shortName is composed of several shortNameFragments.
:ivar long_name: This specifies the long name of the object. Long
name is targeted to human readers and acts like a headline.
:ivar desc: This represents a general but brief (one paragraph)
description what the object in question is about. It is only one
paragraph! Desc is intended to be collected into overview
tables. This property helps a human reader to identify the
object in question. More elaborate documentation, (in particular
how the object is built or used) should go to "introduction".
:ivar category: The category is a keyword that specializes the
semantics of the Identifiable. It affects the expected existence
of attributes and the applicability of constraints.
:ivar admin_data: This represents the administrative data for the
identifiable object.
:ivar introduction: This represents more information about how the
object in question is built or is used. Therefore it is a
DocumentationBlock.
:ivar annotations: Possibility to provide additional notes while
defining a model element (e.g. the ECU Configuration Parameter
Values). These are not intended as documentation but are mere
design notes.
:ivar variation_point: This element was generated/modified due to an
atpVariation stereotype.
:ivar blueprint_policys: This role indicates whether the
blueprintable element will be modifiable or not motifiable.
:ivar short_name_pattern: This attribute represents the pattern
which shall be used to build the shortName of the derived
elements. As of now it is modeled as a String. In general it
should follow the pattern: pattern = (placeholder | namePart)*
placeholder = "{" namePart "}" namePart = identifier | "_" This
is subject to be refined in subsequent versions. Note that this
is marked as obsolete. Use the xml attribute namePattern instead
as it applies to Identifier and CIdentifier (shortName, symbol
etc.)
:ivar memory_allocation_keyword_policy: Enumeration to specify the
name pattern of the Memory Allocation Keyword.
:ivar options:
:ivar section_initialization_policy: Specifies the expected
initialization of the variables (inclusive those which are
implementing VariableDataPrototypes). Therefore this is an
implementation constraint for initialization code of BSW modules
(especially RTE) as well as the start-up code which initializes
the memory segment to which the AutosarDataPrototypes referring
to the SwAddrMethod's are later on mapped. If the attribute is
not defined it has the identical semantic as the attribute value
"INIT"
:ivar section_type: Defines the type of memory sections which can be
associated with this addresssing method.
:ivar s: Checksum calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine if
an ArObject has changed. The checksum has no semantic meaning
for an AUTOSAR model and there is no requirement for AUTOSAR
tools to manage the checksum.
:ivar t: Timestamp calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine
the last change of an ArObject. The timestamp has no semantic
meaning for an AUTOSAR model and there is no requirement for
AUTOSAR tools to manage the timestamp.
:ivar uuid: The purpose of this attribute is to provide a globally
unique identifier for an instance of a meta-class. The values of
this attribute should be globally unique strings prefixed by the
type of identifier. For example, to include a DCE UUID as
defined by The Open Group, the UUID would be preceded by "DCE:".
The values of this attribute may be used to support merging of
different AUTOSAR models. The form of the UUID (Universally
Unique Identifier) is taken from a standard defined by the Open
Group (was Open Software Foundation). This standard is widely
used, including by Microsoft for COM (GUIDs) and by many
companies for DCE, which is based on CORBA. The method for
generating these 128-bit IDs is published in the standard and
the effectiveness and uniqueness of the IDs is not in practice
disputed. If the id namespace is omitted, DCE is assumed. An
example is "DCE:2fac1234-31f8-11b4-a222-08002b34c003". The uuid
attribute has no semantic meaning for an AUTOSAR model and there
is no requirement for AUTOSAR tools to manage the timestamp.
"""
class Meta:
name = "SW-ADDR-METHOD"
short_name: Optional[Identifier] = field(
default=None,
metadata={
"name": "SHORT-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
"required": True,
}
)
short_name_fragments: Optional["SwAddrMethod.ShortNameFragments"] = field(
default=None,
metadata={
"name": "SHORT-NAME-FRAGMENTS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
long_name: Optional[MultilanguageLongName] = field(
default=None,
metadata={
"name": "LONG-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
desc: Optional[MultiLanguageOverviewParagraph] = field(
default=None,
metadata={
"name": "DESC",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
category: Optional[CategoryString] = field(
default=None,
metadata={
"name": "CATEGORY",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
admin_data: Optional[AdminData] = field(
default=None,
metadata={
"name": "ADMIN-DATA",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
introduction: Optional[DocumentationBlock] = field(
default=None,
metadata={
"name": "INTRODUCTION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
annotations: Optional["SwAddrMethod.Annotations"] = field(
default=None,
metadata={
"name": "ANNOTATIONS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
variation_point: Optional[VariationPoint] = field(
default=None,
metadata={
"name": "VARIATION-POINT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
blueprint_policys: Optional["SwAddrMethod.BlueprintPolicys"] = field(
default=None,
metadata={
"name": "BLUEPRINT-POLICYS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
short_name_pattern: Optional[String] = field(
default=None,
metadata={
"name": "SHORT-NAME-PATTERN",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
memory_allocation_keyword_policy: Optional[MemoryAllocationKeywordPolicyType] = field(
default=None,
metadata={
"name": "MEMORY-ALLOCATION-KEYWORD-POLICY",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
options: Optional["SwAddrMethod.Options"] = field(
default=None,
metadata={
"name": "OPTIONS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
section_initialization_policy: Optional[SectionInitializationPolicyType] = field(
default=None,
metadata={
"name": "SECTION-INITIALIZATION-POLICY",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
section_type: Optional[MemorySectionType] = field(
default=None,
metadata={
"name": "SECTION-TYPE",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
s: Optional[str] = field(
default=None,
metadata={
"name": "S",
"type": "Attribute",
}
)
t: Optional[str] = field(
default=None,
metadata={
"name": "T",
"type": "Attribute",
"pattern": r"([0-9]{4}-[0-9]{2}-[0-9]{2})(T[0-9]{2}:[0-9]{2}:[0-9]{2}(Z|([+\-][0-9]{2}:[0-9]{2})))?",
}
)
uuid: Optional[str] = field(
default=None,
metadata={
"name": "UUID",
"type": "Attribute",
}
)
@dataclass
class ShortNameFragments:
short_name_fragment: List[ShortNameFragment] = field(
default_factory=list,
metadata={
"name": "SHORT-NAME-FRAGMENT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class Annotations:
annotation: List[Annotation] = field(
default_factory=list,
metadata={
"name": "ANNOTATION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class BlueprintPolicys:
blueprint_policy_list: List[BlueprintPolicyList] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-LIST",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
blueprint_policy_not_modifiable: List[BlueprintPolicyNotModifiable] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-NOT-MODIFIABLE",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
blueprint_policy_single: List[BlueprintPolicySingle] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-SINGLE",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
@dataclass
class Options:
"""
:ivar option: This attribute introduces the ability to specify
further intended properties of the MemorySection in with the
related objects shall be placed. These properties are
handled as to be selected. The intended options are
mentioned in the list. In the Memory Mapping configuration,
this option list is used to determine an appropriate
MemMapAddressingModeSet.
"""
option: List[Identifier] = field(
default_factory=list,
metadata={
"name": "OPTION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
}
)
class BaseGANVocoderConfig(BaseVocoderConfig):
"""Base config class used among all the GAN based vocoders.
Args:
use_stft_loss (bool):
enable / disable the use of STFT loss. Defaults to True.
use_subband_stft_loss (bool):
enable / disable the use of Subband STFT loss. Defaults to True.
use_mse_gan_loss (bool):
enable / disable the use of Mean Squared Error based GAN loss. Defaults to True.
use_hinge_gan_loss (bool):
enable / disable the use of Hinge GAN loss. Defaults to True.
use_feat_match_loss (bool):
enable / disable feature matching loss. Defaults to True.
use_l1_spec_loss (bool):
enable / disable L1 spectrogram loss. Defaults to True.
stft_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 0.
subband_stft_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 0.
mse_G_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 1.
hinge_G_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 0.
feat_match_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 100.
l1_spec_loss_weight (float):
Loss weight that multiplies the computed loss value. Defaults to 45.
stft_loss_params (dict):
Parameters for the STFT loss. Defaults to `{"n_ffts": [1024, 2048, 512], "hop_lengths": [120, 240, 50], "win_lengths": [600, 1200, 240]}`.
l1_spec_loss_params (dict):
Parameters for the L1 spectrogram loss. Defaults to
`{
"use_mel": True,
"sample_rate": 22050,
"n_fft": 1024,
"hop_length": 256,
"win_length": 1024,
"n_mels": 80,
"mel_fmin": 0.0,
"mel_fmax": None,
}`
target_loss (str):
Target loss name that defines the quality of the model. Defaults to `avg_G_loss`.
gen_clip_grad (float):
Gradient clipping threshold for the generator model. Any value less than 0 disables clipping.
Defaults to -1.
disc_clip_grad (float):
Gradient clipping threshold for the discriminator model. Any value less than 0 disables clipping.
Defaults to -1.
lr_gen (float):
Generator model initial learning rate. Defaults to 0.0002.
lr_disc (float):
Discriminator model initial learning rate. Defaults to 0.0002.
optimizer (torch.optim.Optimizer):
Optimizer used for the training. Defaults to `AdamW`.
optimizer_params (dict):
Optimizer kwargs. Defaults to `{"betas": [0.8, 0.99], "weight_decay": 0.0}`
lr_scheduler_gen (torch.optim.Scheduler):
Learning rate scheduler for the generator. Defaults to `ExponentialLR`.
lr_scheduler_gen_params (dict):
Parameters for the generator learning rate scheduler. Defaults to `{"gamma": 0.999, "last_epoch": -1}`.
lr_scheduler_disc (torch.optim.Scheduler):
Learning rate scheduler for the discriminator. Defaults to `ExponentialLR`.
lr_scheduler_dict_params (dict):
Parameters for the discriminator learning rate scheduler. Defaults to `{"gamma": 0.999, "last_epoch": -1}`.
use_pqmf (bool):
enable / disable PQMF for subband approximation at training. Defaults to False.
steps_to_start_discriminator (int):
Number of steps required to start training the discriminator. Defaults to 0.
diff_samples_for_G_and_D (bool):
enable / disable use of different training samples for the generator and the discriminator iterations.
Enabling it results in slower iterations but faster convergance in some cases. Defaults to False.
"""
# LOSS PARAMETERS
use_stft_loss: bool = True
use_subband_stft_loss: bool = True
use_mse_gan_loss: bool = True
use_hinge_gan_loss: bool = True
use_feat_match_loss: bool = True # requires MelGAN Discriminators (MelGAN and HifiGAN)
use_l1_spec_loss: bool = True
# loss weights
stft_loss_weight: float = 0
subband_stft_loss_weight: float = 0
mse_G_loss_weight: float = 1
hinge_G_loss_weight: float = 0
feat_match_loss_weight: float = 100
l1_spec_loss_weight: float = 45
stft_loss_params: dict = field(
default_factory=lambda: {
"n_ffts": [1024, 2048, 512],
"hop_lengths": [120, 240, 50],
"win_lengths": [600, 1200, 240],
})
l1_spec_loss_params: dict = field(
default_factory=lambda: {
"use_mel": True,
"sample_rate": 22050,
"n_fft": 1024,
"hop_length": 256,
"win_length": 1024,
"n_mels": 80,
"mel_fmin": 0.0,
"mel_fmax": None,
})
target_loss: str = "avg_G_loss" # loss value to pick the best model to save after each epoch
# optimizer
gen_clip_grad: float = -1 # Generator gradient clipping threshold. Apply gradient clipping if > 0
disc_clip_grad: float = -1 # Discriminator gradient clipping threshold.
lr_gen: float = 0.0002 # Initial learning rate.
lr_disc: float = 0.0002 # Initial learning rate.
optimizer: str = "AdamW"
optimizer_params: dict = field(default_factory=lambda: {
"betas": [0.8, 0.99],
"weight_decay": 0.0
})
lr_scheduler_gen: str = "ExponentialLR" # one of the schedulers from https:#pytorch.org/docs/stable/optim.html
lr_scheduler_gen_params: dict = field(default_factory=lambda: {
"gamma": 0.999,
"last_epoch": -1
})
lr_scheduler_disc: str = "ExponentialLR" # one of the schedulers from https:#pytorch.org/docs/stable/optim.html
lr_scheduler_disc_params: dict = field(default_factory=lambda: {
"gamma": 0.999,
"last_epoch": -1
})
use_pqmf: bool = False # enable/disable using pqmf for multi-band training. (Multi-band MelGAN)
steps_to_start_discriminator = 0 # start training the discriminator after this number of steps.
diff_samples_for_G_and_D: bool = False # use different samples for G and D training steps.
class ReadContext:
session: Session
resource: "BaseResource"
uid: int
# request vars
include: Optional[List]
query: Optional[List[FilterItem]]
sort: Optional[list]
resource_id: Optional[int]
limit: Optional[int]
offset: Optional[int]
# runtime vars
q: Select = field(default_factory=Select)
obj_ids: List[int] = field(default_factory=list)
parent_payload: List = field(default_factory=list)
related_payload: Dict = field(default_factory=dict)
total: int = 0
# for Aggregation Resources
time_scale = None
@property
def is_list(self):
return not bool(self.resource_id)
def read__filtering(self):
if not self.query:
return
filters_to_apply = []
for f in self.query:
resource_field = self.resource.fields.get(f.field)
model_field = resource_field.metadata.get("model_field")
if model_field is None:
raise BadFilter(filter=f.field)
value = f.wrapper(f.value)
value = filter_value_to_python(value)
list_deserialization = isinstance(value, list) and not isinstance(
resource_field, ListField
)
if value is not None:
if list_deserialization:
value = [resource_field.deserialize(item) for item in value]
else:
value = resource_field.deserialize(value)
field_expr = getattr(model_field, f.op)(value)
filters_to_apply.append(field_expr)
self.q = self.q.where(sa.and_(*filters_to_apply))
def read__sorting(self):
if self.sort:
# TODO: add support for routes
# example: (xxx.names.display_name)
for sort_item in self.sort:
sort_route, sort_way = get_sort_way(sort_item)
if sort_route in self.resource.fields:
self.q = self.q.order_by(
sort_way(
self.resource.fields[sort_route].metadata.get(
"model_field"
)
)
)
def read__pagination(self):
if self.limit:
self.q = self.q.limit(self.limit)
if self.offset:
self.q = self.q.offset(self.offset)
def read__query(self):
fields_to_select = {}
to_group_by = []
for field_name, field in self.resource.fields.items():
model_field = field.metadata.get("model_field")
if field.load_only:
continue
if model_field is None:
raise Exception(
f"{self.resource.Meta.name}.{field_name} field must have "
f"'model_field' argument"
)
if isinstance(field, ToMany):
fields_to_select[field_name] = sa.func.array_remove(
sa.func.array_agg(model_field), None
)
elif isinstance(field, ToOne):
fields_to_select[field_name] = model_field
else:
fields_to_select[field_name] = model_field
if (
isinstance(model_field, InstrumentedAttribute)
and model_field.class_ is not self.resource.Meta.model
):
to_group_by.append(model_field)
q = sa.select([clm.label(lbl) for lbl, clm in fields_to_select.items()])
joins = getattr(self.resource.Meta, "select_from", None)
if joins is not None:
q = q.select_from(joins)
if not self.is_list:
model_id = get_id_field(self.resource)
q = q.where(model_id == self.resource_id)
if self.resource.Meta.auth:
q = self.resource.Meta.auth.can_read(self, q)
if joins is not None:
model_id = get_id_field(self.resource)
q = q.group_by(model_id, *to_group_by)
self.q = q
def __add_related_payload(self, related_res, related_data: list):
if related_res.Meta.name in self.related_payload:
related_res = related_res()
related_res_id_field = get_id_field(related_res, name_only=True)
existing_record_ids = [
rec[related_res_id_field]
for rec in self.related_payload[related_res.Meta.name]
]
for rec in related_data:
if rec[related_res_id_field] not in existing_record_ids:
self.related_payload[related_res.Meta.name].append(rec)
else:
self.related_payload[related_res.Meta.name] = related_data
def read__includes(self):
for i in self.include or []:
if "." in i:
raise BadRequest()
elif not isinstance(self.resource.fields.get(i), (ToOne, ToMany)):
raise RelationNotFound(f"Relation <{i}> not found")
related_resource_name = self.resource.fields[i].metadata["resource"]
related_res = self.resource.RESOURCES[related_resource_name]
related_field = self.resource.fields[i].metadata.get("model_field")
method_name = f"get_by_{self.resource.Meta.name.lower()}_ids"
if not hasattr(related_res, method_name):
raise BadRequest(
f"Relation {related_resource_name} doesn't ready yet. "
f"Ask developers to add {method_name} method "
f"to {related_resource_name} resource"
)
related_res_obj = related_res()
related_data = getattr(related_res_obj, method_name)(
self.session, self, related_field
)
self.__add_related_payload(related_res, related_data)
def read__serializing(self) -> dict:
response = self.resource.Response(self.resource, self.is_list)
response.set_parent_payload(self.parent_payload)
response.set_related_payload(self.related_payload)
response.set_total(self.total)
serialized_response = response.serialize()
return serialized_response
def read__execute_query(self):
if not self.resource.Meta.disable_total:
self.q.append_column(sa.func.count().over().label("total"))
result = self.session.execute(self.q).fetchall()
serialized_data = self.resource.dump(result, many=True)
id_field = get_id_field(self.resource, name_only=True)
self.obj_ids.extend([_i[id_field] for _i in serialized_data])
self.parent_payload = serialized_data
if serialized_data and not self.resource.Meta.disable_total:
self.total = result[0].total
class _Name_default:
middle_name_1: Optional[str] = field(default=None)
middle_name_2: Optional[str] = field(default=None)
maiden_name: Optional[str] = field(default=None)
divorcée: Optional[str] = field(default=None)
class _Party_default:
party_entry: str = field(default="unknown")
party_exit: str = field(default="unknown")
class Regex:
att: Optional[str] = field(default=None,
metadata={
"type": "Attribute",
"pattern": r"\p{IsHangulJamo}+",
})
class Cnmt():
title_id: str
version: int
type: SwitchContentMetaType
contents: dict[bytes, tuple[int, ContentType]] = field(compare=False)
class CoctMt090003Uv01AssignedEntity:
class Meta:
name = "COCT_MT090003UV01.AssignedEntity"
realm_code: List[Cs] = field(
default_factory=list,
metadata={
"name": "realmCode",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
type_id: Optional[Ii] = field(
default=None,
metadata={
"name": "typeId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
template_id: List[Ii] = field(
default_factory=list,
metadata={
"name": "templateId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
id: List[Ii] = field(
default_factory=list,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
"min_occurs": 1,
}
)
code: Optional[Ce] = field(
default=None,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
addr: List[AdExplicit] = field(
default_factory=list,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
telecom: List[TelExplicit] = field(
default_factory=list,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
assigned_person: Optional[CoctMt090003Uv01Person] = field(
default=None,
metadata={
"name": "assignedPerson",
"type": "Element",
"namespace": "urn:hl7-org:v3",
"nillable": True,
}
)
assigned_device: Optional[CoctMt090003Uv01Device] = field(
default=None,
metadata={
"name": "assignedDevice",
"type": "Element",
"namespace": "urn:hl7-org:v3",
"nillable": True,
}
)
assigned_organization: Optional[CoctMt090003Uv01Organization] = field(
default=None,
metadata={
"name": "assignedOrganization",
"type": "Element",
"namespace": "urn:hl7-org:v3",
"nillable": True,
}
)
represented_organization: Optional[CoctMt150003Uv03Organization] = field(
default=None,
metadata={
"name": "representedOrganization",
"type": "Element",
"namespace": "urn:hl7-org:v3",
"nillable": True,
}
)
null_flavor: Optional[NullFlavor] = field(
default=None,
metadata={
"name": "nullFlavor",
"type": "Attribute",
}
)
class_code: Optional[RoleClassAssignedEntity] = field(
default=None,
metadata={
"name": "classCode",
"type": "Attribute",
"required": True,
}
)
class TransformerLanguageModelConfig(FairseqDataclass):
activation_fn: ChoiceEnum(utils.get_available_activation_fns()) = field(
default="relu", metadata={"help": "activation function to use"})
dropout: float = field(default=0.1,
metadata={"help": "dropout probability"})
attention_dropout: float = field(
default=0.0,
metadata={"help": "dropout probability for attention weights"})
activation_dropout: float = field(
default=0.0,
metadata={"help": "dropout probability after activation in FFN."})
relu_dropout: float = field(
default=0.0,
metadata={"help": "dropout probability after activation in FFN."})
decoder_embed_dim: int = field(
default=512, metadata={"help": "decoder embedding dimension"})
decoder_output_dim: int = field(
default=512, metadata={"help": "decoder output dimension"})
decoder_input_dim: int = field(
default=512, metadata={"help": "decoder input dimension"})
decoder_ffn_embed_dim: int = field(
default=2048, metadata={"help": "decoder embedding dimension for FFN"})
decoder_layers: int = field(default=6,
metadata={"help": "num decoder layers"})
decoder_attention_heads: int = field(
default=8, metadata={"help": "num decoder attention heads"})
decoder_normalize_before: bool = field(
default=False,
metadata={"help": "apply layernorm before each decoder block"})
no_decoder_final_norm: bool = field(
default=False,
metadata={
"help": "don't add an extra layernorm after the last decoder block"
},
)
adaptive_softmax_cutoff: Optional[str] = field(
default=None,
metadata={
"help":
"comma separated list of adaptive softmax cutoff points. "
"Must be used with adaptive_loss criterion"
},
)
adaptive_softmax_dropout: float = field(
default=0,
metadata={
"help": "sets adaptive softmax dropout for the tail projections"
},
)
adaptive_softmax_factor: float = field(
default=4, metadata={"help": "adaptive input factor"})
no_token_positional_embeddings: bool = field(
default=False,
metadata={
"help":
"if set, disables positional embeddings (outside self attention)"
},
)
share_decoder_input_output_embed: bool = field(
default=False,
metadata={"help": "share decoder input and output embeddings"})
character_embeddings: bool = field(
default=False,
metadata={
"help":
"if set, uses character embedding convolutions to produce token embeddings"
},
)
character_filters: str = field(
default=
"[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]",
metadata={"help": "size of character embeddings"},
)
character_embedding_dim: int = field(
default=4, metadata={"help": "size of character embeddings"})
char_embedder_highway_layers: int = field(
default=2,
metadata={
"help": "number of highway layers for character token embeddder"
},
)
adaptive_input: bool = field(
default=False, metadata={"help": "if set, uses adaptive input"})
adaptive_input_factor: float = field(
default=4, metadata={"help": "adaptive input factor"})
adaptive_input_cutoff: Optional[str] = field(
default=None,
metadata={
"help": "comma separated list of adaptive input cutoff points."
},
)
tie_adaptive_weights: bool = field(
default=False,
metadata={
"help":
"if set, ties the weights of adaptive softmax and adaptive input"
},
)
tie_adaptive_proj: bool = field(
default=False,
metadata={
"help":
"if set, ties the projection weights of adaptive softmax and adaptive input"
},
)
decoder_learned_pos: bool = field(
default=False,
metadata={"help": "use learned positional embeddings in the decoder"},
)
layernorm_embedding: bool = field(
default=False, metadata={"help": "add layernorm to embedding"})
no_scale_embedding: bool = field(
default=False, metadata={"help": "if True, dont scale embeddings"})
checkpoint_activations: bool = field(
default=False,
metadata={"help": "checkpoint activations at each layer"})
offload_activations: bool = field(
default=False,
metadata={
"help": "move checkpointed activations to CPU after they are used."
},
)
# config for "Reducing Transformer Depth on Demand with Structured Dropout" (Fan et al., 2019)
decoder_layerdrop: float = field(
default=0.0, metadata={"help": "LayerDrop probability for decoder"})
decoder_layers_to_keep: Optional[str] = field(
default=None,
metadata={
"help":
"which layers to *keep* when pruning as a comma-separated list"
},
)
# config for Training with Quantization Noise for Extreme Model Compression ({Fan*, Stock*} et al., 2020)
quant_noise_pq: float = field(
default=0.0,
metadata={"help": "iterative PQ quantization noise at training time"},
)
quant_noise_pq_block_size: int = field(
default=8,
metadata={"help": "block size of quantization noise at training time"},
)
quant_noise_scalar: float = field(
default=0.0,
metadata={
"help":
"scalar quantization noise and scalar quantization at training time"
},
)
# config for Fully Sharded Data Parallel (FSDP) training
min_params_to_wrap: int = field(
default=DEFAULT_MIN_PARAMS_TO_WRAP,
metadata={
"help":
("minimum number of params for a layer to be wrapped with FSDP() when "
"training with --ddp-backend=fully_sharded. Smaller values will "
"improve memory efficiency, but may make torch.distributed "
"communication less efficient due to smaller input sizes. This option "
"is set to 0 (i.e., always wrap) when --checkpoint-activations or "
"--offload-activations are passed.")
})
# options from other parts of the config
add_bos_token: bool = II("task.add_bos_token")
tokens_per_sample: int = II("task.tokens_per_sample")
max_target_positions: Optional[int] = II("task.max_target_positions")
tpu: bool = II("common.tpu")
class ProjectConfig(JsonSchemaMixin, allow_additional_props=False):
"""Project specific configuration section"""
name: Optional[ProjectName] = field(default=None,
metadata=METADATA["project__name"])
auth: Optional[Dict[Region, str]] = field(default=None,
metadata=METADATA["auth"])
owner: Optional[str] = field(default=None,
metadata=METADATA["project__owner"])
regions: Optional[List[Region]] = field(default=None,
metadata=METADATA["regions"])
az_blacklist: Optional[List[AzId]] = field(default=None,
metadata=METADATA["az_ids"])
package_lambda: Optional[bool] = field(default=None,
metadata=METADATA["package_lambda"])
lambda_zip_path: Optional[str] = field(
default=None, metadata=METADATA["lambda_zip_path"])
lambda_source_path: Optional[str] = field(
default=None, metadata=METADATA["lambda_source_path"])
s3_bucket: Optional[S3BucketName] = field(default=None,
metadata=METADATA["s3_bucket"])
parameters: Optional[Dict[ParameterKey, ParameterValue]] = field(
default=None, metadata=METADATA["parameters"])
build_submodules: Optional[bool] = field(
default=None, metadata=METADATA["build_submodules"])
template: Optional[str] = field(default=None,
metadata=METADATA["template"])
tags: Optional[Dict[TagKey, TagValue]] = field(default=None,
metadata=METADATA["tags"])
s3_enable_sig_v2: Optional[bool] = field(
default=None, metadata=METADATA["enable_sig_v2"])
s3_object_acl: Optional[str] = field(default=None,
metadata=METADATA["s3_object_acl"])
class Armor:
name: str = ''
dur: int = 0
mass: float = 0
mat: Tuple = field(default_factory=tuple)
defl: HarmStats = field(default_factory=HarmStats)
class Context:
Schema: ClassVar[Type[Schema]] = Schema
name: str
account_id: str
region: str
env_tag: str
env_stack_name: str
env_ssm_parameter_name: str
eks_stack_name: str
ssm_parameter_name: str
ssm_dockerhub_parameter_name: str
toolkit: ToolkitManifest
cdk_toolkit: CdkToolkitManifest
networking: NetworkingContext
user_pool_id: Optional[str] = None
scratch_bucket_arn: Optional[str] = None
cognito_users_url: Optional[str] = None
cognito_external_provider_redirect: Optional[str] = None
cognito_external_provider_domain: Optional[str] = None
landing_page_url: Optional[str] = None
k8_dashboard_url: Optional[str] = None
codeartifact_domain: Optional[str] = None
codeartifact_repository: Optional[str] = None
cognito_external_provider: Optional[str] = None
cognito_external_provider_label: Optional[str] = None
eks_cluster_role_arn: Optional[str] = None
eks_fargate_profile_role_arn: Optional[str] = None
eks_env_nodegroup_role_arn: Optional[str] = None
eks_oidc_provider: Optional[str] = None
eks_system_masters_roles: Optional[List[str]] = cast(
List[str], field(default_factory=list))
user_pool_client_id: Optional[str] = None
identity_pool_id: Optional[str] = None
teams: List[TeamContext] = field(default_factory=list)
images: ImagesManifest = ImagesManifest()
elbs: Optional[Dict[str, Dict[str, Any]]] = None
shared_efs_fs_id: Optional[str] = None
shared_efs_sg_id: Optional[str] = None
cluster_sg_id: Optional[str] = None
cluster_pod_sg_id: Optional[str] = None
managed_nodegroups: List[ManagedNodeGroupManifest] = field(
default_factory=list)
policies: Optional[List[str]] = cast(List[str],
field(default_factory=list))
orbit_version: Optional[str] = aws_orbit.__version__
helm_repository: Optional[str] = None
install_ssm_agent: Optional[bool] = False
install_image_replicator: Optional[bool] = False
def get_team_by_name(self, name: str) -> Optional[TeamContext]:
for t in self.teams:
if t.name == name:
return t
return None
def remove_team_by_name(self, name: str) -> None:
self.teams = [t for t in self.teams if t.name != name]
def fetch_env_data(self) -> None:
_logger.debug("Fetching Env data...")
values = ssm.get_parameter(name=self.env_ssm_parameter_name)
self.eks_cluster_role_arn = values["EksClusterRoleArn"]
self.eks_fargate_profile_role_arn = values["EksFargateProfileRoleArn"]
self.eks_env_nodegroup_role_arn = values["EksEnvNodegroupRoleArn"]
self.user_pool_id = values["UserPoolId"]
self.user_pool_client_id = values["UserPoolClientId"]
self.identity_pool_id = values["IdentityPoolId"]
self.cluster_pod_sg_id = values["ClusterPodSecurityGroupId"]
self.fetch_cognito_external_idp_data()
_logger.debug("Env data fetched successfully.")
def fetch_cognito_external_idp_data(self) -> None:
_logger.debug("Fetching Cognito External IdP data...")
if self.cognito_external_provider:
client = boto3_client(service_name="cognito-idp")
response: Dict[str, Any] = client.describe_user_pool(
UserPoolId=self.user_pool_id)
domain: str = response["UserPool"].get("Domain")
self.cognito_external_provider_domain = f"{domain}.auth.{self.region}.amazoncognito.com"
_logger.debug("cognito_external_provider_domain: %s",
self.cognito_external_provider_domain)
response = client.describe_user_pool_client(
UserPoolId=self.user_pool_id,
ClientId=self.user_pool_client_id)
self.cognito_external_provider_redirect = response[
"UserPoolClient"]["CallbackURLs"][0]
_logger.debug("cognito_external_provider_redirect: %s",
self.cognito_external_provider_redirect)
_logger.debug("Cognito External IdP data fetched successfully.")
else:
self.cognito_external_provider_domain = None
self.cognito_external_provider_domain_label = None
_logger.debug("No Cognito External IdP data to fetch.")
def fetch_teams_data(self) -> None:
_logger.debug("Fetching Teams data...")
for team in self.teams:
team.fetch_team_data()
self.fetch_cognito_external_idp_data()
_logger.debug("Env data fetched successfully.")
class DefendStats:
skill: float = 1.0 # combination of experience, combat training, reflex training, observation, spidy-sense...
playerstats: PlayerStats = field(default_factory=PlayerStats)
class ParameterSet(Prototype):
"""
{begin_markdown ParameterSet}
{spell_markdown init inits param params gprior}
# `curvefit.core.parameter.ParameterSet`
## A set of parameters that together specify the functional form of a curve
A `ParameterSet` is a set of parameters that define the functional form for the curve.
For example, if the parametric curve you want
to fit has three parameters then you need 1 `ParameterSet` objects that consists of 3 `Parameter` objects.
A `ParameterSet` is made up of one or more
[`Parameter`](Parameter.md) objects, which are each made up of one or more [`Variable`](Variable.md) objects.
Please refer to their documentation for more details on those objects.
A `ParameterSet` can also encode functional priors -- priors for functions of the parameter list that is
passed into a `ParameterSet`.
## Arguments
- `parameters (List[curvefit.core.parameter.Parameter])`: a list of `Parameter` instances
- `parameter_functions (List[Tuple[Callable, List[float]]]`: a list of tuples which each contain
(0) functions to apply to the `parameters` list and
(1) a prior for the parameter function (mean and standard deviation --
see [`Variable`](Variable.md#arguments) for more details about priors)
## Attributes
All attributes from the `Parameter`s in the list in the `parameters` argument are carried over to
`ParameterSet` but they are put into a list. For example, the `fe_init` attribute for `ParameterSet` is a list of
the `fe_init` attributes for each `Parameter` in the order that they were passed in `parameters` list (which
are lists of `fe_inits` for each `Variable` within a `Parameter` (see [here](Parameter.md#attributes) for more).
*Additional* attributes that are not lists of the individual `Parameter` attributes are listed below.
- `self.num_fe (int)`: total number of effects for the parameter set (number of variables)
## Methods
### `delete_random_effects`
Returns a copy of itself but with random effects bounds set to 0. This means that the parameter set
will not have any random effects in the model. Useful for when the same parameter set will be used to fit
jointly to many groups before being fit to individual groups.
## Usage
```python
from curvefit.core.parameter import Parameter, Variable, ParameterSet
var = Variable(covariate='ones', var_link_fun=lambda x: x, fe_init=0., re_init=0.)
param = Parameter(param_name='alpha', link_fun=lambda x: x, variables=[var])
param_function = ParameterFunction(
param_function_name='alpha_squared',
param_function=lambda params: params[0] ** 2,
param_function_fe_gprior=[0., np.inf]
)
param_set = ParameterSet(
parameters=[param], parameter_functions=[param_function]
)
```
{end_markdown ParameterSet}
"""
parameters: InitVar[List[Parameter]]
parameter_functions: InitVar[List[ParameterFunction]] = None
param_name: List[str] = field(init=False)
num_fe: int = field(init=False)
link_fun: List[Callable] = field(init=False)
covariate: List[List[str]] = field(init=False)
var_link_fun: List[List[Callable]] = field(init=False)
fe_init: List[List[float]] = field(init=False)
re_init: List[List[float]] = field(init=False)
re_zero_sum_std: List[List[float]] = field(init=False)
fe_gprior: List[List[List[float]]] = field(init=False)
re_gprior: List[List[List[float]]] = field(init=False)
fe_bounds: List[List[List[float]]] = field(init=False)
re_bounds: List[List[List[float]]] = field(init=False)
param_function_name: List[str] = field(init=False)
param_function: List[Callable] = field(init=False)
param_function_fe_gprior: List[List[float]] = field(init=False)
def __post_init__(self, parameters, parameter_functions):
for k, v in consolidate(Parameter, parameters, exclude=['num_fe']).items():
self.__setattr__(k, v)
for k, v in consolidate(ParameterFunction, parameter_functions).items():
self.__setattr__(k, v)
if len(set(self.param_name)) != len(self.param_name):
raise RuntimeError("Cannot have duplicate parameters in a set.")
if len(set(self.param_function_name)) != len(self.param_function_name):
raise RuntimeError("Cannot have duplicate parameter functions in a set.")
self.num_fe = 0
for param in parameters:
self.num_fe += param.num_fe
def get_param_index(self, param_name):
try:
param_index = self.param_name.index(param_name)
except ValueError:
raise RuntimeError(f"No {param_name} parameter in this parameter set.")
return param_index
def get_param_function_index(self, param_function_name):
try:
param_function_index = self.param_function_name.index(param_function_name)
except ValueError:
raise RuntimeError(f"No {param_function_name} parameter in this parameter set.")
return param_function_index
def delete_random_effects(self):
param_set = self.clone()
bounds = np.array(self.re_bounds)
bounds[:] = 0.
param_set.re_bounds = bounds.tolist()
return param_set
class TrainingArguments:
"""
TrainingArguments is the subset of the arguments we use in our example scripts **which relate to the training loop
itself**.
Using [`PdArgumentParser`] we can turn this class into
[argparse](https://docs.python.org/3/library/argparse#module-argparse) arguments that can be specified on the
command line.
Parameters:
output_dir (`str`):
The output directory where the model predictions and checkpoints will be written.
overwrite_output_dir (`bool`, *optional*, defaults to `False`):
If `True`, overwrite the content of the output directory. Use this to continue training if `output_dir`
points to a checkpoint directory.
do_train (`bool`, *optional*, defaults to `False`):
Whether to run training or not. This argument is not directly used by [`Trainer`], it's intended to be used
by your training/evaluation scripts instead. See the [example
scripts](https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples) for more details.
do_eval (`bool`, *optional*):
Whether to run evaluation on the validation set or not. Will be set to `True` if `evaluation_strategy` is
different from `"no"`. This argument is not directly used by [`Trainer`], it's intended to be used by your
training/evaluation scripts instead. See the [example
scripts](https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples) for more details.
do_predict (`bool`, *optional*, defaults to `False`):
Whether to run predictions on the test set or not. This argument is not directly used by [`Trainer`], it's
intended to be used by your training/evaluation scripts instead. See the [example
scripts](https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples) for more details.
do_export (`bool`, *optional*, defaults to `False`):
Whether to export inference model or not. This argument is not directly used by [`Trainer`], it's
intended to be used by your training/evaluation scripts instead.
evaluation_strategy (`str` or [`~trainer_utils.IntervalStrategy`], *optional*, defaults to `"no"`):
The evaluation strategy to adopt during training. Possible values are:
- `"no"`: No evaluation is done during training.
- `"steps"`: Evaluation is done (and logged) every `eval_steps`.
- `"epoch"`: Evaluation is done at the end of each epoch.
prediction_loss_only (`bool`, *optional*, defaults to `False`):
When performing evaluation and generating predictions, only returns the loss.
per_device_train_batch_size (`int`, *optional*, defaults to 8):
The batch size per GPU core/CPU for training.
per_device_eval_batch_size (`int`, *optional*, defaults to 8):
The batch size per GPU core/CPU for evaluation.
gradient_accumulation_steps (`int`, *optional*, defaults to 1):
Number of updates steps to accumulate the gradients for, before performing a backward/update pass.
<Tip warning={true}>
When using gradient accumulation, one step is counted as one step with backward pass. Therefore, logging,
evaluation, save will be conducted every `gradient_accumulation_steps * xxx_step` training examples.
</Tip>
learning_rate (`float`, *optional*, defaults to 5e-5):
The initial learning rate for [`AdamW`] optimizer.
weight_decay (`float`, *optional*, defaults to 0):
The weight decay to apply (if not zero) to all layers except all bias and LayerNorm weights in [`AdamW`]
optimizer.
adam_beta1 (`float`, *optional*, defaults to 0.9):
The beta1 hyperparameter for the [`AdamW`] optimizer.
adam_beta2 (`float`, *optional*, defaults to 0.999):
The beta2 hyperparameter for the [`AdamW`] optimizer.
adam_epsilon (`float`, *optional*, defaults to 1e-8):
The epsilon hyperparameter for the [`AdamW`] optimizer.
max_grad_norm (`float`, *optional*, defaults to 1.0):
Maximum gradient norm (for gradient clipping).
num_train_epochs(`float`, *optional*, defaults to 3.0):
Total number of training epochs to perform (if not an integer, will perform the decimal part percents of
the last epoch before stopping training).
max_steps (`int`, *optional*, defaults to -1):
If set to a positive number, the total number of training steps to perform. Overrides `num_train_epochs`.
In case of using a finite iterable dataset the training may stop before reaching the set number of steps
when all data is exhausted
lr_scheduler_type (`str` or [`SchedulerType`], *optional*, defaults to `"linear"`):
The scheduler type to use. See the documentation of [`SchedulerType`] for all possible values.
warmup_ratio (`float`, *optional*, defaults to 0.0):
Ratio of total training steps used for a linear warmup from 0 to `learning_rate`.
warmup_steps (`int`, *optional*, defaults to 0):
Number of steps used for a linear warmup from 0 to `learning_rate`. Overrides any effect of `warmup_ratio`.
log_on_each_node (`bool`, *optional*, defaults to `True`):
In multinode distributed training, whether to log using `log_level` once per node, or only on the main
node.
logging_dir (`str`, *optional*):
log directory. Will default to *output_dir/runs/**CURRENT_DATETIME_HOSTNAME***.
logging_strategy (`str` or [`~trainer_utils.IntervalStrategy`], *optional*, defaults to `"steps"`):
The logging strategy to adopt during training. Possible values are:
- `"no"`: No logging is done during training.
- `"epoch"`: Logging is done at the end of each epoch.
- `"steps"`: Logging is done every `logging_steps`.
logging_first_step (`bool`, *optional*, defaults to `False`):
Whether to log and evaluate the first `global_step` or not.
logging_steps (`int`, *optional*, defaults to 500):
Number of update steps between two logs if `logging_strategy="steps"`.
save_strategy (`str` or [`~trainer_utils.IntervalStrategy`], *optional*, defaults to `"steps"`):
The checkpoint save strategy to adopt during training. Possible values are:
- `"no"`: No save is done during training.
- `"epoch"`: Save is done at the end of each epoch.
- `"steps"`: Save is done every `save_steps`.
save_steps (`int`, *optional*, defaults to 500):
Number of updates steps before two checkpoint saves if `save_strategy="steps"`.
save_total_limit (`int`, *optional*):
If a value is passed, will limit the total amount of checkpoints. Deletes the older checkpoints in
`output_dir`.
save_on_each_node (`bool`, *optional*, defaults to `False`):
When doing multi-node distributed training, whether to save models and checkpoints on each node, or only on
the main one.
This should not be activated when the different nodes use the same storage as the files will be saved with
the same names for each node.
no_cuda (`bool`, *optional*, defaults to `False`):
Whether to not use CUDA even when it is available or not.
seed (`int`, *optional*, defaults to 42):
Random seed that will be set at the beginning of training. To ensure reproducibility across runs, use the
[`~Trainer.model_init`] function to instantiate the model if it has some randomly initialized parameters.
fp16 (`bool`, *optional*, defaults to `False`):
Whether to use fp16 16-bit (mixed) precision training instead of 32-bit training.
fp16_opt_level (`str`, *optional*, defaults to 'O1'):
For `fp16` training, AMP optimization level selected in ['O0', 'O1', 'O2']. See details at
https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/amp/auto_cast_cn.html
scale_loss (`float`, *optional*, defaults to 32768):
The value of initial scale_loss for fp16. (default: 32768)
local_rank (`int`, *optional*, defaults to -1):
Rank of the process during distributed training.
dataloader_drop_last (`bool`, *optional*, defaults to `False`):
Whether to drop the last incomplete batch (if the length of the dataset is not divisible by the batch size)
or not.
eval_steps (`int`, *optional*):
Number of update steps between two evaluations if `evaluation_strategy="steps"`. Will default to the same
value as `logging_steps` if not set.
dataloader_num_workers (`int`, *optional*, defaults to 0):
Number of subprocesses to use for data loading. 0 means that the data will be loaded in the
main process.
past_index (`int`, *optional*, defaults to -1):
Some models like TransformerXL or XLNet can make use of the past hidden states for their predictions.
If this argument is set to a positive int, the `Trainer` will use the corresponding output (usually index 2) as
the past state and feed it to the model at the next training step under the keyword argument `mems`.
run_name (`str`, *optional*):
A descriptor for the run. Typically used for logging.
disable_tqdm (`bool`, *optional*):
Whether or not to disable the tqdm progress bars and table of metrics. Will default to `True` if the logging
level is set to warn or lower (default), `False` otherwise.
remove_unused_columns (`bool`, *optional*, defaults to `True`):
If using `datasets.Dataset` datasets, whether or not to automatically remove the columns unused by the
model forward method.
label_names (`List[str]`, *optional*):
The list of keys in your dictionary of inputs that correspond to the labels.
Will eventually default to `["labels"]` except if the model used is one of the `XxxForQuestionAnswering` in
which case it will default to `["start_positions", "end_positions"]`.
load_best_model_at_end (`bool`, *optional*, defaults to `False`):
Whether or not to load the best model found during training at the end of training.
<Tip>
When set to `True`, the parameters `save_strategy` needs to be the same as `eval_strategy`, and in the case
it is "steps", `save_steps` must be a round multiple of `eval_steps`.
</Tip>
metric_for_best_model (`str`, *optional*):
Use in conjunction with `load_best_model_at_end` to specify the metric to use to compare two different
models. Must be the name of a metric returned by the evaluation with or without the prefix `"eval_"`. Will
default to `"loss"` if unspecified and `load_best_model_at_end=True` (to use the evaluation loss).
If you set this value, `greater_is_better` will default to `True`. Don't forget to set it to `False` if
your metric is better when lower.
greater_is_better (`bool`, *optional*):
Use in conjunction with `load_best_model_at_end` and `metric_for_best_model` to specify if better models
should have a greater metric or not. Will default to:
- `True` if `metric_for_best_model` is set to a value that isn't `"loss"` or `"eval_loss"`.
- `False` if `metric_for_best_model` is not set, or set to `"loss"` or `"eval_loss"`.
ignore_data_skip (`bool`, *optional*, defaults to `False`):
When resuming training, whether or not to skip the epochs and batches to get the data loading at the same
stage as in the previous training. If set to `True`, the training will begin faster (as that skipping step
can take a long time) but will not yield the same results as the interrupted training would have.
optim (`str` or [`training_args.OptimizerNames`], *optional*, defaults to `"adamw"`):
The optimizer to use: adamw, or adafactor.
length_column_name (`str`, *optional*, defaults to `"length"`):
Column name for precomputed lengths. If the column exists, grouping by length will use these values rather
than computing them on train startup. Ignored unless `group_by_length` is `True` and the dataset is an
instance of `Dataset`.
report_to (`str` or `List[str]`, *optional*, defaults to `"visualdl"`):
The list of integrations to report the results and logs to. Supported platforms is `"visualdl"`.
`"none"` for no integrations.
resume_from_checkpoint (`str`, *optional*):
The path to a folder with a valid checkpoint for your model. This argument is not directly used by
[`Trainer`], it's intended to be used by your training/evaluation scripts instead. See the [example
scripts](https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples) for more details.
"""
output_dir: str = field(metadata={
"help":
"The output directory where the model predictions and checkpoints will be written."
}, )
overwrite_output_dir: bool = field(
default=False,
metadata={
"help":
("Overwrite the content of the output directory. "
"Use this to continue training if output_dir points to a checkpoint directory."
)
},
)
do_train: bool = field(default=False,
metadata={"help": "Whether to run training."})
do_eval: bool = field(
default=False,
metadata={"help": "Whether to run eval on the dev set."})
do_predict: bool = field(
default=False,
metadata={"help": "Whether to run predictions on the test set."})
do_export: bool = field(
default=False, metadata={"help": "Whether to export infernece model."})
evaluation_strategy: IntervalStrategy = field(
default="no",
metadata={"help": "The evaluation strategy to use."},
)
prediction_loss_only: bool = field(
default=False,
metadata={
"help":
"When performing evaluation and predictions, only returns the loss."
},
)
per_device_train_batch_size: int = field(
default=8,
metadata={"help": "Batch size per GPU core/CPU for training."})
per_device_eval_batch_size: int = field(
default=8,
metadata={"help": "Batch size per GPU core/CPU for evaluation."})
gradient_accumulation_steps: int = field(
default=1,
metadata={
"help":
"Number of updates steps to accumulate before performing a backward/update pass."
},
)
learning_rate: float = field(
default=5e-5,
metadata={"help": "The initial learning rate for AdamW."})
weight_decay: float = field(
default=0.0,
metadata={"help": "Weight decay for AdamW if we apply some."})
adam_beta1: float = field(default=0.9,
metadata={"help": "Beta1 for AdamW optimizer"})
adam_beta2: float = field(default=0.999,
metadata={"help": "Beta2 for AdamW optimizer"})
adam_epsilon: float = field(
default=1e-8, metadata={"help": "Epsilon for AdamW optimizer."})
max_grad_norm: float = field(default=1.0,
metadata={"help": "Max gradient norm."})
num_train_epochs: float = field(
default=3.0,
metadata={"help": "Total number of training epochs to perform."})
max_steps: int = field(
default=-1,
metadata={
"help":
"If > 0: set total number of training steps to perform. Override num_train_epochs."
},
)
lr_scheduler_type: str = field(
default="linear",
metadata={"help": "The scheduler type to use."},
)
warmup_ratio: float = field(
default=0.0,
metadata={
"help": "Linear warmup over warmup_ratio fraction of total steps."
})
warmup_steps: int = field(
default=0, metadata={"help": "Linear warmup over warmup_steps."})
log_on_each_node: bool = field(
default=True,
metadata={
"help":
"When doing a multinode distributed training, whether to log once per node or just once on the main node."
},
)
logging_dir: Optional[str] = field(default=None,
metadata={"help": "VisualDL log dir."})
logging_strategy: IntervalStrategy = field(
default="steps",
metadata={"help": "The logging strategy to use."},
)
logging_first_step: bool = field(
default=False, metadata={"help": "Log the first global_step"})
logging_steps: int = field(default=500,
metadata={"help": "Log every X updates steps."})
save_strategy: IntervalStrategy = field(
default="steps",
metadata={"help": "The checkpoint save strategy to use."},
)
save_steps: int = field(
default=500,
metadata={"help": "Save checkpoint every X updates steps."})
save_total_limit: Optional[int] = field(
default=None,
metadata={
"help":
("Limit the total amount of checkpoints. "
"Deletes the older checkpoints in the output_dir. Default is unlimited checkpoints"
)
},
)
save_on_each_node: bool = field(
default=False,
metadata={
"help":
"When doing multi-node distributed training, whether to save models and checkpoints on each node, or only on the main one"
},
)
no_cuda: bool = field(
default=False,
metadata={"help": "Do not use CUDA even when it is available"})
seed: int = field(
default=42,
metadata={
"help":
"Random seed that will be set at the beginning of training."
})
fp16: bool = field(
default=False,
metadata={
"help": "Whether to use fp16 (mixed) precision instead of 32-bit"
},
)
fp16_opt_level: str = field(
default="O1",
metadata={
"help":
("For fp16: AMP optimization level selected in ['O0', 'O1', and 'O2']. "
"See details at https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/amp/auto_cast_cn.html"
)
},
)
scale_loss: float = field(
default=2**15,
metadata={"help": "The value of initial scale_loss for fp16."})
minimum_eval_times: int = field(
default=None,
metadata={
"help":
"If under eval_steps, the valid time is less then minimum_eval_times, the config of override eval_steps."
})
local_rank: int = field(
default=-1, metadata={"help": "For distributed training: local_rank"})
dataloader_drop_last: bool = field(
default=False,
metadata={
"help":
"Drop the last incomplete batch if it is not divisible by the batch size."
})
eval_steps: int = field(
default=None, metadata={"help": "Run an evaluation every X steps."})
dataloader_num_workers: int = field(
default=0,
metadata={
"help":
"Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process."
},
)
past_index: int = field(
default=-1,
metadata={
"help":
"If >=0, uses the corresponding part of the output as the past state for next step."
},
)
run_name: Optional[str] = field(
default=None, metadata={"help": "An optional descriptor for the run."})
device: Optional[str] = field(
default="gpu", metadata={"help": "select cpu, gpu, xpu devices."})
disable_tqdm: Optional[bool] = field(
default=None,
metadata={"help": "Whether or not to disable the tqdm progress bars."})
remove_unused_columns: Optional[bool] = field(
default=True,
metadata={
"help":
"Remove columns not required by the model when using an nlp.Dataset."
})
label_names: Optional[List[str]] = field(
default=None,
metadata={
"help":
"The list of keys in your dictionary of inputs that correspond to the labels."
})
load_best_model_at_end: Optional[bool] = field(
default=False,
metadata={
"help":
"Whether or not to load the best model found during training at the end of training."
},
)
metric_for_best_model: Optional[str] = field(
default=None,
metadata={
"help": "The metric to use to compare two different models."
})
greater_is_better: Optional[bool] = field(
default=None,
metadata={
"help":
"Whether the `metric_for_best_model` should be maximized or not."
})
ignore_data_skip: bool = field(
default=False,
metadata={
"help":
"When resuming training, whether or not to skip the first epochs and batches to get to the same training data."
},
)
optim: str = field(
default="adamw",
metadata={"help": "The optimizer to use."},
)
report_to: Optional[List[str]] = field(
default=None,
metadata={
"help":
"The list of integrations to report the results and logs to."
})
resume_from_checkpoint: Optional[str] = field(
default=None,
metadata={
"help":
"The path to a folder with a valid checkpoint for your model."
},
)
def __post_init__(self):
env_local_rank = int(os.environ.get("PADDLE_RANK_IN_NODE", -1))
if env_local_rank != -1 and env_local_rank != self.local_rank and paddle.distributed.get_world_size(
) > 1:
self.local_rank = env_local_rank
# convert to int
self.log_level = -1
self.log_level_replica = -1
# expand paths, if not os.makedirs("~/bar") will make directory
# in the current directory instead of the actual home
if self.output_dir is not None:
self.output_dir = os.path.expanduser(self.output_dir)
if self.logging_dir is None and self.output_dir is not None:
self.logging_dir = os.path.join(self.output_dir, default_logdir())
if self.logging_dir is not None:
self.logging_dir = os.path.expanduser(self.logging_dir)
if self.disable_tqdm is None:
self.disable_tqdm = False # logger.getEffectiveLevel() > logging.WARN
self.evaluation_strategy = IntervalStrategy(self.evaluation_strategy)
self.logging_strategy = IntervalStrategy(self.logging_strategy)
self.save_strategy = IntervalStrategy(self.save_strategy)
self.lr_scheduler_type = SchedulerType(self.lr_scheduler_type)
if self.do_eval is False and self.evaluation_strategy != IntervalStrategy.NO:
self.do_eval = True
if self.do_eval and self.evaluation_strategy == IntervalStrategy.NO:
logger.warning(
"evaluation_strategy reset to IntervalStrategy.STEPS for do_eval is True. you can also set evaluation_strategy='epoch'."
)
self.evaluation_strategy = IntervalStrategy.STEPS
# eval_steps has to be defined and non-zero, fallbacks to logging_steps if the latter is non-zero
if self.evaluation_strategy == IntervalStrategy.STEPS and (
self.eval_steps is None or self.eval_steps == 0):
if self.logging_steps > 0:
logger.info(
f"using `logging_steps` to initialize `eval_steps` to {self.logging_steps}"
)
self.eval_steps = self.logging_steps
else:
raise ValueError(
f"evaluation strategy {self.evaluation_strategy} requires either non-zero --eval_steps or --logging_steps"
)
# logging_steps must be non-zero for logging_strategy that is other than 'no'
if self.logging_strategy == IntervalStrategy.STEPS and self.logging_steps == 0:
raise ValueError(
f"logging strategy {self.logging_strategy} requires non-zero --logging_steps"
)
# Sanity checks for load_best_model_at_end: we require save and eval strategies to be compatible.
if self.load_best_model_at_end:
if self.evaluation_strategy != self.save_strategy:
raise ValueError(
"--load_best_model_at_end requires the save and eval strategy to match, but found\n- Evaluation "
f"strategy: {self.evaluation_strategy}\n- Save strategy: {self.save_strategy}"
)
if self.evaluation_strategy == IntervalStrategy.STEPS and self.save_steps % self.eval_steps != 0:
raise ValueError(
"--load_best_model_at_end requires the saving steps to be a round multiple of the evaluation "
f"steps, but found {self.save_steps}, which is not a round multiple of {self.eval_steps}."
)
if self.load_best_model_at_end and self.metric_for_best_model is None:
self.metric_for_best_model = "loss"
if self.greater_is_better is None and self.metric_for_best_model is not None:
self.greater_is_better = self.metric_for_best_model not in [
"loss", "eval_loss"
]
if self.run_name is None:
self.run_name = self.output_dir
self.optim = OptimizerNames(self.optim)
if self.report_to is None:
logger.info(
"The default value for the training argument `--report_to` will change in v5 (from all installed "
"integrations to none). In v5, you will need to use `--report_to all` to get the same behavior as "
"now. You should start updating your code and make this info disappear :-)."
)
self.report_to = "all"
if self.report_to == "all" or self.report_to == ["all"]:
# Import at runtime to avoid a circular import.
from .integrations import get_available_reporting_integrations
self.report_to = get_available_reporting_integrations()
elif self.report_to == "none" or self.report_to == ["none"]:
self.report_to = []
elif not isinstance(self.report_to, list):
self.report_to = [self.report_to]
if self.warmup_ratio < 0 or self.warmup_ratio > 1:
raise ValueError("warmup_ratio must lie in range [0,1]")
elif self.warmup_ratio > 0 and self.warmup_steps > 0:
logger.info(
"Both warmup_ratio and warmup_steps given, warmup_steps will override any effect of warmup_ratio during training"
)
def __str__(self):
self_as_dict = asdict(self)
self_as_dict = {
k: f"<{k.upper()}>" if k.endswith("_token") else v
for k, v in self_as_dict.items()
}
attrs_as_str = [f"{k}={v},\n" for k, v in sorted(self_as_dict.items())]
return f"{self.__class__.__name__}(\n{''.join(attrs_as_str)})"
__repr__ = __str__
@property
def train_batch_size(self) -> int:
"""
The actual batch size for training.
"""
train_batch_size = self.per_device_train_batch_size
return train_batch_size
@property
def eval_batch_size(self) -> int:
"""
The actual batch size for evaluation.
"""
eval_batch_size = self.per_device_eval_batch_size
return eval_batch_size
@property
def current_device(self) -> "paddle.device":
"""
The device used by this process.
"""
return paddle.device.get_device()
@property
def world_size(self):
"""
The number of processes used in parallel.
"""
if self.local_rank != -1:
return paddle.distributed.get_world_size()
return 1
@property
def process_index(self):
"""
The index of the current process used.
"""
if self.local_rank != -1:
return paddle.distributed.get_rank()
return 0
@property
def local_process_index(self):
"""
The index of the local process used.
"""
if self.local_rank != -1:
return self.local_rank
return 0
@property
def should_log(self):
"""
Whether or not the current process should produce log.
"""
if self.log_on_each_node:
return self.local_process_index == 0
else:
return self.process_index == 0
@property
def should_save(self):
"""
Whether or not the current process should write to disk, e.g., to save models and checkpoints.
"""
if self.save_on_each_node:
return self.local_process_index == 0
else:
return self.process_index == 0
@property
def _no_sync_in_gradient_accumulation(self):
"""
Whether or not to use no_sync for the gradients when doing gradient accumulation.
"""
return True
@contextlib.contextmanager
def main_process_first(self, local=True, desc="work"):
"""
A context manager for paddle distributed environment where on needs to do something on the main process, while
blocking replicas, and when it's finished releasing the replicas.
One such use is for `datasets`'s `map` feature which to be efficient should be run once on the main process,
which upon completion saves a cached version of results and which then automatically gets loaded by the
replicas.
Args:
local (`bool`, *optional*, defaults to `True`):
if `True` first means process of rank 0 of each node if `False` first means process of rank 0 of node
rank 0 In multi-node environment with a shared filesystem you most likely will want to use
`local=False` so that only the main process of the first node will do the processing. If however, the
filesystem is not shared, then the main process of each node will need to do the processing, which is
the default behavior.
desc (`str`, *optional*, defaults to `"work"`):
a work description to be used in debug logs
"""
if self.world_size > 1:
if local:
is_main_process = self.local_process_index == 0
main_process_desc = "main local process"
else:
is_main_process = self.process_index == 0
main_process_desc = "main process"
try:
if not is_main_process:
# tell all replicas to wait
logger.debug(
f"{self.process_index}: waiting for the {main_process_desc} to perform {desc}"
)
paddle.distributed.barrier()
yield
finally:
if is_main_process:
# the wait is over
logger.debug(
f"{self.process_index}: {main_process_desc} completed {desc}, releasing all replicas"
)
paddle.distributed.barrier()
else:
yield
def get_warmup_steps(self, num_training_steps: int):
"""
Get number of steps used for a linear warmup.
"""
warmup_steps = (self.warmup_steps if self.warmup_steps > 0 else
math.ceil(num_training_steps * self.warmup_ratio))
return warmup_steps
def to_dict(self):
"""
Serializes this instance while replace `Enum` by their values (for JSON serialization support). It obfuscates
the token values by removing their value.
"""
d = asdict(self)
for k, v in d.items():
if isinstance(v, Enum):
d[k] = v.value
if isinstance(v, list) and len(v) > 0 and isinstance(v[0], Enum):
d[k] = [x.value for x in v]
if k.endswith("_token"):
d[k] = f"<{k.upper()}>"
return d
def to_json_string(self):
"""
Serializes this instance to a JSON string.
"""
return json.dumps(self.to_dict(), indent=2)
def to_sanitized_dict(self) -> Dict[str, Any]:
"""
Sanitized serialization
"""
d = self.to_dict()
d = {
**d,
**{
"train_batch_size": self.train_batch_size,
"eval_batch_size": self.eval_batch_size
}
}
valid_types = [bool, int, float, str]
valid_types.append(paddle.Tensor)
return {
k: v if type(v) in valid_types else str(v)
for k, v in d.items()
}
def print_config(self, args=None, key=""):
"""
print all config values.
"""
logger.info("=" * 60)
if args is None:
args = self
key = "Training"
logger.info('{:^40}'.format("{} Configuration Arguments".format(key)))
logger.info('{:30}:{}'.format("paddle commit id",
paddle.version.commit))
for a in dir(args):
if a[:2] != "__": #don't print double underscore methods
v = getattr(args, a)
if not isinstance(v, types.MethodType):
logger.info('{:30}:{}'.format(a, v))
logger.info("")
class Variable:
"""
{begin_markdown Variable}
{spell_markdown init gprior}
# `curvefit.core.parameter.Variable`
## A variable to be estimated during the fit
A `Variable` is the most detailed unit to be estimated for curve fitting. A `Variable` corresponds
to an effect on a parameter -- and can contain a combination of both a fixed effect and random effects.
A `Variable` needs a `"covariate"`, but the covariate in the data can be just a column of 1's, in which
case a `Variable` is equivalent to a [`Parameter`](Parameter.md). If instead the values of
the `"covariate"` argument differ for different rows of the data, `Variable` multiplies the covariate
to get the `Parameter` for that data row.
A `curvefit` model is made up of multiple parameters. For more information, see
[`Parameter`](Parameter.md) and [`ParameterSet`](ParameterSet.md).
## Arguments
- `covariate (str)`: name of the covariate for this variable (corresponds to what it will be in the data
that is eventually used to fit the model)
- `var_link_fun (Callable)`: link function for the variable
- `fe_init (float)`: initial value to be used in the optimization for the fixed effect
- `re_init (float)`: initial value to be used in the optimization for the random effect
- `re_zero_sum_std (float)`: standard deviation of the zero sum prior
for he random effects corresponding to this variable.
- `fe_gprior (optional, List[float])`: list of Gaussian priors
the fixed effect where the first element is the prior
mean and the second element is the prior standard deviation
- `re_gprior (optional, List[float])`: list of Gaussian priors
the random effect where the first element is the prior
mean and the second element is the prior standard deviation
- `fe_bounds (optional, List[float])`: list of box constraints
for the fixed effects during the optimization where the first element is the lower bound
and the second element is the upper bound
- `re_bounds (optional, List[float])`: list of box constraints
for the fixed effects during the optimization where the first element is the lower bound
and the second element is the upper bound
## Usage
```python
from curvefit.core.parameter import Variable
var = Variable(covariate='ones', var_link_fun=lambda x: x, fe_init=0., re_init=0.)
```
{end_markdown Variable}
"""
covariate: str
var_link_fun: Callable
fe_init: float
re_init: float
re_zero_sum_std: float = field(default=np.inf)
fe_gprior: List[float] = field(default_factory=lambda: [0.0, np.inf])
re_gprior: List[float] = field(default_factory=lambda: [0.0, np.inf])
fe_bounds: List[float] = field(default_factory=lambda: [-np.inf, np.inf])
re_bounds: List[float] = field(default_factory=lambda: [-np.inf, np.inf])
def __post_init__(self):
assert isinstance(self.covariate, str)
assert len(self.fe_gprior) == 2
assert len(self.re_gprior) == 2
assert len(self.fe_bounds) == 2
assert len(self.re_bounds) == 2
assert self.fe_gprior[1] > 0.0
assert self.re_gprior[1] > 0.0
class TrainingArguments:
output_dir: str = field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."},
)
overwrite_output_dir: bool = field(
default=False,
metadata={
"help": (
"Overwrite the content of the output directory. "
"Use this to continue training if output_dir points to a checkpoint directory."
)
},
)
do_train: bool = field(default=False, metadata={"help": "Whether to run training."})
do_eval: bool = field(default=False, metadata={"help": "Whether to run eval on the dev set."})
do_predict: bool = field(default=False, metadata={"help": "Whether to run predictions on the test set."})
per_device_train_batch_size: int = field(
default=8, metadata={"help": "Batch size per GPU/TPU core/CPU for training."}
)
per_device_eval_batch_size: int = field(
default=8, metadata={"help": "Batch size per GPU/TPU core/CPU for evaluation."}
)
learning_rate: float = field(default=5e-5, metadata={"help": "The initial learning rate for AdamW."})
weight_decay: float = field(default=0.0, metadata={"help": "Weight decay for AdamW if we apply some."})
adam_beta1: float = field(default=0.9, metadata={"help": "Beta1 for AdamW optimizer"})
adam_beta2: float = field(default=0.999, metadata={"help": "Beta2 for AdamW optimizer"})
adam_epsilon: float = field(default=1e-8, metadata={"help": "Epsilon for AdamW optimizer."})
adafactor: bool = field(default=False, metadata={"help": "Whether or not to replace AdamW by Adafactor."})
num_train_epochs: float = field(default=3.0, metadata={"help": "Total number of training epochs to perform."})
warmup_steps: int = field(default=0, metadata={"help": "Linear warmup over warmup_steps."})
logging_steps: int = field(default=500, metadata={"help": "Log every X updates steps."})
save_steps: int = field(default=500, metadata={"help": "Save checkpoint every X updates steps."})
eval_steps: int = field(default=None, metadata={"help": "Run an evaluation every X steps."})
seed: int = field(default=42, metadata={"help": "Random seed that will be set at the beginning of training."})
push_to_hub: bool = field(
default=False, metadata={"help": "Whether or not to upload the trained model to the model hub after training."}
)
hub_model_id: str = field(
default=None, metadata={"help": "The name of the repository to keep in sync with the local `output_dir`."}
)
hub_token: str = field(default=None, metadata={"help": "The token to use to push to the Model Hub."})
def __post_init__(self):
if self.output_dir is not None:
self.output_dir = os.path.expanduser(self.output_dir)
def to_dict(self):
"""
Serializes this instance while replace `Enum` by their values (for JSON serialization support). It obfuscates
the token values by removing their value.
"""
d = asdict(self)
for k, v in d.items():
if isinstance(v, Enum):
d[k] = v.value
if isinstance(v, list) and len(v) > 0 and isinstance(v[0], Enum):
d[k] = [x.value for x in v]
if k.endswith("_token"):
d[k] = f"<{k.upper()}>"
return d
class StructuralPred:
"""Class that organises Structural Predictions output for single protein
Attributes
----------
DIS_threshold : float with values between 0. and 1. (default=0.50)
Threshold used for disorder class definition of All predictors.
ACC_threshold: probability threshold for Scratch1D (only) rellative
solvent prediction class definition, default=0.20
RaptorXDir : Path
Path of the folder where RaptorX prediction output is being stored.
PsiPredDir : Path
Path of the folder where PsiPred prediction output is being stored.
DisoPredDir : Path
Path of the folder where DisoPred prediction output is being stored.
Scratch1dDir : Path
Path of the folder where Scratch1d prediction output is being stored.
RaptorXPred : RaptorX object
Contains parsed predictions output of RaptorX
Scratch1dPred : Scratch1D object
Contains parsed predictions output of Scratch1D
PsiPredPred : PsiPred object
Contains parsed predictions output of PsiPred
DisoPredPred : DisoPred object
Contains parsed predictions output of DisoPred
Public Methods
--------------
parsestructural()
Parses all the prediction output files and add the data inside the
above attribute data structures.
printSingleProtVertical( self )
Prints all structural predictors in a vertical layout
"""
DIS_threshold: float = 0.5
ACC_threshold: float = 0.2
predictions: dict = field(default_factory=dict)
paths: dict = field(default_factory=dict)
@staticmethod
def parseall(paths: dict,
DIS_threshold: float = None,
ACC_threshold: float = None) -> StructuralPred:
"""
Parses all the prediction output files and add the data inside the
above attribute data structures.
"""
data = StructuralPred()
data.DIS_threshold = DIS_threshold if (DIS_threshold is not None and 0.0 < DIS_threshold < 1.0) \
else 0.50
data.ACC_threshold = ACC_threshold if (ACC_threshold is not None and 0.0 < ACC_threshold < 1.0) \
else 0.20
for prot in paths:
for predictor in paths[prot]:
if predictor.lower() == "raptorx":
preddata = RaptorXData.parse(paths[prot][predictor])
elif predictor.lower() == "psipred":
preddata = PsiPredData.parse(paths[prot][predictor])
elif predictor.lower() == "disopred":
preddata = DisoPredData.parse(paths[prot][predictor])
elif predictor.lower() == "scratch1d":
preddata = Scratch1dData.parse(paths[prot][predictor])
else:
print("Unknown predictor key: ", predictor)
raise
if prot not in data.predictions:
data.predictions[prot] = {}
data.predictions[prot][predictor] = preddata
return data
def printSingleProtVertical(self: StructuralPred, protname: str):
try:
data = self.predictions[protname]
# Print header
print("#Protname: ", protname, sep='\t')
print("#DIS threshold: ", self.DIS_threshold, sep='\t')
print("#ACC threshold (Scratch only): ",
self.ACC_threshold,
sep='\t')
print("\n")
print("#resid", "aa", \
"_", "RaptorX_SS3", "Scratch1D_SS3", "PsiPred_SS3", \
"_", "RaptorX_SS8", "Scratch1D_SS8", \
"_", "RaptorX_ACC3", "Scratch1D__ACC2", \
"_", "RaptorX_DIS", "DisoPred_DIS", sep='\t')
seq = data["raptorx"].sequence
protsize = len(seq)
for id in range(protsize):
print(id + 1, seq[id], \
"_", data["raptorx"].SS3_classes[id], data["scratch1d"].SS3_classes[id], \
data["psipred"].SS3_classes[id], \
"_", data["raptorx"].SS8_classes[id], data["scratch1d"].SS8_classes[id], \
"_", data["raptorx"].ACC3_classes[id], data["scratch1d"].ACC2_classes[id], \
"_", data["raptorx"].DIS_classes[id], data["disopred"].DIS_classes[id], sep='\t')
except:
print("Unexpected error:", sys.exc_info()[0])
raise
class RestEndpointDelete:
"""
This meta-class represents the ability to model a REST endpoint with DELETE
semantics.
:ivar short_name: This specifies an identifying shortName for the
object. It needs to be unique within its context and is intended
for humans but even more for technical reference.
:ivar short_name_fragments: This specifies how the
Referrable.shortName is composed of several shortNameFragments.
:ivar long_name: This specifies the long name of the object. Long
name is targeted to human readers and acts like a headline.
:ivar desc: This represents a general but brief (one paragraph)
description what the object in question is about. It is only one
paragraph! Desc is intended to be collected into overview
tables. This property helps a human reader to identify the
object in question. More elaborate documentation, (in particular
how the object is built or used) should go to "introduction".
:ivar category: The category is a keyword that specializes the
semantics of the Identifiable. It affects the expected existence
of attributes and the applicability of constraints.
:ivar admin_data: This represents the administrative data for the
identifiable object.
:ivar introduction: This represents more information about how the
object in question is built or is used. Therefore it is a
DocumentationBlock.
:ivar annotations: Possibility to provide additional notes while
defining a model element (e.g. the ECU Configuration Parameter
Values). These are not intended as documentation but are mere
design notes.
:ivar blueprint_policys: This role indicates whether the
blueprintable element will be modifiable or not motifiable.
:ivar short_name_pattern: This attribute represents the pattern
which shall be used to build the shortName of the derived
elements. As of now it is modeled as a String. In general it
should follow the pattern: pattern = (placeholder | namePart)*
placeholder = "{" namePart "}" namePart = identifier | "_" This
is subject to be refined in subsequent versions. Note that this
is marked as obsolete. Use the xml attribute namePattern instead
as it applies to Identifier and CIdentifier (shortName, symbol
etc.)
:ivar arguments: Some endpoints can require a list of arguments.
:ivar s: Checksum calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine if
an ArObject has changed. The checksum has no semantic meaning
for an AUTOSAR model and there is no requirement for AUTOSAR
tools to manage the checksum.
:ivar t: Timestamp calculated by the user's tool environment for an
ArObject. May be used in an own tool environment to determine
the last change of an ArObject. The timestamp has no semantic
meaning for an AUTOSAR model and there is no requirement for
AUTOSAR tools to manage the timestamp.
:ivar uuid: The purpose of this attribute is to provide a globally
unique identifier for an instance of a meta-class. The values of
this attribute should be globally unique strings prefixed by the
type of identifier. For example, to include a DCE UUID as
defined by The Open Group, the UUID would be preceded by "DCE:".
The values of this attribute may be used to support merging of
different AUTOSAR models. The form of the UUID (Universally
Unique Identifier) is taken from a standard defined by the Open
Group (was Open Software Foundation). This standard is widely
used, including by Microsoft for COM (GUIDs) and by many
companies for DCE, which is based on CORBA. The method for
generating these 128-bit IDs is published in the standard and
the effectiveness and uniqueness of the IDs is not in practice
disputed. If the id namespace is omitted, DCE is assumed. An
example is "DCE:2fac1234-31f8-11b4-a222-08002b34c003". The uuid
attribute has no semantic meaning for an AUTOSAR model and there
is no requirement for AUTOSAR tools to manage the timestamp.
"""
class Meta:
name = "REST-ENDPOINT-DELETE"
short_name: Optional[Identifier] = field(
default=None,
metadata={
"name": "SHORT-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
"required": True,
})
short_name_fragments: Optional[
"RestEndpointDelete.ShortNameFragments"] = field(
default=None,
metadata={
"name": "SHORT-NAME-FRAGMENTS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
long_name: Optional[MultilanguageLongName] = field(
default=None,
metadata={
"name": "LONG-NAME",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
desc: Optional[MultiLanguageOverviewParagraph] = field(
default=None,
metadata={
"name": "DESC",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
category: Optional[CategoryString] = field(
default=None,
metadata={
"name": "CATEGORY",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
admin_data: Optional[AdminData] = field(
default=None,
metadata={
"name": "ADMIN-DATA",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
introduction: Optional[DocumentationBlock] = field(
default=None,
metadata={
"name": "INTRODUCTION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
annotations: Optional["RestEndpointDelete.Annotations"] = field(
default=None,
metadata={
"name": "ANNOTATIONS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
blueprint_policys: Optional["RestEndpointDelete.BlueprintPolicys"] = field(
default=None,
metadata={
"name": "BLUEPRINT-POLICYS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
short_name_pattern: Optional[String] = field(
default=None,
metadata={
"name": "SHORT-NAME-PATTERN",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
arguments: Optional["RestEndpointDelete.Arguments"] = field(
default=None,
metadata={
"name": "ARGUMENTS",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
s: Optional[str] = field(default=None,
metadata={
"name": "S",
"type": "Attribute",
})
t: Optional[str] = field(
default=None,
metadata={
"name":
"T",
"type":
"Attribute",
"pattern":
r"([0-9]{4}-[0-9]{2}-[0-9]{2})(T[0-9]{2}:[0-9]{2}:[0-9]{2}(Z|([+\-][0-9]{2}:[0-9]{2})))?",
})
uuid: Optional[str] = field(default=None,
metadata={
"name": "UUID",
"type": "Attribute",
})
@dataclass
class ShortNameFragments:
short_name_fragment: List[ShortNameFragment] = field(
default_factory=list,
metadata={
"name": "SHORT-NAME-FRAGMENT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
@dataclass
class Annotations:
annotation: List[Annotation] = field(
default_factory=list,
metadata={
"name": "ANNOTATION",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
@dataclass
class BlueprintPolicys:
blueprint_policy_list: List[BlueprintPolicyList] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-LIST",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
blueprint_policy_not_modifiable: List[
BlueprintPolicyNotModifiable] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-NOT-MODIFIABLE",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
blueprint_policy_single: List[BlueprintPolicySingle] = field(
default_factory=list,
metadata={
"name": "BLUEPRINT-POLICY-SINGLE",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
@dataclass
class Arguments:
rest_endpoint_argument: List[RestEndpointArgument] = field(
default_factory=list,
metadata={
"name": "REST-ENDPOINT-ARGUMENT",
"type": "Element",
"namespace": "http://autosar.org/schema/r4.0",
})
class ModelArguments:
"""
Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
"""
model_name_or_path: Optional[str] = field(
default=None,
metadata={
"help":
("The model checkpoint for weights initialization.Don't set if you want to train a model from scratch."
)
},
)
model_type: Optional[str] = field(
default=None,
metadata={
"help":
"If training from scratch, pass a model type from the list: " +
", ".join(MODEL_TYPES)
},
)
config_overrides: Optional[str] = field(
default=None,
metadata={
"help":
("Override some existing default config settings when a model is trained from scratch. Example: "
"n_embd=10,resid_pdrop=0.2,scale_attn_weights=false,summary_type=cls_index"
)
},
)
config_name: Optional[str] = field(
default=None,
metadata={
"help":
"Pretrained config name or path if not the same as model_name"
})
tokenizer_name: Optional[str] = field(
default=None,
metadata={
"help":
"Pretrained tokenizer name or path if not the same as model_name"
})
cache_dir: Optional[str] = field(
default=None,
metadata={
"help":
"Where do you want to store the pretrained models downloaded from huggingface.co"
},
)
use_fast_tokenizer: bool = field(
default=True,
metadata={
"help":
"Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."
},
)
model_revision: str = field(
default="main",
metadata={
"help":
"The specific model version to use (can be a branch name, tag name or commit id)."
},
)
use_auth_token: bool = field(
default=False,
metadata={
"help":
("Will use the token generated when running `transformers-cli login` (necessary to use this script "
"with private models).")
},
)
def __post_init__(self):
if self.config_overrides is not None and (self.config_name is not None
or self.model_name_or_path
is not None):
raise ValueError(
"--config_overrides can't be used in combination with --config_name or --model_name_or_path"
)
class CoctMt090003Uv01Person:
class Meta:
name = "COCT_MT090003UV01.Person"
realm_code: List[Cs] = field(
default_factory=list,
metadata={
"name": "realmCode",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
type_id: Optional[Ii] = field(
default=None,
metadata={
"name": "typeId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
template_id: List[Ii] = field(
default_factory=list,
metadata={
"name": "templateId",
"type": "Element",
"namespace": "urn:hl7-org:v3",
}
)
name: List[EnExplicit] = field(
default_factory=list,
metadata={
"type": "Element",
"namespace": "urn:hl7-org:v3",
"min_occurs": 1,
}
)
null_flavor: Optional[NullFlavor] = field(
default=None,
metadata={
"name": "nullFlavor",
"type": "Attribute",
}
)
class_code: EntityClass = field(
init=False,
default=EntityClass.PSN,
metadata={
"name": "classCode",
"type": "Attribute",
"required": True,
}
)
determiner_code: EntityDeterminer = field(
init=False,
default=EntityDeterminer.INSTANCE,
metadata={
"name": "determinerCode",
"type": "Attribute",
"required": True,
}
)
def field(number: int, *args, **kwargs) -> dataclasses.Field:
"""
Convenience method to assign field numbers.
Calls the standard ``dataclasses.field`` function with the metadata assigned.
"""
return dataclasses.field(*args, metadata={'number': number}, **kwargs)
class Spectrum(SpectralData):
r"""Create a spectrum object.
A spectrum array is a subclass of a record or structured numpy array,
with one axis representing the wavelength vector (w) and other the
reflectance (r). The optional
Attributes
----------
w: numpy array
array corresponding to the wavelength vector
r: numpy array
array corresponding to the relative reflectance of
the asteroid
unit: str
The wavelength units. Default is 'microns'.
r_unc: numpy array (optional)
array corresponding to the relative reflectance
uncertainty of the asteroid
path: None or str (optional)
The path of the spectrum file
label: None or str
The spectrum label
res: float
The spectrum resolution (number of points).
Methods
-------
trim
fit
autofit
estimate_rms
clean_spec
mad
rebin
normalize
mask_region
save
plot
"""
w: np.ndarray
r: np.ndarray
r_unc: np.ndarray = field(default=None)
label: str = field(default='asteroid')
unit: str = field(default='micron')
def __post_init__(self):
r"""Inicialize class."""
self.w = np.array(self.w, ndmin=1)
self.r = np.array(self.r, ndmin=1)
assert self.w.size == self.r.size
if self.r_unc is None:
dtype = ('w', 'r')
else:
self.r_unc = np.array(self.r_unc, ndmin=1)
assert self.r_unc.size == self.r.size
dtype = ('w', 'r', 'r_unc')
SpectralData.__init__(self, self.w, dtype, self.label, unit=self.unit)
def angstrom2micron(self):
r"""Convert wavenlength axis from angstrom to micron."""
self.w = self.w / 10000.0
self.unit = 'micron'
return self
def micron2angstrom(self):
r"""Convert wavenlength axis from micron to angstrom."""
self.w = self.w * 10000.
self.unit = 'angstrom'
return self
def fit(self, order=4, ftype='spline'):
r"""
Fit the spectrum using a polynomial or a smooth spline.
Parameters
----------
order: int
Order of the fitting.
ftype: str
Type of algorithm to use for the fitting.
Options are: 'spline' or 'polynomial'.
Returns
-------
fspec: Spectrum Object
The fitted spectrum array
fcoefs: array-like
the fitting coefficients
"""
# Performing the fit
if ftype == 'spline':
fcoefs = UnivariateSpline(self.w, self.r, k=order)
fspec_y = fcoefs(self.w)
elif ftype == 'polynomial':
fcoefs = np.polyfit(self.w, self.r, order)
fspec_y = np.polyval(fcoefs, self.w)
y_err = np.abs(fspec_y - self.r)
# building new array
fspec = self.__class__(w=self.w, r=fspec_y, r_unc=y_err, unit=self.unit,
label=self.label + '_fit')
return fspec, fcoefs
def autofit(self, degree_min=1, degree_max=12):
r"""
Find the best order of the polynomial fitting of the spectrum.
Parameters
----------
degree_min: int
The minimal order for a fit
degree_max: int
The maximal order for a fit
Returns
-------
fspec: Spectrum Object
The fitted spectrum array
fcoefs: array-like
the fitting coefficients
"""
# degrees which we will test the fitting
degrees = np.arange(degree_min, degree_max+1)
# calculation cross-validation score and error for each degree
cross, _ = np.array([self._autofit_aux(deg) for deg in degrees]).T
# getting minimun cross validation score
aux = np.argmin(cross)
bestfit = degrees[aux]
# fitting spectrum
fspec, fcoefs = self.fit(order=bestfit, ftype='polynomial')
return fspec, fcoefs
def _autofit_aux(self, degree):
r"""Auxiliary funtion for the autofit method."""
# creating polynomial and reshaping array for model validation
polynom = PolynomialFeatures(degree=degree, include_bias=False)
x_train = self.w.reshape((-1, 1))
x_train_trans = polynom.fit_transform(x_train)
# Create the linear regression model and train
model = LinearRegression()
model.fit(x_train_trans, self.r)
# Calculate the cross validation score
cross_valid = cross_val_score(model, x_train_trans, self.r,
scoring='neg_mean_squared_error', cv=10)
# Training predictions and error
y_train_predictions = model.predict(x_train_trans)
training_error = mean_squared_error(self.r, y_train_predictions)
return -np.mean(cross_valid), training_error
def estimate_rms(self, ftype='auto', order=5):
r"""
Estimate the signal-to-noise ratio in a spectrum.
Parameters
----------
ftype: str
Type of fitting for the snr estimation. Options are:
'spline', 'polynomial', 'auto'. Default is 'auto'.
order: int
Order of the adjust. Ignored if ftype is 'auto'.
Returns
-------
rms: float
The estimated snr value
"""
# fitting the spectrum
if ftype == 'auto':
fspec, _ = self.autofit()
else:
fspec, _ = self.fit(order, ftype)
# Estimating the SNR
std_arr = np.abs(self.r - fspec.r)
rms = np.std(std_arr)
return rms
def clean_spec(self, method='sigmaclip', sigma=3, fit='auto'):
r"""
Remove outliers from the spectrum.
Parameters
----------
method: str
Method for detecting outliers. Currently only 'sigmaclip' available
Default is 'sigmaclip'.
sigma: int
Remove points higher than sigma.
fit: 'auto' or integer
The order of the polynomial fit. If auto it will try to find
automaticaly. Default is 'auto'.
Returns
-------
spec: Spectrum
Spectrum object with outliers removed.
"""
if fit == 'auto':
fspec, _ = self.autofit()
else:
fspec, _ = self.fit(order=fit, ftype='polynomial')
if method == 'sigmaclip':
cspec = np.divide(self.r, fspec.r)
cspec_index = [self._sigma_clip(val, sigma=sigma, cspec=cspec)
for val in cspec]
aux = self[cspec_index]
spec = self.__class__(aux.w, aux.r, r_unc=aux.r_unc, unit=self.unit,
label=self.label + '_cleaned')
return spec
def _sigma_clip(self, val, sigma, cspec):
r"""Auxiliary method to perform sigma-clipping on array elements."""
if (np.median(cspec) - self._mad(cspec) * sigma < val) and \
(val < np.median(cspec) + self._mad(cspec)*sigma):
return True
return False
def mad(self, axis=None):
r"""
Calculate the median absolute deviation.
Parameters
----------
axis: str
'wave', 'ref' or None.
It will return the mad in the defined axis.
If None, than returns the mad in both axis
Returns
-------
The median absolute deviation
"""
if axis is not None:
return self._mad(axis)
return self._mad('w'), self._mad('r')
@staticmethod
def _mad(arr):
r"""Auxiliary function for calculating the MAD."""
return np.median(np.abs(arr - np.median(arr)))
def rebin(self, binsize=11, method='median', std=True,
rem_trend=False):
r"""
Rebin the spectrum.
Parameters
----------
binsize: int
The number of points in the bin.
method: str
The method for the rebinning.
Options are:'mean and 'median'.
std: boolean
If True, also returns the deviation.
In the case of the median, returns the
MAD (median absolute deviation).
rem_trend: boolean
Returns
-------
The rebined spectrum
"""
spec_size = len(self.w)
y_stop = (spec_size//binsize)*binsize
wave_arr = self.w[:y_stop]
ref_arr = self.r[:y_stop]
if method == 'median':
func = np.median
std_func = np.std # ->>>>>>change later
if method == 'mean':
func = np.mean
std_func = np.std
wave_reb = func(wave_arr.reshape(spec_size // binsize, binsize),
axis=-1).T
ref_reb = func(ref_arr.reshape(spec_size // binsize, binsize),
axis=-1).T
if rem_trend:
fspec = self.autofit()[0]
ref_arr = np.divide(ref_arr, fspec.r[:y_stop])
if std:
std = std_func(ref_arr.reshape(spec_size // binsize, binsize),
axis=-1).T
std = np.array(std)
else:
std = None
return self.__class__(w=wave_reb, r=ref_reb, r_unc=std, unit=self.unit,
label=self.label + '_binned')
def ref_from_wavelength(self, w, interpolate=True):
r"""
Get the spectrum reflectance in a particular wavelength.
Parameters
----------
w: float
Wavelength value
If interpolate=False, The code will search the closest
value.
interpolate: boolean (optional)
If interpolate=False, The code will search the closest
value. If True it will interpolate the value of w.
Returns
-------
The reflectance value
"""
if not interpolate:
aux = find_nearest(self.w, w)[0]
ref = self.r[aux]
else:
ref = np.interp(w, self.w, self.r)
return ref
def normalize(self, wnorm=0.55, window=None, interpolate=True):
r"""
Normalize the spectrum in a particular wavelength.
Parameters
----------
wnorm: float
Wavelength value to normalize the spectrum.
If interpolate=False, The code will search the closest
value.
window: None or float (optional)
The wavelenght window size for normalizing.
If None it will normalize in the wnorm point only.
interpolate: boolean (optional)
If interpolate=False, The code will search the closest
value. If True it will interpolate the value of wnorm.
Returns
-------
The normalized Spectrum
"""
if window is None:
norm_factor = self.ref_from_wavelength(wnorm,
interpolate=interpolate)
else:
aux = np.argwhere((self.w > wnorm-window) &
(self.w < wnorm+window))
norm_factor = np.mean(self.r[aux])
self.r = self.r / norm_factor
if self.r_unc is not None:
self.r_unc = self.r_unc / norm_factor
return self
def mask_region(self, region=[(1.3, 1.45), (1.8, 1.95)]):
r"""
Exclude a region of the spectrum.
Parameters
----------
w_min: float
Wavelength lower limit of the masked region
w_max: float
Wavelength upper limit of the masked region
Returns
-------
masked_spec: Spectrum
The Spectrum array without the masked region
"""
if isinstance(region[0], (float, int)):
masked_spec = self.mask_region_aux(self, wmin=region[0],
wmax=region[1])
else:
masked_spec = self
for rr in region:
masked_spec = self.mask_region_aux(masked_spec,
wmin=rr[0],
wmax=rr[1])
return masked_spec
def mask_region_aux(self, spec, wmin, wmax):
r"""
Exclude a region of the spectrum.
Parameters
----------
w_min: float
Wavelength lower limit of the masked region
w_max: float
Wavelength upper limit of the masked region
Returns
-------
The Spectrum array without the masked region
"""
aux = np.argwhere((spec.w > wmin) & (spec.w < wmax))
mask = np.ones(len(spec.w), dtype=bool)
mask[aux] = 0
w = spec.w[mask]
r = spec.r[mask]
r_unc = spec.r_unc
if r_unc is not None:
r_unc = spec.r_unc[mask]
return self.__class__(w=w, r=r, r_unc=r_unc, unit=spec.unit,
label=spec.label)
def plot(self, fax=None, show=False, savefig=None,
axistitles=True, speckwargs=None, legendkwargs=None):
r"""
Quick plot of the Spectrum.
Parameters
----------
fax (Optional): matplotlib.axes
If desired to subplot image in a figure. Default is 'None', which
will open a new plt.figure()
show (Optional): boolean
True if want to plt.show(). Default is True.
savefig (Optional): str
The path to save the figure. If set to None, wont save the figure.
Default is None
axistitles: boolean
If True will label the axis. Default is True.
speckwargs: dict
Arguments for matplotlib plot function.
default values: {'c':'0.9', 'lw':'1'}.
legendkwargs: dict
Arguments for matplotlib legend function.
default values: {'loc':'best'}.
Returns
-------
the matplotlib.axes of the figure
"""
# checking if plot in another frame
if fax is None:
fig = plt.figure()
fax = plt.gca()
# setting default values for image plot with matplotlib
specsty_defaults = {'c': '0.1', 'lw': 1}
legendsty_defaults = {'loc': 'best'}
# updating plot styles
speckwargs = kwargupdate(specsty_defaults, speckwargs)
legendkwargs = kwargupdate(legendsty_defaults, legendkwargs)
# Ploting the spec
fax.errorbar(self.w, self.r, yerr=self.r_unc, **speckwargs)
# Checking if desired to plot the axis labels
if axistitles:
if self.unit == 'micron':
unit_label = '$\mu$m'
fax.set_xlabel('Wavelength (%s)' % unit_label)
fax.set_ylabel('Reflectance')
# plot legend?
if 'label' in speckwargs:
fax.legend(**legendkwargs)
# check if save the image
if savefig is not None:
plt.savefig(savefig)
if not show:
plt.clf()
matplotlib.use('TkAgg')
# show in the matplotlib window?
if show:
plt.show()
