abstractflow 0.3.0__py3-none-any.whl → 0.3.1__py3-none-any.whl

abstractflow/__init__.py

@@ -39,9 +39,9 @@ For agent-based flows:
     >>> result = runner.run({"context": {"task": "Build a REST API"}})
 """
 
-__version__ = "0.3.0"
+__version__ = "0.3.1"
 __author__ = "Laurent-Philippe Albou"
-__email__ = "contact@abstractcore.ai"
+__email__ = "contact@abstractflow.ai"
 __license__ = "MIT"
 
 # Core classes

abstractflow/adapters/agent_adapter.py

@@ -1,124 +1,5 @@
-"""Adapter for using AbstractAgent agents as flow nodes."""
+"""Re-export: AbstractRuntime VisualFlow compiler adapter."""
 
 from __future__ import annotations
 
-from typing import Any, Callable, Optional, TYPE_CHECKING
-
-if TYPE_CHECKING:
-    from abstractruntime.core.models import RunState, StepPlan
-    from abstractagent.agents.base import BaseAgent
-
-
-def create_agent_node_handler(
-    node_id: str,
-    agent: "BaseAgent",
-    next_node: Optional[str],
-    input_key: Optional[str] = None,
-    output_key: Optional[str] = None,
-) -> Callable:
-    """Create a node handler that runs an agent as a subworkflow.
-
-    Agent nodes execute the full agent workflow (ReAct loop, CodeAct, etc.)
-    as a subworkflow. The agent runs to completion before transitioning.
-
-    Args:
-        node_id: Unique identifier for this node
-        agent: The agent instance to run
-        next_node: ID of the next node to transition to (None for terminal)
-        input_key: Key in run.vars to read task/input from
-        output_key: Key in run.vars to write agent output to
-
-    Returns:
-        A node handler function compatible with AbstractRuntime
-
-    Example:
-        >>> from abstractagent import create_react_agent
-        >>> planner = create_react_agent(provider="ollama", model="qwen3:4b")
-        >>> handler = create_agent_node_handler("plan", planner, "search")
-    """
-    from abstractruntime.core.models import Effect, EffectType, StepPlan
-
-    def handler(run: "RunState", ctx: Any) -> "StepPlan":
-        """Start the agent as a subworkflow."""
-        # Determine task for the agent
-        task = ""
-
-        if input_key:
-            input_data = run.vars.get(input_key, {})
-            if isinstance(input_data, dict):
-                task = input_data.get("task", "") or input_data.get("query", "")
-                if not task:
-                    # Use the whole input as context
-                    task = str(input_data)
-            else:
-                task = str(input_data)
-
-        # Fallback to flow's main task
-        if not task:
-            context = run.vars.get("context", {})
-            if isinstance(context, dict):
-                task = context.get("task", "")
-
-        if not task:
-            task = f"Execute {node_id} step"
-
-        # Build initial vars for the agent subworkflow
-        max_iterations = getattr(agent, "_max_iterations", 25)
-        max_history_messages = getattr(agent, "_max_history_messages", -1)
-        max_tokens = getattr(agent, "_max_tokens", None)
-        if not isinstance(max_tokens, int) or max_tokens <= 0:
-            try:
-                runtime = getattr(agent, "runtime", None)
-                config = getattr(runtime, "config", None)
-                base = config.to_limits_dict() if config is not None else {}
-                max_tokens = int(base.get("max_tokens", 32768) or 32768)
-            except Exception:
-                max_tokens = 32768
-
-        agent_vars = {
-            "context": {
-                "task": task,
-                "messages": [],
-            },
-            "scratchpad": {
-                "iteration": 0,
-                "max_iterations": max_iterations,
-                "max_history_messages": max_history_messages,
-            },
-            "_runtime": {"inbox": []},
-            "_temp": {},
-            # Canonical _limits namespace for runtime awareness
-            "_limits": {
-                "max_iterations": max_iterations,
-                "current_iteration": 0,
-                "max_tokens": max_tokens,
-                "max_history_messages": max_history_messages,
-                "estimated_tokens_used": 0,
-                "warn_iterations_pct": 80,
-                "warn_tokens_pct": 80,
-            },
-        }
-
-        # Inject any additional context from the flow
-        if input_key and isinstance(run.vars.get(input_key), dict):
-            # Merge additional context
-            for k, v in run.vars.get(input_key, {}).items():
-                if k not in ("task", "query"):
-                    agent_vars["context"][k] = v
-
-        # Use START_SUBWORKFLOW effect to run agent durably
-        return StepPlan(
-            node_id=node_id,
-            effect=Effect(
-                type=EffectType.START_SUBWORKFLOW,
-                payload={
-                    "workflow_id": agent.workflow.workflow_id,
-                    "vars": agent_vars,
-                    "async": False,  # Sync: wait for completion
-                },
-                result_key=output_key or f"_flow.{node_id}.result",
-            ),
-            next_node=next_node,
-        )
-
-    return handler
+from abstractruntime.visualflow_compiler.adapters.agent_adapter import *  # noqa: F401,F403