argus-code 0.2.0__py3-none-any.whl

argus/schema/types.py

@@ -0,0 +1,157 @@
+"""Core data model types — pydantic v2 BaseModels.
+
+Same shape as the TS `interface`s in src/schema/types.ts. The `metadata`
+field stays as a plain dict and is serialized to JSON on disk via
+``json.dumps`` (not pydantic's serializer) to keep the byte representation
+identical to what the TS version writes.
+"""
+from __future__ import annotations
+
+from typing import Any, Literal
+
+from pydantic import BaseModel, ConfigDict, Field
+
+# `agent` is an open string (not a closed Literal) so future adapters
+# (codex, openclaw, hermes) can self-register without editing this file.
+# The Adapter registry enforces "must match a registered adapter" at the
+# repository boundary.
+AgentName = str
+
+SegmentRole = Literal["user", "assistant", "thinking", "tool_result"]
+
+
+class NormalizedCacheFields(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    fresh_input_tokens: int
+    cache_read_tokens: int
+    cache_write_tokens: int
+    cache_write_5m_tokens: int | None
+    cache_write_1h_tokens: int | None
+
+
+class Session(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    id: str
+    agent: AgentName
+    agent_version: str | None
+    project_path: str
+    started_at: str
+    ended_at: str | None
+    duration_sec: int | None
+    total_fresh_input_tokens: int = 0
+    total_output_tokens: int = 0
+    total_cache_read_tokens: int = 0
+    total_cache_write_tokens: int = 0
+    total_cost_usd: float = 0.0
+    primary_model: str
+    turn_count: int = 0
+    pricing_table_version: str
+    computed_at: str
+    agent_reported_cost_usd: float | None
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
+class Turn(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    id: str
+    session_id: str
+    sequence: int
+    timestamp: str
+    model: str
+    model_raw: str
+    fresh_input_tokens: int = 0
+    output_tokens: int = 0
+    cache_read_tokens: int = 0
+    cache_write_tokens: int = 0
+    cache_write_5m_tokens: int | None = None
+    cache_write_1h_tokens: int | None = None
+    tool_calls_count: int = 0
+    cost_usd: float = 0.0
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
+class RawTurnEvent(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    native_turn_id: str
+    sequence: int
+    timestamp: str
+    model: str
+    model_raw: str
+    fresh_input_tokens: int = 0
+    output_tokens: int = 0
+    cache_read_tokens: int = 0
+    cache_write_tokens: int = 0
+    cache_write_5m_tokens: int | None = None
+    cache_write_1h_tokens: int | None = None
+    tool_calls_count: int = 0
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
+class RawSessionHeader(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    native_session_id: str
+    agent: AgentName
+    agent_version: str | None
+    project_path: str
+    started_at: str
+    ended_at: str | None
+    agent_reported_cost_usd: float | None
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
+class ToolCall(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    id: str
+    session_id: str
+    turn_index: int
+    tool_name: str
+    is_error: int  # 0 | 1
+    input_size: int
+    subagent_type: str | None
+    timestamp: str
+
+
+class Prompt(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    id: int | None = None  # autoincrement; None on insert
+    timestamp_ms: int
+    project_path: str
+    display: str
+    pasted_chars: int = 0
+    is_slash: int = 0  # 0 | 1
+
+
+class TranscriptSegment(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    uid: str
+    session_id: str
+    timestamp: str
+    role: SegmentRole
+    text: str
+
+
+AlertSeverity = Literal["info", "warning", "critical"]
+
+
+class Alert(BaseModel):
+    model_config = ConfigDict(frozen=True)
+
+    id: int | None = None
+    detector: str
+    dedup_key: str
+    severity: AlertSeverity
+    title: str
+    message: str
+    metadata: dict[str, Any] = Field(default_factory=dict)
+    first_seen_at: str
+    last_seen_at: str
+    seen_at: str | None = None
+    resolved_at: str | None = None