enkryptai-sdk 0.1.6__py3-none-any.whl → 0.1.7__py3-none-any.whl

enkryptai_sdk/dto/guardrails.py

@@ -0,0 +1,1261 @@
+from enum import Enum
+from .base import BaseDTO
+from dataclasses import dataclass, field
+from typing import List, Dict, Any, Set
+
+
+class GuardrailsPIIModes(str, Enum):
+    REQUEST = "request"
+    RESPONSE = "response"
+
+
+@dataclass
+class GuardrailsHealthResponse(BaseDTO):
+    status: str
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsHealthResponse":
+        return cls(
+            status=data.get("status", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "status": self.status
+        }
+
+
+@dataclass
+class GuardrailsModelsResponse(BaseDTO):
+    models: List[str] = field(default_factory=list)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsModelsResponse":
+        return cls(
+            models=data.get("models", [])
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "models": self.models
+        }
+
+
+# -------------------------------------
+# Guardrails Detect Request
+# -------------------------------------
+
+
+@dataclass
+class TopicDetector(BaseDTO):
+    enabled: bool = False
+    topic: List[str] = field(default_factory=list)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "TopicDetector":
+        return cls(
+            enabled=data.get("enabled", False),
+            topic=data.get("topic", [])
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled,
+            "topic": self.topic
+        }
+
+
+@dataclass
+class NSFWDetector(BaseDTO):
+    enabled: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "NSFWDetector":
+        return cls(
+            enabled=data.get("enabled", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled
+        }
+
+
+@dataclass
+class ToxicityDetector(BaseDTO):
+    enabled: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "ToxicityDetector":
+        return cls(
+            enabled=data.get("enabled", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled
+        }
+
+
+@dataclass
+class PIIDetector(BaseDTO):
+    enabled: bool = False
+    entities: List[str] = field(default_factory=list)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "PIIDetector":
+        return cls(
+            enabled=data.get("enabled", False),
+            entities=data.get("entities", [])
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled,
+            "entities": self.entities
+        }
+
+
+@dataclass
+class InjectionAttackDetector(BaseDTO):
+    enabled: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "InjectionAttackDetector":
+        return cls(
+            enabled=data.get("enabled", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled
+        }
+
+
+@dataclass
+class KeywordDetector(BaseDTO):
+    enabled: bool = False
+    banned_keywords: List[str] = field(default_factory=list)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "KeywordDetector":
+        return cls(
+            enabled=data.get("enabled", False),
+            banned_keywords=data.get("banned_keywords", [])
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled,
+            "banned_keywords": self.banned_keywords
+        }
+
+
+@dataclass
+class PolicyViolationDetector(BaseDTO):
+    enabled: bool = False
+    policy_text: str = ""
+    need_explanation: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "PolicyViolationDetector":
+        return cls(
+            enabled=data.get("enabled", False),
+            policy_text=data.get("policy_text", ""),
+            need_explanation=data.get("need_explanation", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled,
+            "policy_text": self.policy_text,
+            "need_explanation": self.need_explanation
+        }
+
+
+@dataclass
+class BiasDetector(BaseDTO):
+    enabled: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "BiasDetector":
+        return cls(
+            enabled=data.get("enabled", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled
+        }
+
+
+@dataclass
+class CopyrightIPDetector(BaseDTO):
+    enabled: bool = False
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "CopyrightIPDetector":
+        return cls(
+            enabled=data.get("enabled", False)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled
+        }
+
+
+@dataclass
+class SystemPromptDetector(BaseDTO):
+    enabled: bool = False
+    index: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "SystemPromptDetector":
+        return cls(
+            enabled=data.get("enabled", False),
+            index=data.get("index", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "enabled": self.enabled,
+            "index": self.index
+        }
+
+
+@dataclass
+class GuardrailDetectors(BaseDTO):
+    topic_detector: TopicDetector = field(default_factory=TopicDetector)
+    nsfw: NSFWDetector = field(default_factory=NSFWDetector)
+    toxicity: ToxicityDetector = field(default_factory=ToxicityDetector)
+    pii: PIIDetector = field(default_factory=PIIDetector)
+    injection_attack: InjectionAttackDetector = field(default_factory=InjectionAttackDetector)
+    keyword_detector: KeywordDetector = field(default_factory=KeywordDetector)
+    policy_violation: PolicyViolationDetector = field(default_factory=PolicyViolationDetector)
+    bias: BiasDetector = field(default_factory=BiasDetector)
+    copyright_ip: CopyrightIPDetector = field(default_factory=CopyrightIPDetector)
+    system_prompt: SystemPromptDetector = field(default_factory=SystemPromptDetector)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailDetectors":
+        return cls(
+            topic_detector=TopicDetector.from_dict(data.get("topic_detector", {})),
+            nsfw=NSFWDetector.from_dict(data.get("nsfw", {})),
+            toxicity=ToxicityDetector.from_dict(data.get("toxicity", {})),
+            pii=PIIDetector.from_dict(data.get("pii", {})),
+            injection_attack=InjectionAttackDetector.from_dict(data.get("injection_attack", {})),
+            keyword_detector=KeywordDetector.from_dict(data.get("keyword_detector", {})),
+            policy_violation=PolicyViolationDetector.from_dict(data.get("policy_violation", {})),
+            bias=BiasDetector.from_dict(data.get("bias", {})),
+            copyright_ip=CopyrightIPDetector.from_dict(data.get("copyright_ip", {})),
+            system_prompt=SystemPromptDetector.from_dict(data.get("system_prompt", {}))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "topic_detector": self.topic_detector.to_dict(),
+            "nsfw": self.nsfw.to_dict(),
+            "toxicity": self.toxicity.to_dict(),
+            "pii": self.pii.to_dict(),
+            "injection_attack": self.injection_attack.to_dict(),
+            "keyword_detector": self.keyword_detector.to_dict(),
+            "policy_violation": self.policy_violation.to_dict(),
+            "bias": self.bias.to_dict(),
+            "copyright_ip": self.copyright_ip.to_dict(),
+            "system_prompt": self.system_prompt.to_dict()
+        }
+
+
+@dataclass
+class GuardrailsDetectRequest(BaseDTO):
+    text: str
+    detectors: GuardrailDetectors = field(default_factory=GuardrailDetectors)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsDetectRequest":
+        detectors_data = data.get("detectors", {})
+        return cls(
+            text=data.get("text", ""),
+            detectors=GuardrailDetectors.from_dict(detectors_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "text": self.text,
+            "detectors": self.detectors.to_dict()
+        }
+
+
+@dataclass
+class GuardrailsPolicyDetectRequest(BaseDTO):
+    text: str
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPolicyDetectRequest":
+        return cls(
+            text=data.get("text", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "text": self.text
+        }
+    
+
+# -------------------------------------
+# Guardrails Detect Response
+# -------------------------------------
+
+
+@dataclass
+class TopicDetectorDetail(BaseDTO):
+    _extra_fields: Dict[str, float] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "TopicDetectorDetail":
+        obj = cls()
+        obj._extra_fields = {k: float(v) for k, v in data.items()}
+        return obj
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return self._extra_fields
+
+
+@dataclass
+class NSFWDetail(BaseDTO):
+    sfw: float = 0.0
+    nsfw: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "NSFWDetail":
+        return cls(
+            sfw=float(data.get("sfw", 0.0)),
+            nsfw=float(data.get("nsfw", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "sfw": self.sfw,
+            "nsfw": self.nsfw
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class ToxicityDetail(BaseDTO):
+    toxicity: float = 0.0
+    severe_toxicity: float = 0.0
+    obscene: float = 0.0
+    threat: float = 0.0
+    insult: float = 0.0
+    identity_hate: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "ToxicityDetail":
+        return cls(
+            toxicity=float(data.get("toxicity", 0.0)),
+            severe_toxicity=float(data.get("severe_toxicity", 0.0)),
+            obscene=float(data.get("obscene", 0.0)),
+            threat=float(data.get("threat", 0.0)),
+            insult=float(data.get("insult", 0.0)),
+            identity_hate=float(data.get("identity_hate", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "toxicity": self.toxicity,
+            "severe_toxicity": self.severe_toxicity,
+            "obscene": self.obscene,
+            "threat": self.threat,
+            "insult": self.insult,
+            "identity_hate": self.identity_hate
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class PIIEntityDetail(BaseDTO):
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "PIIEntityDetail":
+        obj = cls()
+        obj._extra_fields = data
+        return obj
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return self._extra_fields
+
+
+@dataclass
+class PIIDetail(BaseDTO):
+    pii: PIIEntityDetail = field(default_factory=PIIEntityDetail)
+    secrets: PIIEntityDetail = field(default_factory=PIIEntityDetail)
+    ip_address: PIIEntityDetail = field(default_factory=PIIEntityDetail)
+    url: PIIEntityDetail = field(default_factory=PIIEntityDetail)
+    person: PIIEntityDetail = field(default_factory=PIIEntityDetail)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "PIIDetail":
+        return cls(
+            pii=PIIEntityDetail.from_dict(data.get("pii", {})),
+            secrets=PIIEntityDetail.from_dict(data.get("secrets", {})),
+            ip_address=PIIEntityDetail.from_dict(data.get("ip_address", {})),
+            url=PIIEntityDetail.from_dict(data.get("url", {})),
+            person=PIIEntityDetail.from_dict(data.get("person", {}))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "pii": self.pii.to_dict(),
+            "secrets": self.secrets.to_dict(),
+            "ip_address": self.ip_address.to_dict(),
+            "url": self.url.to_dict(),
+            "person": self.person.to_dict()
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class InjectionAttackDetail(BaseDTO):
+    safe: str = "0.0"
+    attack: str = "0.0"
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "InjectionAttackDetail":
+        return cls(
+            safe=data.get("safe", "0.0"),
+            attack=data.get("attack", "0.0")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "safe": self.safe,
+            "attack": self.attack
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class KeywordDetectorDetail(BaseDTO):
+    detected_keywords: List[str] = field(default_factory=list)
+    detected_counts: Dict[str, int] = field(default_factory=dict)
+    redacted_text: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "KeywordDetectorDetail":
+        return cls(
+            detected_keywords=data.get("detected_keywords", []),
+            detected_counts=data.get("detected_counts", {}),
+            redacted_text=data.get("redacted_text", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "detected_keywords": self.detected_keywords,
+            "detected_counts": self.detected_counts,
+            "redacted_text": self.redacted_text
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class PolicyViolationDetail(BaseDTO):
+    violating_policy: str = ""
+    explanation: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "PolicyViolationDetail":
+        return cls(
+            violating_policy=data.get("violating_policy", ""),
+            explanation=data.get("explanation", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "violating_policy": self.violating_policy,
+            "explanation": self.explanation
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class BiasDetail(BaseDTO):
+    bias_detected: bool = False
+    debiased_text: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "BiasDetail":
+        return cls(
+            bias_detected=data.get("bias_detected", False),
+            debiased_text=data.get("debiased_text", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "bias_detected": self.bias_detected,
+            "debiased_text": self.debiased_text
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class CopyrightIPDetail(BaseDTO):
+    similarity_score: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "CopyrightIPDetail":
+        return cls(
+            similarity_score=float(data.get("similarity_score", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "similarity_score": self.similarity_score
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class SystemPromptDetail(BaseDTO):
+    similarity_score: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "SystemPromptDetail":
+        return cls(
+            similarity_score=float(data.get("similarity_score", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "similarity_score": self.similarity_score
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class DetectResponseSummary(BaseDTO):
+    on_topic: int = 0
+    nsfw: int = 0
+    toxicity: List[str] = field(default_factory=list)
+    pii: int = 0
+    injection_attack: int = 0
+    keyword_detected: int = 0
+    policy_violation: int = 0
+    bias: int = 0
+    copyright_ip_similarity: int = 0
+    system_prompt_similarity: int = 0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "DetectResponseSummary":
+        return cls(
+            on_topic=data.get("on_topic", 0),
+            nsfw=data.get("nsfw", 0),
+            toxicity=data.get("toxicity", []),
+            pii=data.get("pii", 0),
+            injection_attack=data.get("injection_attack", 0),
+            keyword_detected=data.get("keyword_detected", 0),
+            policy_violation=data.get("policy_violation", 0),
+            bias=data.get("bias", 0),
+            copyright_ip_similarity=data.get("copyright_ip_similarity", 0),
+            system_prompt_similarity=data.get("system_prompt_similarity", 0)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "on_topic": self.on_topic,
+            "nsfw": self.nsfw,
+            "toxicity": self.toxicity,
+            "pii": self.pii,
+            "injection_attack": self.injection_attack,
+            "keyword_detected": self.keyword_detected,
+            "policy_violation": self.policy_violation,
+            "bias": self.bias,
+            "copyright_ip_similarity": self.copyright_ip_similarity,
+            "system_prompt_similarity": self.system_prompt_similarity
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class DetectResponseDetails(BaseDTO):
+    topic_detector: TopicDetectorDetail = field(default_factory=TopicDetectorDetail)
+    nsfw: NSFWDetail = field(default_factory=NSFWDetail)
+    toxicity: ToxicityDetail = field(default_factory=ToxicityDetail)
+    pii: PIIDetail = field(default_factory=PIIDetail)
+    injection_attack: InjectionAttackDetail = field(default_factory=InjectionAttackDetail)
+    keyword_detector: KeywordDetectorDetail = field(default_factory=KeywordDetectorDetail)
+    policy_violation: PolicyViolationDetail = field(default_factory=PolicyViolationDetail)
+    bias: BiasDetail = field(default_factory=BiasDetail)
+    copyright_ip: CopyrightIPDetail = field(default_factory=CopyrightIPDetail)
+    system_prompt: SystemPromptDetail = field(default_factory=SystemPromptDetail)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "DetectResponseDetails":
+        return cls(
+            topic_detector=TopicDetectorDetail.from_dict(data.get("topic_detector", {})),
+            nsfw=NSFWDetail.from_dict(data.get("nsfw", {})),
+            toxicity=ToxicityDetail.from_dict(data.get("toxicity", {})),
+            pii=PIIDetail.from_dict(data.get("pii", {})),
+            injection_attack=InjectionAttackDetail.from_dict(data.get("injection_attack", {})),
+            keyword_detector=KeywordDetectorDetail.from_dict(data.get("keyword_detector", {})),
+            policy_violation=PolicyViolationDetail.from_dict(data.get("policy_violation", {})),
+            bias=BiasDetail.from_dict(data.get("bias", {})),
+            copyright_ip=CopyrightIPDetail.from_dict(data.get("copyright_ip", {})),
+            system_prompt=SystemPromptDetail.from_dict(data.get("system_prompt", {}))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "topic_detector": self.topic_detector.to_dict(),
+            "nsfw": self.nsfw.to_dict(),
+            "toxicity": self.toxicity.to_dict(),
+            "pii": self.pii.to_dict(),
+            "injection_attack": self.injection_attack.to_dict(),
+            "keyword_detector": self.keyword_detector.to_dict(),
+            "policy_violation": self.policy_violation.to_dict(),
+            "bias": self.bias.to_dict(),
+            "copyright_ip": self.copyright_ip.to_dict(),
+            "system_prompt": self.system_prompt.to_dict()
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsDetectResponse(BaseDTO):
+    summary: DetectResponseSummary
+    details: DetectResponseDetails
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsDetectResponse":
+        summary_data = data.get("summary", {})
+        details_data = data.get("details", {})
+        
+        return cls(
+            summary=DetectResponseSummary.from_dict(summary_data),
+            details=DetectResponseDetails.from_dict(details_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "summary": self.summary.to_dict(),
+            "details": self.details.to_dict()
+        }
+        result.update(self._extra_fields)
+        return result
+
+    def has_violations(self) -> bool:
+        """
+        Check if any detectors found violations in the content.
+        
+        Returns:
+            bool: True if any detector reported a violation (score > 0), False otherwise
+        """
+        summary = self.summary.to_dict()
+        for key, value in summary.items():
+            if key == "toxicity" and isinstance(value, list) and len(value) > 0:
+                return True
+            elif isinstance(value, (int, float)) and value > 0:
+                return True
+        return False
+    
+    def get_violations(self) -> list[str]:
+        """
+        Get a list of detector names that found violations.
+        
+        Returns:
+            list[str]: Names of detectors that reported violations
+        """
+        summary = self.summary.to_dict()
+        violations = []
+        for detector, value in summary.items():
+            if detector == "toxicity" and isinstance(value, list) and len(value) > 0:
+                violations.append(detector)
+            elif isinstance(value, (int, float)) and value > 0:
+                violations.append(detector)
+        return violations
+
+    def is_safe(self) -> bool:
+        """
+        Check if the content is safe (no violations detected).
+        
+        Returns:
+            bool: True if no violations were detected, False otherwise
+        """
+        return not self.has_violations()
+    
+    def is_attack(self) -> bool:
+        """
+        Check if the content is attacked (violations detected).
+        
+        Returns:
+            bool: True if violations were detected, False otherwise
+        """
+        return self.has_violations()
+    
+    def __str__(self) -> str:
+        """
+        String representation of the response.
+        
+        Returns:
+            str: A formatted string showing summary and violation status
+        """
+        violations = self.get_violations()
+        status = "UNSAFE" if violations else "SAFE"
+        
+        if violations:
+            violation_str = f"Violations detected: {', '.join(violations)}"
+        else:
+            violation_str = "No violations detected"
+            
+        return f"Response Status: {status}\n{violation_str}"
+
+
+# -------------------------------------
+# Guardrails PII
+# -------------------------------------
+
+
+@dataclass
+class GuardrailsPIIRequest(BaseDTO):
+    text: str
+    # mode: GuardrailsPIIModes = GuardrailsPIIModes.REQUEST
+    mode: str = "request"
+    key: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPIIRequest":
+        return cls(
+            text=data.get("text", ""),
+            # mode=data.get("mode", GuardrailsPIIModes.REQUEST),
+            mode=data.get("mode", "request"),
+            key=data.get("key", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "text": self.text,
+            "mode": self.mode,
+            "key": self.key
+        }
+
+
+@dataclass
+class GuardrailsPIIResponse(BaseDTO):
+    text: str
+    key: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPIIResponse":
+        return cls(
+            text=data.get("text", ""),
+            key=data.get("key", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "text": self.text,
+            "key": self.key
+        }
+
+
+# -------------------------------------
+# Guardrails Hallucination
+# -------------------------------------
+
+@dataclass
+class GuardrailsHallucinationRequest(BaseDTO):
+    request_text: str
+    response_text: str
+    context: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsHallucinationRequest":
+        return cls(
+            request_text=data.get("request_text", ""),
+            response_text=data.get("response_text", ""),
+            context=data.get("context", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "request_text": self.request_text,
+            "response_text": self.response_text,
+            "context": self.context
+        }
+
+
+@dataclass
+class HallucinationSummary(BaseDTO):
+    is_hallucination: int
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "HallucinationSummary":
+        return cls(
+            is_hallucination=data.get("is_hallucination", 0)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "is_hallucination": self.is_hallucination
+        }
+
+
+@dataclass
+class HallucinationDetails(BaseDTO):
+    prompt_based: float
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "HallucinationDetails":
+        return cls(
+            prompt_based=float(data.get("prompt_based", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "prompt_based": self.prompt_based
+        }
+
+
+@dataclass
+class GuardrailsHallucinationResponse(BaseDTO):
+    summary: HallucinationSummary
+    details: HallucinationDetails
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsHallucinationResponse":
+        summary_data = data.get("summary", {})
+        details_data = data.get("details", {})
+        
+        return cls(
+            summary=HallucinationSummary.from_dict(summary_data),
+            details=HallucinationDetails.from_dict(details_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "summary": self.summary.to_dict(),
+            "details": self.details.to_dict()
+        }
+
+
+# -------------------------------------
+# Guardrails Adherence
+# -------------------------------------
+
+
+@dataclass
+class GuardrailsAdherenceRequest(BaseDTO):
+    llm_answer: str
+    context: str
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsAdherenceRequest":
+        return cls(
+            llm_answer=data.get("llm_answer", ""),
+            context=data.get("context", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "llm_answer": self.llm_answer,
+            "context": self.context
+        }
+
+
+@dataclass
+class AdherenceSummary(BaseDTO):
+    adherence_score: float
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "AdherenceSummary":
+        return cls(
+            adherence_score=float(data.get("adherence_score", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "adherence_score": self.adherence_score
+        }
+
+
+@dataclass
+class AdherenceDetails(BaseDTO):
+    atomic_facts: List[str] = field(default_factory=list)
+    adherence_list: List[int] = field(default_factory=list)
+    adherence_response: str = ""
+    adherence_latency: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "AdherenceDetails":
+        return cls(
+            atomic_facts=data.get("atomic_facts", []),
+            adherence_list=data.get("adherence_list", []),
+            adherence_response=data.get("adherence_response", ""),
+            adherence_latency=float(data.get("adherence_latency", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "atomic_facts": self.atomic_facts,
+            "adherence_list": self.adherence_list,
+            "adherence_response": self.adherence_response,
+            "adherence_latency": self.adherence_latency
+        }
+
+
+@dataclass
+class GuardrailsAdherenceResponse(BaseDTO):
+    summary: AdherenceSummary
+    details: AdherenceDetails
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsAdherenceResponse":
+        summary_data = data.get("summary", {})
+        details_data = data.get("details", {})
+        
+        return cls(
+            summary=AdherenceSummary.from_dict(summary_data),
+            details=AdherenceDetails.from_dict(details_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "summary": self.summary.to_dict(),
+            "details": self.details.to_dict()
+        }
+
+
+# -------------------------------------
+# Guardrails Relevancy
+# -------------------------------------
+
+@dataclass
+class GuardrailsRelevancyRequest(BaseDTO):
+    question: str
+    llm_answer: str
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsRelevancyRequest":
+        return cls(
+            question=data.get("question", ""),
+            llm_answer=data.get("llm_answer", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "question": self.question,
+            "llm_answer": self.llm_answer
+        }
+
+
+@dataclass
+class RelevancySummary(BaseDTO):
+    relevancy_score: float
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "RelevancySummary":
+        return cls(
+            relevancy_score=float(data.get("relevancy_score", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "relevancy_score": self.relevancy_score
+        }
+
+
+@dataclass
+class RelevancyDetails(BaseDTO):
+    atomic_facts: List[str] = field(default_factory=list)
+    relevancy_list: List[int] = field(default_factory=list)
+    relevancy_response: str = ""
+    relevancy_latency: float = 0.0
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "RelevancyDetails":
+        return cls(
+            atomic_facts=data.get("atomic_facts", []),
+            relevancy_list=data.get("relevancy_list", []),
+            relevancy_response=data.get("relevancy_response", ""),
+            relevancy_latency=float(data.get("relevancy_latency", 0.0))
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "atomic_facts": self.atomic_facts,
+            "relevancy_list": self.relevancy_list,
+            "relevancy_response": self.relevancy_response,
+            "relevancy_latency": self.relevancy_latency
+        }
+
+
+@dataclass
+class GuardrailsRelevancyResponse(BaseDTO):
+    summary: RelevancySummary
+    details: RelevancyDetails
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsRelevancyResponse":
+        summary_data = data.get("summary", {})
+        details_data = data.get("details", {})
+        
+        return cls(
+            summary=RelevancySummary.from_dict(summary_data),
+            details=RelevancyDetails.from_dict(details_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "summary": self.summary.to_dict(),
+            "details": self.details.to_dict()
+        }
+
+
+# -------------------------------------
+# Guardrails Policy Request
+# -------------------------------------
+
+
+@dataclass
+class GuardrailsPolicyRequest(BaseDTO):
+    name: str
+    detectors: GuardrailDetectors = field(default_factory=GuardrailDetectors)
+    description: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPolicyRequest":
+        detectors_data = data.get("detectors", {})
+        return cls(
+            name=data.get("name", ""),
+            detectors=GuardrailDetectors.from_dict(detectors_data),
+            description=data.get("description", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "name": self.name,
+            "detectors": self.detectors.to_dict(),
+            "description": self.description
+        }
+
+
+# -------------------------------------
+# Guardrails Policy Response
+# -------------------------------------
+
+
+@dataclass
+class GuardrailsPolicyData(BaseDTO):
+    created_at: str
+    name: str
+    updated_at: str
+    description: str
+    policy_id: int
+    project_name: str = ""
+    is_sample: bool = False
+    detectors: GuardrailDetectors = field(default_factory=GuardrailDetectors)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPolicyData":
+        detectors_data = data.get("detectors", {})
+        return cls(
+            created_at=data.get("created_at", ""),
+            name=data.get("name", ""),
+            updated_at=data.get("updated_at", ""),
+            description=data.get("description", ""),
+            policy_id=data.get("policy_id", 0),
+            project_name=data.get("project_name", ""),
+            detectors=GuardrailDetectors.from_dict(detectors_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "created_at": self.created_at,
+            "name": self.name,
+            "updated_at": self.updated_at,
+            "description": self.description,
+            "policy_id": self.policy_id,
+            "project_name": self.project_name,
+            "detectors": self.detectors.to_dict(),
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsaPolicyResponse(BaseDTO):
+    message: str
+    data: GuardrailsPolicyData
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsaPolicyResponse":
+        policy_data = data.get("data", {})
+        return cls(
+            message=data.get("message", ""),
+            data=GuardrailsPolicyData.from_dict(policy_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "message": self.message,
+            "data": self.data.to_dict()
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsDeletePolicyData(BaseDTO):
+    policy_id: int
+    project_name: str = ""
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsDeletePolicyData":
+        return cls(
+            policy_id=data.get("policy_id", 0),
+            project_name=data.get("project_name", "")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "policy_id": self.policy_id,
+            "project_name": self.project
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsDeletePolicyResponse(BaseDTO):
+    message: str
+    data: GuardrailsDeletePolicyData
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsDeletePolicyResponse":
+        policy_data = data.get("data", {})
+        return cls(
+            message=data.get("message", ""),
+            data=GuardrailsDeletePolicyData.from_dict(policy_data)
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "message": self.message,
+            "data": self.data.to_dict()
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsPolicyListItem(BaseDTO):
+    policy_id: str
+    name: str
+    description: str
+    created_at: str
+    updated_at: str
+    project_name: str = "default"
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsPolicyListItem":
+        return cls(
+            policy_id=data.get("policy_id", ""),
+            name=data.get("name", ""),
+            description=data.get("description", ""),
+            created_at=data.get("created_at", ""),
+            updated_at=data.get("updated_at", ""),
+            project_name=data.get("project_name", "default")
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "policy_id": self.policy_id,
+            "name": self.name,
+            "description": self.description,
+            "created_at": self.created_at,
+            "updated_at": self.updated_at,
+            "project_name": self.project_name
+        }
+        result.update(self._extra_fields)
+        return result
+
+
+@dataclass
+class GuardrailsListPoliciesResponse(BaseDTO):
+    policies: List[GuardrailsPolicyListItem] = field(default_factory=list)
+    _extra_fields: Dict[str, Any] = field(default_factory=dict)
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "GuardrailsListPoliciesResponse":
+        policies_data = data.get("policies", [])
+        return cls(
+            policies=[GuardrailsPolicyListItem.from_dict(policy) for policy in policies_data]
+        )
+    
+    def to_dict(self) -> Dict[str, Any]:
+        result = {
+            "policies": [policy.to_dict() for policy in self.policies]
+        }
+        result.update(self._extra_fields)
+        return result
+