coreason-manifest 0.10.0__py3-none-any.whl → 0.13.0__py3-none-any.whl

coreason_manifest/definitions/audit.py

@@ -1,181 +0,0 @@
-# Copyright (c) 2025 CoReason, Inc.
-#
-# This software is proprietary and dual-licensed.
-# Licensed under the Prosperity Public License 3.0 (the "License").
-# A copy of the license is available at https://prosperitylicense.com/versions/3.0.0
-# For details, see the LICENSE file.
-# Commercial use beyond a 30-day trial requires a separate license.
-#
-# Source Code: https://github.com/CoReason-AI/coreason-manifest
-
-import hashlib
-import json
-import uuid
-from datetime import datetime
-from enum import Enum
-from typing import Any, Dict, List, Optional
-from uuid import UUID
-
-from pydantic import ConfigDict, Field
-
-from coreason_manifest.definitions.base import CoReasonBaseModel
-
-from .message import ChatMessage
-
-
-class GenAITokenUsage(CoReasonBaseModel):
-    """Token consumption stats aligned with OTel conventions."""
-
-    model_config = ConfigDict(extra="ignore")
-
-    input: int = Field(0, description="Number of input tokens (prompt).")
-    output: int = Field(0, description="Number of output tokens (completion).")
-    total: int = Field(0, description="Total number of tokens used.")
-
-    # Backward compatibility fields (mapped to new fields in logic if needed, but kept for schema)
-    prompt_tokens: int = 0
-    completion_tokens: int = 0
-    total_tokens: int = 0
-    details: Dict[str, Any] = Field(default_factory=dict)
-
-    def __add__(self, other: "GenAITokenUsage") -> "GenAITokenUsage":
-        """Add two TokenUsage objects."""
-        new_details = self.details.copy()
-        new_details.update(other.details)
-        return GenAITokenUsage(
-            input=self.input + other.input,
-            output=self.output + other.output,
-            total=self.total + other.total,
-            prompt_tokens=self.prompt_tokens + other.prompt_tokens,
-            completion_tokens=self.completion_tokens + other.completion_tokens,
-            total_tokens=self.total_tokens + other.total_tokens,
-            details=new_details,
-        )
-
-    def __iadd__(self, other: "GenAITokenUsage") -> "GenAITokenUsage":
-        """In-place add two TokenUsage objects."""
-        self.input += other.input
-        self.output += other.output
-        self.total += other.total
-        self.prompt_tokens += other.prompt_tokens
-        self.completion_tokens += other.completion_tokens
-        self.total_tokens += other.total_tokens
-        self.details.update(other.details)
-        return self
-
-
-class GenAIOperation(CoReasonBaseModel):
-    """An atomic operation in the reasoning process (e.g., one LLM call), aligning with OTel Spans."""
-
-    model_config = ConfigDict(extra="ignore")
-
-    span_id: str = Field(..., description="Unique identifier for the operation/span.")
-    trace_id: str = Field(..., description="Trace ID this operation belongs to.")
-    parent_id: Optional[str] = Field(None, description="Parent span ID.")
-
-    operation_name: str = Field(..., description="Name of the operation (e.g., chat, embedding).")
-    provider: str = Field(..., description="GenAI provider (e.g., openai, anthropic).")
-    model: str = Field(..., description="Model name used.")
-
-    start_time: datetime = Field(default_factory=datetime.now)
-    end_time: Optional[datetime] = None
-    duration_ms: float = 0.0
-
-    # Context
-    input_messages: List[ChatMessage] = Field(default_factory=list)
-    output_messages: List[ChatMessage] = Field(default_factory=list)
-
-    # Metrics
-    token_usage: Optional[GenAITokenUsage] = None
-
-    # Metadata
-    status: str = "pending"  # success, error
-    error_type: Optional[str] = None
-    metadata: Dict[str, Any] = Field(default_factory=dict)
-
-    @classmethod
-    def thought(cls, content: str, **kwargs: Any) -> "GenAIOperation":
-        """Factory method to create a simplified thought/reasoning step."""
-        defaults = {
-            "span_id": str(uuid.uuid4()),
-            "trace_id": str(uuid.uuid4()),
-            "operation_name": "thought",
-            "provider": "internal",
-            "model": "internal",
-        }
-        defaults.update(kwargs)
-        # Ensure output_messages is not duplicated if passed in kwargs
-        defaults.pop("output_messages", None)
-        return cls(
-            **defaults,
-            output_messages=[ChatMessage.assistant(content)],
-        )
-
-
-class ReasoningTrace(CoReasonBaseModel):
-    """The full audit trail of an Agent's execution session.
-
-    Aligned with OpenTelemetry for trace identification.
-    """
-
-    model_config = ConfigDict(extra="ignore")
-
-    trace_id: str = Field(..., description="Trace ID (OTel format).")
-    agent_id: str
-    session_id: Optional[str] = None
-
-    start_time: datetime
-    end_time: Optional[datetime] = None
-
-    # The chain of thought (Ordered list of operations)
-    steps: List[GenAIOperation] = Field(default_factory=list)
-
-    # Final outcome
-    status: str = "pending"  # options: success, failure, pending
-    final_result: Optional[str] = None
-    error: Optional[str] = None
-
-    # Aggregated stats
-    total_tokens: GenAITokenUsage = Field(default_factory=GenAITokenUsage)
-    total_cost: float = 0.0
-
-    # Flexible metadata
-    metadata: Dict[str, Any] = Field(default_factory=dict)
-
-
-class AuditEventType(str, Enum):
-    SYSTEM_CHANGE = "system_change"
-    PREDICTION = "prediction"
-    GUARDRAIL_TRIGGER = "guardrail_trigger"
-
-
-class AuditLog(CoReasonBaseModel):
-    """Tamper-evident legal record.
-
-    IDs aligned with OpenTelemetry:
-    - audit_id: Unique record ID.
-    - trace_id: OTel Trace ID.
-    """
-
-    audit_id: UUID = Field(..., description="Unique identifier.")
-    trace_id: str = Field(..., description="Trace ID for OTel correlation.")
-    timestamp: datetime = Field(..., description="ISO8601 timestamp.")
-    actor: str = Field(..., description="User ID or Agent Component ID.")
-    event_type: AuditEventType = Field(..., description="Type of event.")
-    safety_metadata: Dict[str, Any] = Field(..., description="Safety metadata (e.g., PII detected).")
-    previous_hash: str = Field(..., description="Hash of the previous log entry.")
-    integrity_hash: str = Field(..., description="SHA256 hash of this record + previous_hash.")
-
-    def compute_hash(self) -> str:
-        """Computes the integrity hash of the record."""
-        # Use model_dump to get a dict, but exclude integrity_hash as it is the target
-        data = self.model_dump(exclude={"integrity_hash"}, mode="json")
-        # Ensure deterministic serialization
-        json_str = json.dumps(data, sort_keys=True, default=str)
-        return hashlib.sha256(json_str.encode("utf-8")).hexdigest()
-
-
-# --- Backward Compatibility ---
-# Adapters or Aliases
-CognitiveStep = GenAIOperation
-TokenUsage = GenAITokenUsage

coreason_manifest/definitions/events.py

@@ -1,423 +0,0 @@
-# Copyright (c) 2025 CoReason, Inc.
-#
-# This software is proprietary and dual-licensed.
-# Licensed under the Prosperity Public License 3.0 (the "License").
-# A copy of the license is available at https://prosperitylicense.com/versions/3.0.0
-# For details, see the LICENSE file.
-# Commercial use beyond a 30-day trial requires a separate license.
-#
-# Source Code: https://github.com/CoReason-AI/coreason-manifest
-
-from datetime import datetime, timezone
-from typing import Annotated, Any, Dict, Generic, Literal, Optional, Protocol, TypeVar, Union, runtime_checkable
-from uuid import uuid4
-
-from pydantic import ConfigDict, Field
-
-from coreason_manifest.definitions.base import CoReasonBaseModel
-from coreason_manifest.definitions.topology import RuntimeVisualMetadata
-
-# --- CloudEvents v1.0 Implementation ---
-
-T = TypeVar("T")
-
-
-@runtime_checkable
-class CloudEventSource(Protocol):
-    def as_cloud_event_payload(self) -> Any: ...
-
-
-class CloudEvent(CoReasonBaseModel, Generic[T]):
-    """Standard CloudEvent v1.0 Envelope."""
-
-    model_config = ConfigDict(extra="allow", populate_by_name=True)
-
-    specversion: Literal["1.0"] = "1.0"
-    type: str = Field(..., description="Type of occurrence (e.g. ai.coreason.node.started)")
-    source: str = Field(..., description="URI identifying the event producer")
-    id: str = Field(default_factory=lambda: str(uuid4()), description="Unique event identifier")
-    time: datetime = Field(
-        default_factory=lambda: datetime.now(timezone.utc), description="Timestamp of when the occurrence happened"
-    )
-    datacontenttype: Literal["application/json"] = "application/json"
-
-    data: Optional[T] = Field(None, description="The event payload")
-
-    # Distributed Tracing Extensions (W3C)
-    traceparent: Optional[str] = Field(None, description="W3C Trace Context traceparent")
-    tracestate: Optional[str] = Field(None, description="W3C Trace Context tracestate")
-
-
-# --- OTel Semantic Conventions ---
-
-
-class GenAIUsage(CoReasonBaseModel):
-    """GenAI Usage metrics."""
-
-    input_tokens: Optional[int] = Field(None, alias="input_tokens")
-    output_tokens: Optional[int] = Field(None, alias="output_tokens")
-
-    model_config = ConfigDict(populate_by_name=True)
-
-
-class GenAIRequest(CoReasonBaseModel):
-    """GenAI Request details."""
-
-    model: Optional[str] = None
-    temperature: Optional[float] = None
-    top_p: Optional[float] = None
-
-
-class GenAICompletion(CoReasonBaseModel):
-    """GenAI Completion details."""
-
-    chunk: Optional[str] = None
-    finish_reason: Optional[str] = None
-
-
-class GenAISemantics(CoReasonBaseModel):
-    """OpenTelemetry GenAI Semantic Conventions."""
-
-    system: Optional[str] = None
-    usage: Optional[GenAIUsage] = None
-    request: Optional[GenAIRequest] = None
-    completion: Optional[GenAICompletion] = None
-
-
-# --- Base Models ---
-
-
-class BaseNodePayload(CoReasonBaseModel):
-    """Base model for node-related events."""
-
-    model_config = ConfigDict(extra="ignore")
-    node_id: str
-
-    # Common OTel attributes that might appear in payload
-    model: Optional[str] = None
-    system: Optional[str] = None
-
-    def as_cloud_event_payload(self) -> Any:
-        return self
-
-
-# --- Payload Models ---
-
-
-class NodeInit(BaseNodePayload):
-    """Payload for NODE_INIT event."""
-
-    type: str = "DEFAULT"
-    visual_cue: str = "IDLE"
-
-
-class NodeStarted(BaseNodePayload):
-    """Payload for NODE_START event."""
-
-    timestamp: float
-    status: Literal["RUNNING"] = "RUNNING"
-    visual_cue: str = "PULSE"
-    input_tokens: Optional[int] = None
-
-    def as_cloud_event_payload(self) -> Any:
-        semantics = GenAISemantics()
-        has_semantics = False
-
-        if self.input_tokens is not None:
-            semantics.usage = GenAIUsage(input_tokens=self.input_tokens)
-            has_semantics = True
-
-        if self.model:
-            semantics.request = GenAIRequest(model=self.model)
-            has_semantics = True
-
-        if self.system:
-            semantics.system = self.system
-            has_semantics = True
-
-        return StandardizedNodeStarted(
-            node_id=self.node_id,
-            status=self.status,
-            gen_ai=semantics if has_semantics else None,
-        )
-
-
-class NodeCompleted(BaseNodePayload):
-    """Payload for NODE_DONE event."""
-
-    output_summary: str
-    status: Literal["SUCCESS"] = "SUCCESS"
-    visual_cue: str = "GREEN_GLOW"
-    cost: Optional[float] = None
-
-    def as_cloud_event_payload(self) -> Any:
-        semantics = GenAISemantics()
-        has_semantics = False
-
-        if self.model:
-            semantics.request = GenAIRequest(model=self.model)
-            has_semantics = True
-
-        return StandardizedNodeCompleted(
-            node_id=self.node_id,
-            output_summary=self.output_summary,
-            status=self.status,
-            gen_ai=semantics if has_semantics else None,
-        )
-
-
-class NodeRestored(BaseNodePayload):
-    """Payload for NODE_RESTORED event."""
-
-    output_summary: str
-    status: Literal["RESTORED"] = "RESTORED"
-    visual_cue: str = "INSTANT_GREEN"
-
-
-class NodeSkipped(BaseNodePayload):
-    """Payload for NODE_SKIPPED event."""
-
-    status: Literal["SKIPPED"] = "SKIPPED"
-    visual_cue: str = "GREY_OUT"
-
-
-class NodeStream(BaseNodePayload):
-    """Payload for NODE_STREAM event."""
-
-    chunk: str
-    visual_cue: str = "TEXT_BUBBLE"
-
-    def as_cloud_event_payload(self) -> Any:
-        semantics = GenAISemantics(completion=GenAICompletion(chunk=self.chunk))
-        if self.model:
-            if not semantics.request:
-                semantics.request = GenAIRequest()
-            semantics.request.model = self.model
-
-        return StandardizedNodeStream(node_id=self.node_id, gen_ai=semantics)
-
-
-class ArtifactGenerated(BaseNodePayload):
-    """Payload for ARTIFACT_GENERATED event."""
-
-    artifact_type: str = "PDF"
-    url: str
-
-
-class EdgeTraversed(CoReasonBaseModel):
-    """Payload for EDGE_ACTIVE event."""
-
-    model_config = ConfigDict(extra="ignore")
-    source: str
-    target: str
-    animation_speed: str = "FAST"
-
-    def as_cloud_event_payload(self) -> Any:
-        return self
-
-
-class CouncilVote(BaseNodePayload):
-    """Payload for COUNCIL_VOTE event."""
-
-    votes: Dict[str, str]
-
-
-class WorkflowError(BaseNodePayload):
-    """Payload for ERROR event."""
-
-    error_message: str
-    stack_trace: str
-    input_snapshot: Dict[str, Any]
-    status: Literal["ERROR"] = "ERROR"
-    visual_cue: str = "RED_FLASH"
-
-
-# --- Standardized Payloads ---
-
-
-class StandardizedNodeStarted(BaseNodePayload):
-    """Standardized payload for node start with OTel support."""
-
-    gen_ai: Optional[GenAISemantics] = None
-    status: Literal["RUNNING"] = "RUNNING"
-
-
-class StandardizedNodeCompleted(BaseNodePayload):
-    """Standardized payload for node completion with OTel support."""
-
-    gen_ai: Optional[GenAISemantics] = None
-    output_summary: str
-    status: Literal["SUCCESS"] = "SUCCESS"
-
-
-class StandardizedNodeStream(BaseNodePayload):
-    """Standardized payload for node stream with OTel support."""
-
-    gen_ai: Optional[GenAISemantics] = None
-
-
-# --- Aliases for Backward Compatibility ---
-# These ensure that code importing `NodeInitPayload` still works.
-NodeInitPayload = NodeInit
-NodeStartedPayload = NodeStarted
-NodeCompletedPayload = NodeCompleted
-NodeSkippedPayload = NodeSkipped
-NodeStreamPayload = NodeStream
-EdgeTraversedPayload = EdgeTraversed
-ArtifactGeneratedPayload = ArtifactGenerated
-CouncilVotePayload = CouncilVote
-WorkflowErrorPayload = WorkflowError
-
-
-# --- Graph Event Wrapper ---
-
-
-class BaseGraphEvent(CoReasonBaseModel):
-    """Base class for GraphEvents.
-
-    Standardized IDs:
-    - run_id: Workflow execution ID.
-    - trace_id: OpenTelemetry Distributed Trace ID.
-    """
-
-    model_config = ConfigDict(extra="forbid")
-
-    run_id: str
-    trace_id: str = Field(
-        default_factory=lambda: "unknown"
-    )  # Default for compatibility if missing in some legacy calls
-    node_id: str  # Required per BRD and existing tests
-    timestamp: float
-    sequence_id: Optional[int] = None  # Optional for internal use
-
-    # Visual Metadata drives the Flutter animation engine
-    visual_metadata: RuntimeVisualMetadata = Field(
-        ..., description="Hints for UI: color='#00FF00', animation='pulse', progress='0.5'"
-    )
-
-
-class GraphEventNodeInit(BaseGraphEvent):
-    event_type: Literal["NODE_INIT"] = "NODE_INIT"
-    payload: NodeInit = Field(..., description="The logic output")
-
-
-class GraphEventNodeStart(BaseGraphEvent):
-    event_type: Literal["NODE_START"] = "NODE_START"
-    payload: NodeStarted = Field(..., description="The logic output")
-
-
-class GraphEventNodeStream(BaseGraphEvent):
-    event_type: Literal["NODE_STREAM"] = "NODE_STREAM"
-    payload: NodeStream = Field(..., description="The logic output")
-
-
-class GraphEventNodeDone(BaseGraphEvent):
-    event_type: Literal["NODE_DONE"] = "NODE_DONE"
-    payload: NodeCompleted = Field(..., description="The logic output")
-
-
-class GraphEventNodeSkipped(BaseGraphEvent):
-    event_type: Literal["NODE_SKIPPED"] = "NODE_SKIPPED"
-    payload: NodeSkipped = Field(..., description="The logic output")
-
-
-class GraphEventEdgeActive(BaseGraphEvent):
-    event_type: Literal["EDGE_ACTIVE"] = "EDGE_ACTIVE"
-    payload: EdgeTraversed = Field(..., description="The logic output")
-
-
-class GraphEventCouncilVote(BaseGraphEvent):
-    event_type: Literal["COUNCIL_VOTE"] = "COUNCIL_VOTE"
-    payload: CouncilVote = Field(..., description="The logic output")
-
-
-class GraphEventError(BaseGraphEvent):
-    event_type: Literal["ERROR"] = "ERROR"
-    payload: WorkflowError = Field(..., description="The logic output")
-
-
-class GraphEventNodeRestored(BaseGraphEvent):
-    event_type: Literal["NODE_RESTORED"] = "NODE_RESTORED"
-    payload: NodeRestored = Field(..., description="The logic output")
-
-
-class GraphEventArtifactGenerated(BaseGraphEvent):
-    event_type: Literal["ARTIFACT_GENERATED"] = "ARTIFACT_GENERATED"
-    payload: ArtifactGenerated = Field(..., description="The logic output")
-
-
-GraphEvent = Annotated[
-    Union[
-        GraphEventNodeInit,
-        GraphEventNodeStart,
-        GraphEventNodeStream,
-        GraphEventNodeDone,
-        GraphEventNodeSkipped,
-        GraphEventEdgeActive,
-        GraphEventCouncilVote,
-        GraphEventError,
-        GraphEventNodeRestored,
-        GraphEventArtifactGenerated,
-    ],
-    Field(discriminator="event_type", description="Polymorphic graph event definition."),
-]
-
-
-# --- Migration Logic ---
-
-
-def migrate_graph_event_to_cloud_event(event: GraphEvent) -> CloudEvent[Any]:
-    """Migrates a legacy GraphEvent to a CloudEvent v1.0."""
-
-    ce_type_map = {
-        "NODE_START": "ai.coreason.node.started",
-        "NODE_STREAM": "ai.coreason.node.stream",
-        "NODE_DONE": "ai.coreason.node.completed",
-    }
-    ce_type = ce_type_map.get(event.event_type, f"ai.coreason.legacy.{event.event_type.lower()}")
-    ce_source = f"urn:node:{event.node_id}"
-
-    data: Any = None
-
-    # event.payload is already a strictly typed Pydantic model (from GraphEvent union).
-    # All supported payload models implement CloudEventSource protocol (duck-typed or via BaseNodePayload).
-    if isinstance(event.payload, CloudEventSource):
-        data = event.payload.as_cloud_event_payload()
-    else:
-        # Fallback for models that might not implement the protocol (e.g. unknown future extensions)
-        data = event.payload
-
-    # UI Metadata as extension
-    payload_visual_cue = getattr(event.payload, "visual_cue", None)
-
-    visual_dict = event.visual_metadata.model_dump(exclude_none=True)
-    extensions = {
-        "com_coreason_ui_cue": event.visual_metadata.animation or payload_visual_cue,
-        "com_coreason_ui_metadata": visual_dict,
-    }
-
-    # Filter out None values in extensions
-    # For dictionaries, we want to filter out empty dicts too.
-    filtered_extensions = {}
-    for k, v in extensions.items():
-        if v is None:
-            continue
-        if isinstance(v, dict) and not v:
-            continue
-        if isinstance(v, str) and not v:
-            continue
-        # Also check if it's a dict containing only empty strings (recursive check not needed for now, just 1 level)
-        if isinstance(v, dict) and all(isinstance(val, str) and not val for val in v.values()):
-            continue
-
-        filtered_extensions[k] = v
-
-    extensions = filtered_extensions
-
-    return CloudEvent(
-        type=ce_type,
-        source=ce_source,
-        time=datetime.fromtimestamp(event.timestamp, timezone.utc),
-        data=data,
-        **extensions,
-    )