Pertinência de equipo fotossensível

Overview

Pertinence agent for photosensitive equipment (AI004). Evaluates whether photosensitive equipment charges are clinically justified by checking if the same invoice contains at least one medication from the photosensitive_equip_allowlist BigQuery artifact. Equipment in invoices without any justifying medication is flagged.

Architecture

Field	Value
Name	Pertinência de equipo fotossensível
Agent ID	AI004
Version	0.0.0
Nature	audit
Layer	—

Inheritance

PhotosensitiveEquipPertinenceRule → PertinenceRule → BaseAgent

Core Functionality

load_artifacts
prepare_data
rule_assessment

Business Rules

Pertinência da cobrança

Consultar código-fonte para regras de negócio detalhadas.

Usage

from agents.photosensitiveequippertinencerule import PhotosensitiveEquipPertinenceRule

rule = PhotosensitiveEquipPertinenceRule()
violations_df = rule.rule_assessment(claims_df)

Configuration

rule = PhotosensitiveEquipPertinenceRule(
    evaluated=False,
    evaluated=evaluated,
    artifact_loader=ArtifactLoader(self.logger),
    DataFrame=pd.DataFrame(),
    DataFrame=pd.DataFrame(),
    allowlist_keys=set(),
    taxonomy_raw=self.artifact_loader.load_taxonomy(,
    familia_tax_3=[TARGET_FAMILIA_TAX_3],
)

Input Requirements

Column	Type	Description
`EVENT_COLUMNS.id`	-	-
`EVENT_COLUMNS.id_fatura`	-	-
`EVENT_COLUMNS.id_evento`	-	-
`EVENT_COLUMNS.evento`	-	-
`EVENT_COLUMNS.vl_pago`	-	-
`EVENT_COLUMNS.qt_paga`	-	-
`EVENT_COLUMNS.categoria`	-	-
`ARVO_COLUMNS.id_evento_predicted`	-	-
`ARVO_COLUMNS.code_type_predicted`	-	-

Output Format

Glosa DataFrame Columns

Column	Description
`agent_id`	AI004
`version`	0.0.0
`id_arvo`	Original record ID
`vl_glosa_arvo`	Glosa value
`qt_glosa_arvo`	Glosa quantity
`score_arvo`	Confidence score
`motivo_glosa_arvo`	Detailed reason in Portuguese
`motivo_glosa_ANS`	ANS standard code
`data_dict`	JSON metadata with calculation details

Error Handling

DataFrames vazios retornam estrutura vazia com colunas obrigatórias.
Colunas ausentes geram warning mas não causam crash.
Conversões numéricas usam errors="coerce" para evitar falhas.