agentra 0.3.0__py3-none-any.whl

agentra/__init__.py

@@ -0,0 +1,149 @@
+"""
+agentra — Red-team, eval, and monitor your LLMs. Security-first, Python-native.
+
+Quick start:
+    import agentra
+    agentra.init()
+
+    # Red team your chatbot
+    report = agentra.red_team(my_chatbot, plugins=["jailbreak", "pii"])
+    report.summary()
+
+    # Attack heatmap across models
+    fp = agentra.guard.fingerprint({"gpt-4o-mini": fn1, "claude-haiku": fn2})
+    fp.heatmap()
+
+    # Auto-generate test cases
+    ds = agentra.auto_dataset(my_chatbot, n=50, focus="adversarial")
+"""
+from __future__ import annotations
+
+__version__ = "0.3.0"
+__author__ = "agentra"
+
+# Guard (primary — security)
+from agentra.guard.red_team import red_team, RedTeamReport
+from agentra.guard.fingerprint import fingerprint, ModelFingerprint
+from agentra.guard.auto_dataset import auto_dataset
+from agentra.guard.swarm import scan_swarm, SwarmScanReport
+from agentra.guard.toolchain import scan_toolchain, ToolchainReport
+from agentra.guard.prompt_leakage import prompt_leakage_score, LeakageReport
+from agentra.guard.multilingual import scan_multilingual, MultilingualReport
+from agentra.guard.mcp_scanner import scan_mcp, MCPScanReport
+from agentra.guard.mcp_static import analyze_mcp_tools, ToolRiskReport
+
+# Eval
+from agentra.eval.dataset import Dataset, DatasetItem
+from agentra.eval.experiment import Experiment, ExperimentResults
+from agentra.eval import scorers
+from agentra.eval.compare import compare_models, prompt_ab_test
+
+# Monitor
+from agentra.monitor.tracer import trace, span
+from agentra.monitor.drift import DriftDetector, DriftReport
+
+# Sub-packages
+from agentra import guard, eval, monitor
+
+__all__ = [
+    # Core
+    "init",
+    "dataset",
+    "experiment",
+    # Guard
+    "red_team",
+    "RedTeamReport",
+    "fingerprint",
+    "ModelFingerprint",
+    "auto_dataset",
+    "scan_swarm",
+    "SwarmScanReport",
+    "scan_toolchain",
+    "ToolchainReport",
+    "prompt_leakage_score",
+    "LeakageReport",
+    "scan_multilingual",
+    "MultilingualReport",
+    "scan_mcp",
+    "MCPScanReport",
+    "analyze_mcp_tools",
+    "ToolRiskReport",
+    "guard",
+    # Eval
+    "Dataset",
+    "DatasetItem",
+    "Experiment",
+    "ExperimentResults",
+    "scorers",
+    "compare_models",
+    "prompt_ab_test",
+    "eval",
+    # Monitor
+    "trace",
+    "span",
+    "DriftDetector",
+    "DriftReport",
+    "monitor",
+]
+
+
+def init(
+    persist: bool = True,
+    db_path: str | None = None,
+    offline: bool = False,
+    local_judge_model: str = "llama3",
+    judge_model: str = "gpt-4o-mini",
+) -> None:
+    """
+    Initialize agentra — enable persistence and activate SDK interceptors.
+
+    Args:
+        persist: write results to SQLite (default True)
+        db_path: custom path for agentra.db (default: ~/.agentra/data.db)
+        offline: use local Ollama model for judging (no external API calls)
+        local_judge_model: Ollama model to use when offline=True
+        judge_model: default judge model when offline=False
+
+    Example:
+        agentra.init()                              # Standard
+        agentra.init(offline=True)                  # Fully offline with Ollama
+        agentra.init(db_path="/tmp/agentra.db")     # Custom DB path
+    """
+    from agentra import providers, db
+
+    # Configure providers
+    providers.configure(
+        offline=offline,
+        local_judge_model=local_judge_model,
+        judge_model=judge_model,
+    )
+
+    # Configure DB path
+    if db_path:
+        db.set_db_path(db_path)
+
+    # Initialize DB schema
+    if persist:
+        db.init_db(db_path)
+
+    # Activate SDK interceptors
+    try:
+        from agentra import interceptor
+        interceptor.activate()
+    except Exception:
+        pass
+
+
+def dataset(name: str, description: str = "") -> Dataset:
+    """Create or load a named dataset."""
+    return Dataset(name, description)
+
+
+def experiment(
+    name: str,
+    dataset: Dataset | str,
+    fn,
+    scorers: list | None = None,
+) -> Experiment:
+    """Create an experiment."""
+    return Experiment(name=name, dataset=dataset, fn=fn, scorers=scorers or [])