common.py•4.95 kB
import datetime
from enum import Enum
from typing import Annotated
try:
from pydantic import BaseModel, BeforeValidator, Field
except ImportError:
from pydantic import BaseModel, Field
# For pydantic v1 compatibility, create dummy BeforeValidator
def BeforeValidator(func):
return func
# Helper for probability distributions (list of floats summing to 1.0)
def _validate_probability_distribution(v: list[float]) -> list[float]:
"""
Validates that a list of floats represents a probability distribution.
Checks that all values are between 0.0 and 1.0. Allows empty lists and does not enforce that the sum equals 1.0, assuming normalization may occur elsewhere.
Args:
v: List of floats to validate.
Returns:
The validated list of floats if all values are within the valid range.
Raises:
ValueError: If any value is outside the range [0.0, 1.0].
"""
if not v: # Empty list is valid if optional, but if provided, must sum to 1
return v
if not all(0.0 <= p <= 1.0 for p in v):
raise ValueError("All probabilities must be between 0.0 and 1.0")
if abs(sum(v) - 1.0) > 1e-6 and sum(v) != 0: # Allow sum to be 0 for uninitialized
# For flexibility, we might allow non-normalized lists and normalize them later
# For now, strict validation. Could also add a normalization step here.
# raise ValueError("Probabilities must sum to 1.0")
pass # Relaxing this for now, assume normalization happens elsewhere if critical
return v
ProbabilityDistribution = Annotated[
list[float], BeforeValidator(_validate_probability_distribution)
]
# Confidence Vector based on P1.5
# Using a class for better type hinting and potential methods later
class ConfidenceVector(BaseModel):
empirical_support: float = Field(
default=0.5, description="Must be between 0.0 and 1.0"
)
theoretical_basis: float = Field(
default=0.5, description="Must be between 0.0 and 1.0"
)
methodological_rigor: float = Field(
default=0.5, description="Must be between 0.0 and 1.0"
)
consensus_alignment: float = Field(
default=0.5, description="Must be between 0.0 and 1.0"
)
def to_list(self) -> list[float]:
"""
Returns the confidence vector as a list of four float values in a fixed order.
"""
return [
self.empirical_support,
self.theoretical_basis,
self.methodological_rigor,
self.consensus_alignment,
]
@classmethod
def from_list(cls, values: list[float]) -> "ConfidenceVector":
"""
Creates a ConfidenceVector instance from a list of four float values.
Args:
values: A list of four floats representing empirical support, theoretical basis, methodological rigor, and consensus alignment, in that order.
Returns:
A ConfidenceVector initialized with the provided values.
Raises:
ValueError: If the input list does not contain exactly four elements.
"""
if len(values) != 4:
raise ValueError("Confidence list must have exactly 4 values.")
return cls(
empirical_support=values[0],
theoretical_basis=values[1],
methodological_rigor=values[2],
consensus_alignment=values[3],
)
@property
def average_confidence(self) -> float:
return sum(self.to_list()) / 4.0
# Single scalar certainty/confidence if needed
CertaintyScore = Annotated[float, Field(description="Value between 0.0 and 1.0")]
# Impact Score (P1.28)
ImpactScore = Annotated[
float, Field(description="Value between 0.0 and 1.0")
] # Assuming normalized impact
class TimestampedModel(BaseModel):
created_at: datetime.datetime
updated_at: datetime.datetime
def __init__(self, **data):
now = datetime.datetime.now()
if "created_at" not in data:
data["created_at"] = now
if "updated_at" not in data:
data["updated_at"] = now
super().__init__(**data)
def touch(self):
"""Updates the updated_at timestamp."""
self.updated_at = datetime.datetime.now()
# Standardized way to represent a probability distribution for discrete outcomes
class DiscreteProbabilityDistribution(BaseModel):
outcomes: list[str] # The labels for each probability
probabilities: ProbabilityDistribution
class EpistemicStatus(str, Enum):
"""P1.12: Epistemic status of a node/claim."""
ASSUMPTION = "assumption"
HYPOTHESIS = "hypothesis"
EVIDENCE_SUPPORTED = "evidence_supported"
EVIDENCE_CONTRADICTED = "evidence_contradicted"
THEORETICALLY_DERIVED = "theoretically_derived"
WIDELY_ACCEPTED = "widely_accepted"
DISPUTED = "disputed"
UNKNOWN = "unknown"
INFERRED = "inferred"
SPECULATION = "speculation"