"""Validation result data models."""
from datetime import datetime
from pydantic import Field
from .base import BaseModel, ConfidenceScore
class DateField(BaseModel):
"""A date field extracted from a document."""
field_name: str
value: datetime
source: str # "DOC-123, Page 5"
document_id: str
confidence: ConfidenceScore
class DateAlignmentValidation(BaseModel):
"""Result of date alignment validation between two fields."""
validation_id: str
validation_type: str = "date_alignment"
date1: DateField
date2: DateField
delta_days: int
max_allowed_days: int
status: str # "pass", "fail", "warning"
message: str
flagged_for_review: bool = False
class LandTenureField(BaseModel):
"""Land tenure information extracted from a document."""
owner_name: str
area_hectares: float | None = None
tenure_type: str | None = None # "lease", "ownership", "easement"
source: str # "DOC-123, Page 8"
document_id: str
confidence: ConfidenceScore
class LandTenureValidation(BaseModel):
"""Result of cross-document land tenure validation."""
validation_id: str
validation_type: str = "land_tenure"
fields: list[LandTenureField]
owner_name_match: bool
owner_name_similarity: float = Field(..., ge=0.0, le=1.0)
area_consistent: bool
tenure_type_consistent: bool
status: str # "pass", "fail", "warning"
message: str
discrepancies: list[str] = Field(default_factory=list)
flagged_for_review: bool = False
class ProjectIDOccurrence(BaseModel):
"""Single occurrence of a project ID in a document."""
project_id: str
document_id: str
document_name: str
page: int | None = None
section: str | None = None
class ProjectIDValidation(BaseModel):
"""Result of project ID consistency validation."""
validation_id: str
validation_type: str = "project_id"
expected_pattern: str # r"^C\d{2}-\d+$"
found_ids: list[str]
primary_id: str | None = None
occurrences: list[ProjectIDOccurrence]
total_occurrences: int
documents_with_id: int
total_documents: int
status: str # "pass", "fail", "warning"
message: str
flagged_for_review: bool = False
class ContradictionCheck(BaseModel):
"""Result of checking for contradictions between two pieces of information."""
validation_id: str
validation_type: str = "contradiction"
field_name: str
value1: str
source1: str
value2: str
source2: str
is_contradiction: bool
severity: str # "high", "medium", "low"
message: str
flagged_for_review: bool = False
class ValidationSummary(BaseModel):
"""Summary statistics for validation results."""
total_validations: int
validations_passed: int
validations_failed: int
validations_warning: int
items_flagged: int
pass_rate: float = Field(..., ge=0.0, le=1.0)
extraction_method: str = "regex" # "llm", "regex", "llm_fallback", "regex_fallback"
class ValidationResult(BaseModel):
"""Complete validation results for a session."""
session_id: str
validated_at: datetime
date_alignments: list[DateAlignmentValidation] = Field(default_factory=list)
land_tenure: list[LandTenureValidation] = Field(default_factory=list)
project_ids: list[ProjectIDValidation] = Field(default_factory=list)
contradictions: list[ContradictionCheck] = Field(default_factory=list)
summary: ValidationSummary
all_passed: bool
__all__ = [
"DateField",
"DateAlignmentValidation",
"LandTenureField",
"LandTenureValidation",
"ProjectIDOccurrence",
"ProjectIDValidation",
"ContradictionCheck",
"ValidationSummary",
"ValidationResult",
]
