synth-ai 0.2.9.dev5__py3-none-any.whl → 0.2.9.dev6__py3-none-any.whl

examples/swe/task_app/hosted/envs/crafter/policy.py

@@ -0,0 +1,478 @@
+from __future__ import annotations
+
+from abc import ABC, abstractmethod
+from typing import Any
+
+from .react_agent import CrafterReActAgent
+from .tools import TOOLS_SCHEMA
+
+
+# Define Policy base class here to avoid circular import
+class Policy(ABC):
+    """Base class for environment-specific policies."""
+
+    @abstractmethod
+    def prepare_inference_request(
+        self, observation: dict[str, Any], history: list[dict[str, Any]] = None
+    ) -> tuple[list[dict[str, Any]], list[dict[str, Any]] | None]:
+        """Prepare an inference request."""
+        pass
+
+    @abstractmethod
+    def parse_model_response(
+        self, response: str, observation: dict[str, Any]
+    ) -> list[dict[str, Any]]:
+        """Parse model response into tool calls."""
+        pass
+
+
+# (imports moved to top of file to satisfy linter)
+
+
+class CrafterPolicy(Policy):
+    """Thin policy scaffold for Crafter using the ReAct agent prompts.
+
+    This class does not run inference itself. It prepares an inference request
+    (messages and optional tools schema) that the Task App can send to the
+    inference service, and provides helpers to parse the model response into
+    environment tool calls.
+    """
+
+    name: str = "crafter-react"
+
+    def __init__(self, inference_url: str, model: str | None = None) -> None:
+        self.inference_url = inference_url
+        self.model = model
+        self.use_tools = True
+        # Sampling parameters (populated via initialize(config))
+        self.temperature: float | None = None
+        self.top_p: float | None = None
+        self.max_tokens: int | None = None
+        # Thinking controls (populated via initialize(config))
+        self.thinking_mode: str | None = None
+        self.thinking_budget: int | None = None
+        # Rolling conversation and action history for non-Markov policies
+        self.history_messages: list[dict[str, str]] = []  # chat-style without system
+        self.turn_index: int = 0
+        self.trajectory_history: list[dict[str, Any]] = []  # env/policy step records
+
+    async def initialize(self, config: dict[str, Any]) -> None:
+        if "inference_url" in config:
+            self.inference_url = config["inference_url"]
+        if "model" in config:
+            self.model = config["model"]
+        if "use_tools" in config:
+            self.use_tools = bool(config["use_tools"])
+        # Adopt sampling params from policy config (trainer passes these through)
+        if "temperature" in config:
+            self.temperature = float(config["temperature"])  # fail fast on bad types
+        if "top_p" in config:
+            self.top_p = float(config["top_p"])  # fail fast on bad types
+        if "max_tokens" in config:
+            self.max_tokens = int(config["max_tokens"])  # fail fast on bad types
+        # Thinking mode/budget forwarded into vLLM request (mirrors Wordle policy)
+        if "thinking_mode" in config:
+            self.thinking_mode = str(config["thinking_mode"])  # expect "think" or "no_think"
+        if "thinking_budget" in config and config["thinking_budget"] is not None:
+            self.thinking_budget = int(config["thinking_budget"])  # number of tokens inside <think>
+        if self.thinking_budget is None:
+            try:
+                if "openai.com" not in (self.inference_url or "").lower():
+                    self.thinking_budget = 1028
+            except Exception:
+                self.thinking_budget = 1028
+        # Reset state on (re)initialize
+        self.history_messages = []
+        self.turn_index = 0
+        self.trajectory_history = []
+
+    def _append_user_observation(self, observation_text: str) -> None:
+        self.history_messages.append({"role": "user", "content": observation_text})
+        self.turn_index += 1
+
+    def _append_assistant_turn(
+        self,
+        assistant_text: str | None,
+        tool_calls: list[dict[str, Any]] | None,
+        env_result: dict[str, Any] | None,
+    ) -> None:
+        # Record assistant content (if any)
+        if assistant_text is not None:
+            self.history_messages.append({"role": "assistant", "content": assistant_text})
+        # Keep structured step record for training/analysis
+        record: dict[str, Any] = {"turn": self.turn_index}
+        if tool_calls is not None:
+            record["tool_calls"] = tool_calls
+        if env_result is not None:
+            record["env_result"] = env_result
+        self.trajectory_history.append(record)
+
+    def build_inference_request(
+        self,
+        observation_text: str,
+        history: list[dict[str, Any]] | None = None,
+        turn: int | None = None,
+        image_parts: list[dict[str, Any]] | None = None,
+    ) -> dict[str, Any]:
+        messages = CrafterReActAgent.build_messages(
+            observation=observation_text,
+            history=history,
+            turn=turn,
+            image_parts=image_parts,
+        )
+        payload: dict[str, Any] = {
+            "messages": messages,
+        }
+        if self.model is not None:
+            payload["model"] = self.model
+        # Thinking controls
+        if self.thinking_mode is None and "openai.com" not in (self.inference_url or "").lower():
+            self.thinking_mode = "think"
+        if self.thinking_mode is not None:
+            payload["thinking_mode"] = self.thinking_mode
+        if self.thinking_budget is None and "openai.com" not in (self.inference_url or "").lower():
+            self.thinking_budget = 1028
+        if self.thinking_budget is not None:
+            payload["thinking_budget"] = self.thinking_budget
+        # Inject sampling parameters if set via initialize(config)
+        if self.max_tokens is not None:
+            # Use max_completion_tokens for newer models, max_tokens for older ones
+            if self.model and ("gpt-5" in self.model):
+                payload["max_completion_tokens"] = self.max_tokens
+            else:
+                payload["max_tokens"] = self.max_tokens
+        if self.temperature is not None:
+            payload["temperature"] = self.temperature
+        if self.top_p is not None:
+            payload["top_p"] = self.top_p
+        if self.use_tools:
+            payload["tools"] = TOOLS_SCHEMA
+            payload["tool_choice"] = "required"
+            # Ensure the inference server injects family-specific stop sequences
+            # to terminate immediately after the first tool call for compliance.
+            payload["stop_after_tool_calls"] = 1
+        return payload
+
+    @staticmethod
+    def parse_response_to_tool_calls(
+        response: dict[str, Any],
+        use_tools: bool = True,
+    ) -> list[dict[str, Any]]:
+        """Turn an inference response into environment tool calls.
+
+        - If tools were used, expect tool_calls-compatible output and forward as-is
+          in our simple JSON format: {"tool_name": str, "arguments": {...}}.
+        - If no tools, parse plain-text actions using CrafterReActAgent parser and
+          wrap them into a single interact_many tool call.
+        """
+        # First check if we got actual tool calls
+        choices = response.get("choices", [])
+        tool_calls: list[dict[str, Any]] = []
+
+        for choice in choices:
+            msg = choice.get("message", {})
+            if "tool_calls" in msg and msg["tool_calls"] is not None:
+                for tc in msg["tool_calls"]:
+                    if tc is None:
+                        continue
+                    # Handle both OpenAI format and simplified format
+                    if "function" in tc:
+                        # Standard OpenAI format
+                        tool_calls.append(
+                            {
+                                "tool_name": tc["function"]["name"],
+                                "arguments": tc["function"]["arguments"],
+                            }
+                        )
+                    elif "name" in tc:
+                        # Simplified format from our vLLM service
+                        tool_calls.append(
+                            {
+                                "tool_name": tc["name"],
+                                "arguments": tc["arguments"],
+                            }
+                        )
+
+        # If we got tool calls, return them
+        if tool_calls:
+            # Normalize common degenerate pattern ["move_right", "do"] when nothing is nearby.
+            # If previous env_result indicates no interaction target, drop trailing 'do'.
+            normalized: list[dict[str, Any]] = []
+            for tc in tool_calls:
+                if tc and isinstance(tc, dict) and tc.get("tool_name") == "interact_many":
+                    args = tc.get("arguments")
+                    if isinstance(args, str):
+                        try:
+                            import json
+
+                            args = json.loads(args)
+                        except (json.JSONDecodeError, ValueError):
+                            args = {}
+                    actions = []
+                    if isinstance(args, dict):
+                        maybe_actions = args.get("actions")
+                        if isinstance(maybe_actions, list):
+                            actions = maybe_actions
+                    # Simple heuristic: avoid repeating same pair; avoid 'do' with no context
+                    if len(actions) == 2 and actions[0] == "move_right" and actions[1] == "do":
+                        actions = ["move_right"]
+                    normalized.append(
+                        {"tool_name": "interact_many", "arguments": {"actions": actions or []}}
+                    )
+                else:
+                    normalized.append(tc)
+            return normalized
+
+        # Otherwise, parse plain text content for actions
+        text = ""
+        for choice in choices:
+            msg = choice.get("message", {})
+            content = msg.get("content", "")
+            if content:
+                text = content
+                break
+
+        if text:
+            # Try to parse actions from the text
+            from .shared import parse_actions
+
+            actions = parse_actions(text)
+            if actions:
+                # Wrap actions in interact_many tool call
+                return [{"tool_name": "interact_many", "arguments": {"actions": actions}}]
+
+        # No actions found
+        return []
+
+    async def step(
+        self,
+        observation_text: str,
+        state: dict[str, Any] | None = None,
+        metadata: dict[str, Any] | None = None,
+    ) -> tuple[list[dict[str, Any]], dict[str, Any]]:
+        """Stateful step: update policy history and prepare inference request.
+
+        Inputs (via metadata, optional):
+        - "prev_assistant_text": str — assistant text from prior turn
+        - "prev_tool_calls": List[Dict] — tool calls executed last turn
+        - "prev_env_result": Dict — env step result for prior tool calls
+        - "prev_inference_response": Dict — raw LLM response; if present and
+          use_tools=False, we record assistant_text parsed from content.
+
+        Returns (tool_calls, meta):
+        - tool_calls: empty list; coordinator should call inference and then use
+          parse_response_to_tool_calls() to derive tool_calls
+        - meta: { inference_url, inference_request, turn_index, history_len }
+        """
+        # If caller provided results from previous cycle, record them first
+        if metadata is not None:
+            prev_assistant_text: str | None = None
+            prev_tool_calls: list[dict[str, Any]] | None = None
+            prev_env_result: dict[str, Any] | None = None
+            if "prev_assistant_text" in metadata:
+                prev_assistant_text = metadata["prev_assistant_text"]
+            if "prev_tool_calls" in metadata:
+                prev_tool_calls = metadata["prev_tool_calls"]
+            if "prev_env_result" in metadata:
+                prev_env_result = metadata["prev_env_result"]
+            if (
+                prev_assistant_text is not None
+                or prev_tool_calls is not None
+                or prev_env_result is not None
+            ):
+                self._append_assistant_turn(prev_assistant_text, prev_tool_calls, prev_env_result)
+
+        # Append current observation as the next user message (internal history only)
+        self._append_user_observation(observation_text)
+
+        # Build user message by combining the current observation text
+        # (formatted surroundings/inventory) with the previous 3 tool calls as context.
+        # Most recent first.
+        lines: list[str] = []
+
+        def _format_tool_call_line_for_context(
+            tool_name: str, arguments: Any, max_chars: int = 500
+        ) -> str:
+            import json as _json
+
+            # Render arguments compactly, then clip to max_chars
+            if isinstance(arguments, dict | list):
+                try:
+                    rendered = _json.dumps(arguments, ensure_ascii=False, separators=(",", ":"))
+                except Exception:
+                    rendered = str(arguments)
+            elif isinstance(arguments, str):
+                rendered = arguments
+            else:
+                rendered = str(arguments)
+            if isinstance(rendered, str) and len(rendered) > max_chars:
+                rendered = rendered[:max_chars]
+            return f"- {tool_name}: {rendered}"
+
+        # Prefer pulling from trajectory_history (accumulates over turns)
+        for record in reversed(self.trajectory_history):
+            if len(lines) >= 3:
+                break
+            tc_list = record.get("tool_calls")
+            if not tc_list:
+                continue
+            # Use the first tool call for that turn if multiple exist
+            tc = tc_list[0] if isinstance(tc_list, list) and tc_list else None
+            if not isinstance(tc, dict):
+                continue
+            name = tc.get("tool_name") or tc.get("name") or "unknown"
+            args = tc.get("arguments")
+            lines.append(_format_tool_call_line_for_context(name, args))
+
+        # If trajectory history is empty (first few turns), fall back to metadata once
+        if not lines and metadata is not None and metadata.get("prev_tool_calls"):
+            calls: list[dict[str, Any]] = metadata["prev_tool_calls"]
+            for call in reversed(calls):
+                if len(lines) >= 3:
+                    break
+                if not isinstance(call, dict):
+                    continue
+                name = call.get("tool_name") or call.get("name") or "unknown"
+                args = call.get("arguments")
+                lines.append(_format_tool_call_line_for_context(name, args))
+
+        context_text = "Previous tool calls (most recent first):\n" + (
+            "\n".join(lines) if lines else "- none"
+        )
+
+        # Combine observation with context so the model always sees surroundings/inventory
+        combined_text = f"{observation_text}\n\n{context_text}"
+
+        raw_observation: dict[str, Any] | None = None
+        if metadata is not None:
+            raw_candidate = metadata.get("raw_observation")
+            if isinstance(raw_candidate, dict):
+                raw_observation = raw_candidate
+        image_parts = self._extract_image_parts(raw_observation)
+
+        payload = self.build_inference_request(
+            combined_text,
+            history=[],  # no prior user/assistant history
+            turn=self.turn_index,
+            image_parts=image_parts,
+        )
+        # print("Debugging only:; ", payload)
+        meta_out = {
+            "inference_url": self.inference_url,
+            "inference_request": payload,
+            "turn_index": self.turn_index,
+            "history_len": len(self.history_messages),
+        }
+        return [], meta_out
+
+    def state_dict(self) -> dict[str, Any]:
+        return {
+            "turn_index": self.turn_index,
+            "history_messages": self.history_messages,
+            "trajectory_history": self.trajectory_history,
+        }
+
+    def load_state_dict(self, state: dict[str, Any]) -> None:
+        self.turn_index = int(state["turn_index"])
+        self.history_messages = state["history_messages"]
+        self.trajectory_history = state["trajectory_history"]
+
+    async def serialize(self) -> dict[str, Any]:
+        return {
+            "name": self.name,
+            "config": {
+                "inference_url": self.inference_url,
+                "model": self.model,
+                "use_tools": self.use_tools,
+            },
+            "state": self.state_dict(),
+        }
+
+    @classmethod
+    async def deserialize(cls, payload: dict[str, Any]) -> CrafterPolicy:
+        config = payload["config"]
+        state = payload["state"]
+        policy = cls(
+            inference_url=config["inference_url"],
+            model=config.get("model"),
+        )
+        policy.use_tools = bool(config["use_tools"])
+        policy.load_state_dict(state)
+        return policy
+
+    async def terminate(self) -> None:
+        return None
+
+    def prepare_inference_request(
+        self, observation: dict[str, Any], history: list[dict[str, Any]] = None
+    ) -> tuple[list[dict[str, Any]], list[dict[str, Any]] | None]:
+        """Prepare an inference request (implementing abstract method)."""
+        # Format observation with rich contextual information
+        observation_text = self._format_observation_for_llm(observation)
+        image_parts = self._extract_image_parts(observation)
+
+        # Build messages (observation_text already formatted; no raw matrices)
+        messages = CrafterReActAgent.build_messages(
+            observation=observation_text,
+            history=history,
+            turn=self.turn_index,
+            image_parts=image_parts,
+        )
+
+        # Return messages and tools schema
+        tools = TOOLS_SCHEMA if self.use_tools else None
+        return messages, tools
+
+    def _format_observation_for_llm(self, observation: dict[str, Any]) -> str:
+        """Format observation with rich contextual information for the LLM using the shared formatter."""
+        from .shared import format_observation
+
+        # Get the observation data (could be nested)
+        obs_data = observation.get("observation", observation)
+
+        # Ensure obs_data is a dict for safe access
+        if not isinstance(obs_data, dict):
+            return f"Observation: {str(observation)}"
+
+        # Use the shared format_observation function with step information
+        step_idx = observation.get("step_idx", 0)
+        max_steps = 100  # Default max steps, could be made configurable
+
+        # Get additional info from the observation wrapper
+        info = observation.get("info", {})
+        if isinstance(info, dict) and "health" in info and "health" not in obs_data:
+            obs_data = dict(obs_data)  # Make a copy
+            obs_data["health"] = info["health"]
+
+        return format_observation(obs_data, step_count=step_idx, max_steps=max_steps)
+
+    def _extract_image_parts(
+        self, observation: dict[str, Any] | None
+    ) -> list[dict[str, Any]]:
+        """Crafter policy uses text-only prompts; do not attach image parts."""
+
+        return []
+
+    def parse_model_response(
+        self, response: str, observation: dict[str, Any]
+    ) -> list[dict[str, Any]]:
+        """Parse model response into tool calls (implementing abstract method).
+
+        Note: Despite the type hint, vLLM actually returns a dict response,
+        not a string. We handle both cases.
+        """
+        # Handle dict response from vLLM (the actual case)
+        if isinstance(response, dict):
+            return self.parse_response_to_tool_calls(response, self.use_tools)
+
+        # Handle string response (fallback case for raw text)
+        if isinstance(response, str):
+            actions = CrafterReActAgent.parse_actions_from_response(response)
+            if actions:
+                return [{"tool_name": "interact_many", "arguments": {"actions": actions}}]
+
+        # Default empty response
+        return []
+
+
+__all__ = ["CrafterPolicy"]

examples/swe/task_app/hosted/envs/crafter/react_agent.py

@@ -0,0 +1,108 @@
+"""Crafter ReAct agent: system prompt and message assembly.
+
+This agent encapsulates the Crafter-specific system prompt and helpers to
+construct OpenAI-style message lists. Response parsing delegates to shared
+utilities to keep a single parser.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from .shared import parse_actions
+
+
+class CrafterReActAgent:
+    """Lightweight ReAct-style agent surface for Crafter prompts."""
+
+    @staticmethod
+    def get_system_prompt() -> str:
+        return (
+            "You are playing Crafter, a survival game by Danijar Hafner. Your goal is to collect resources, "
+            "craft tools, survive, and unlock achievements.\n\n"
+            "Core rules:\n"
+            "- The world contains trees (wood), stone, coal, iron, plants, cows, zombies, and water.\n"
+            "- Movement constraints: you cannot walk onto blocking tiles: tree, stone, water, lava, coal, iron. Navigate around obstacles.\n"
+            "- You start with empty hands and low health/hunger.\n"
+            "- Interact ('do') only when adjacent to a resource (tree, stone, cow, zombie, etc.).\n"
+            "- Movement is essential: you can and should move multiple steps in one turn to explore effectively.\n"
+            "- Achievements are unlocked by collecting resources, crafting tools, placing objects, fighting, and surviving longer.\n\n"
+            "Key strategies:\n"
+            "1. Begin by moving around to find trees. Use 'do' to collect wood when adjacent.\n"
+            "2. Craft a wood pickaxe as soon as you have enough wood ('make_wood_pickaxe').\n"
+            "3. Use the pickaxe to gather stone, then craft a stone pickaxe. Progress to iron tools as you find iron.\n"
+            "4. Build a table ('place_table') to unlock more crafting options (furnace, sword, etc.).\n"
+            "5. Manage hunger by collecting and eating plants or interacting with cows.\n"
+            "6. Fight zombies with a sword for achievements and resources.\n"
+            "7. Survive by balancing exploration, combat, and resource gathering.\n\n"
+            "8. Keep moving to discover new resources and stay alive. If you're in the middle of nowhere, take 5-8 consecutive move-related actions to explore and see what's outside your field of view. Don't delay exploration when it's the right move.\n\n"
+            "Achievements to aim for:\n"
+            "- Collecting resources (wood, stone, coal, iron, plants).\n"
+            "- Crafting tools (wood/stone/iron pickaxe, wood/stone/iron sword).\n"
+            "- Placing structures (table, furnace, plant).\n"
+            "- Combat (killing a cow or zombie).\n"
+            "- Survival milestones (staying alive over time).\n\n"
+            "Action policy:\n"
+            "- Always return a single tool call: interact_many({actions: [...]})\n"
+            "- Use 2–5 actions per call; prefer long movement sequences to explore.\n"
+            "- Mix in 'do' only when it makes sense (tree, stone, animal, enemy nearby).\n"
+            "- Do not spam the same exact sequence twice in a row—explore in varied directions.\n\n"
+            "Available actions: noop, move_up, move_down, move_left, move_right, do (interact), sleep, "
+            "place_stone, place_table, place_furnace, place_plant, make_wood_pickaxe, make_stone_pickaxe, "
+            "make_iron_pickaxe, make_wood_sword, make_stone_sword, make_iron_sword\n"
+        )
+
+    @staticmethod
+    def get_system_prompt_with_tools() -> str:
+        """System prompt for tool-based interaction (e.g., Qwen3 models)."""
+        return (
+            "You are playing Crafter, a survival game by Danijar Hafner. Your goal is to collect resources, "
+            "craft tools, survive, and unlock achievements.\n\n"
+            "Rules & world:\n"
+            "- Explore by chaining multiple movement actions in one turn.\n"
+            "- You cannot walk onto blocking tiles: tree, stone, water, lava, coal, iron. Plan routes around obstacles.\n"
+            "- Use 'do' intentionally when standing next to resources (trees, stone, cows, zombies, etc.).\n"
+            "- Achievements come from collecting, crafting, building, fighting, and surviving.\n\n"
+            "Strategy path:\n"
+            "1. Move around to find trees → 'do' to collect wood.\n"
+            "2. Craft a wood pickaxe.\n"
+            "3. Gather stone → craft stone pickaxe.\n"
+            "4. Place a table → unlock furnace and swords.\n"
+            "5. Fight enemies (cow/zombie) with swords for achievements.\n"
+            "6. Keep moving to discover new resources and stay alive. If you're in the middle of nowhere, take 5-8 consecutive move-related actions to explore and see what's outside your field of view. Don't delay exploration when it's the right move.\n\n"
+            "You must use the 'interact_many' tool to perform actions in the game. "
+            "This tool accepts an array of 1–5 actions to execute sequentially. Prefer sequences like "
+            "[move_up, move_up, move_left, do] instead of single steps.\n\n"
+            "Available actions: noop, move_up, move_down, move_left, move_right, do (interact), sleep, "
+            "place_stone, place_table, place_furnace, place_plant, make_wood_pickaxe, make_stone_pickaxe, "
+            "make_iron_pickaxe, make_wood_sword, make_stone_sword, make_iron_sword\n\n"
+            "Always call the interact_many tool with your chosen actions. Do not write plain text actions.\n"
+        )
+
+    @staticmethod
+    def build_messages(
+        observation: str,
+        history: list[dict[str, Any]] | None = None,
+        turn: int | None = None,
+        image_parts: list[dict[str, Any]] | None = None,
+    ) -> list[dict[str, Any]]:
+        """Construct OpenAI-style messages list for vLLM generation."""
+        msgs: list[dict[str, Any]] = [
+            {"role": "system", "content": CrafterReActAgent.get_system_prompt()}
+        ]
+        if history:
+            msgs.extend(history)
+        user_content: Any
+        if image_parts:
+            user_content = [{"type": "text", "text": observation}] + list(image_parts)
+        else:
+            user_content = observation
+        msgs.append({"role": "user", "content": user_content})
+        return msgs
+
+    @staticmethod
+    def parse_actions_from_response(response_text: str) -> list[str]:
+        return parse_actions(response_text)
+
+
+__all__ = ["CrafterReActAgent"]