algorhino-anemone 0.1.2__tar.gz → 0.1.4__tar.gz

{algorhino_anemone-0.1.2/src/algorhino_anemone.egg-info → algorhino_anemone-0.1.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: algorhino-anemone
-Version: 0.1.2
+Version: 0.1.4
 Summary: anemone searches trees
 Author-email: Victor Gabillon <victorgabillon@gmail.com>
 License-Expression: GPL-3.0-only
@@ -12,8 +12,8 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: valanga
-Requires-Dist: atomheart
+Requires-Dist: valanga>=0.1.4
+Requires-Dist: atomheart>=0.1.2
 Requires-Dist: rich
 Requires-Dist: sortedcollections>=2.1.0
 Requires-Dist: graphviz

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "algorhino-anemone"
-version = "0.1.2"
+version = "0.1.4"
 description = "anemone searches trees"
 keywords = ["tree", "search"]
 classifiers = [
@@ -14,9 +14,8 @@ classifiers = [
 ]
 requires-python = ">=3.13"
 dependencies = [
-    "valanga",
-"atomheart",
-
+    "valanga>=0.1.4",
+    "atomheart>=0.1.2",
     "rich",
     "sortedcollections>=2.1.0",
     "graphviz",
@@ -236,9 +235,17 @@ extraPaths = ["src"]
 
 [tool.tox]
 min_version = "4.32.0"
-env_list = ["py313", "lint", "typecheck"]
+env_list = ["py313", "lint", "typecheck", "build"]
 isolated_build = true
 
+[tool.tox.env.build]
+deps = ["build", "twine"]
+commands = [
+  ["python", "-m", "build"],
+  ["python", "-m", "twine", "check", "dist/*"],
+]
+
+
 [tool.tox.env_run_base]
 base_python = ["python3.13"]
 
@@ -258,8 +265,8 @@ description = "ruff + pylint (fast, no package build/install)"
 package = "skip"
 deps = [
   # runtime deps needed for pylint imports (match [project].dependencies)
-  "valanga",
-  "atomheart",
+  "valanga>=0.1.4",
+  "atomheart>=0.1.2",
   "rich",
   "sortedcollections>=2.1.0",
   "graphviz",
@@ -283,8 +290,8 @@ description = "mypy + pyright (fast, no package build/install)"
 package = "skip"
 deps = [
   # runtime deps for import resolution
-  "valanga",
-  "atomheart",
+  "valanga>=0.1.4",
+  "atomheart>=0.1.2",
   "rich",
   "sortedcollections>=2.1.0",
   "graphviz",

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4/src/algorhino_anemone.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: algorhino-anemone
-Version: 0.1.2
+Version: 0.1.4
 Summary: anemone searches trees
 Author-email: Victor Gabillon <victorgabillon@gmail.com>
 License-Expression: GPL-3.0-only
@@ -12,8 +12,8 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: valanga
-Requires-Dist: atomheart
+Requires-Dist: valanga>=0.1.4
+Requires-Dist: atomheart>=0.1.2
 Requires-Dist: rich
 Requires-Dist: sortedcollections>=2.1.0
 Requires-Dist: graphviz

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/algorhino_anemone.egg-info/requires.txt

@@ -1,5 +1,5 @@
-valanga
-atomheart
+valanga>=0.1.4
+atomheart>=0.1.2
 rich
 sortedcollections>=2.1.0
 graphviz

algorhino_anemone-0.1.4/src/anemone/basics.py

@@ -0,0 +1,23 @@
+"""Basic types and protocols for Anemone."""
+
+from typing import Annotated, Protocol
+
+from valanga import Color, HasTurn, State
+
+type Seed = Annotated[int, "seed"]
+type TreeDepth = Annotated[int, "Depth level of a node in a tree structure"]
+
+
+class StateWithTurn(State, HasTurn, Protocol):
+    """A `valanga.State` that also exposes turn information."""
+
+    ...
+
+
+class HasBlackAndWhiteTurn(Protocol):
+    """Protocol for state that has black and white turns."""
+
+    @property
+    def turn(self) -> Color:
+        """Return the current player's turn color."""
+        ...

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/factory.py

@@ -56,7 +56,8 @@ def create_tree_and_value_branch_selector[StateT: TurnState](
     args: TreeAndValuePlayerArgs,
     random_generator: Random,
     master_state_evaluator: MasterStateEvaluator,
-    state_representation_factory: RepresentationFactory[ContentRepresentation] | None,
+    state_representation_factory: RepresentationFactory[StateT, ContentRepresentation]
+    | None,
     queue_progress_player: Queue[IsDataclass] | None,
 ) -> TreeAndValueBranchSelector[StateT]:
     """Convenience constructor using the default minmax tree evaluation.
@@ -85,7 +86,8 @@ def create_tree_and_value_branch_selector_with_tree_eval_factory[StateT: TurnSta
     args: TreeAndValuePlayerArgs,
     random_generator: Random,
     master_state_evaluator: MasterStateEvaluator,
-    state_representation_factory: RepresentationFactory[ContentRepresentation] | None,
+    state_representation_factory: RepresentationFactory[StateT, ContentRepresentation]
+    | None,
     node_tree_evaluation_factory: NodeTreeEvaluationFactory[StateT],
     queue_progress_player: Queue[IsDataclass] | None,
 ) -> TreeAndValueBranchSelector[StateT]:

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/node_evaluation/node_direct_evaluation/node_direct_evaluator.py

@@ -89,7 +89,7 @@ class MasterStateEvaluator(Protocol):
 class NodeDirectEvaluator[StateT: State = State]:
     """
     The NodeEvaluator class is responsible for evaluating the value of nodes in a tree structure.
-    It uses a board evaluator and a syzygy evaluator to calculate the value of the nodes.
+    It uses a state evaluator and a syzygy evaluator to calculate the value of the nodes.
     """
 
     master_state_evaluator: MasterStateEvaluator

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/node_evaluation/node_tree_evaluation/node_minmax_evaluation.py

@@ -21,12 +21,12 @@ from random import choice
 from typing import Any, Protocol, Self, runtime_checkable
 
 from valanga import (
-    BoardEvaluation,
     BranchKey,
     Color,
     FloatyStateEvaluation,
     ForcedOutcome,
     OverEvent,
+    StateEvaluation,
     TurnState,
 )
 
@@ -140,7 +140,7 @@ class NodeMinmaxEvaluation[
         return self.value_white_minmax
 
     def set_evaluation(self, evaluation: float) -> None:
-        """sets the evaluation from the board evaluator
+        """Set the evaluation from the state evaluator.
 
         Args:
             evaluation (float): The evaluation value to be set.
@@ -874,8 +874,9 @@ class NodeMinmaxEvaluation[
                     best_branches.append(branch_key)
         return best_branches
 
-    def evaluate(self) -> BoardEvaluation:
-        """Build a BoardEvaluation from current minmax state."""
+    def evaluate(self) -> StateEvaluation:
+        """Build a StateEvaluation from current minmax state."""
+
         if self.over_event.is_over():
             return ForcedOutcome(
                 outcome=self.over_event,

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/node_evaluation/node_tree_evaluation/node_tree_evaluation.py

@@ -1,10 +1,10 @@
 from typing import TYPE_CHECKING, Protocol, Self
 
 from valanga import (
-    BoardEvaluation,
     BranchKey,
     OverEvent,
     State,
+    StateEvaluation,
 )
 
 type BranchSortValue = tuple[float, int, int]
@@ -43,7 +43,7 @@ class NodeTreeEvaluation[StateT: State = State](Protocol):
     value_white_minmax: float | None = None
 
     def set_evaluation(self, evaluation: float) -> None:
-        """sets the evaluation from the board evaluator
+        """Set the evaluation from the state evaluator.
 
         Args:
             evaluation (float): The evaluation value to be set.
@@ -90,8 +90,8 @@ class NodeTreeEvaluation[StateT: State = State](Protocol):
         """Update terminal state based on updated branches."""
         ...
 
-    def evaluate(self) -> BoardEvaluation:
-        """Return a board evaluation for this node."""
+    def evaluate(self) -> StateEvaluation:
+        """Return a state evaluation for this node."""
         ...
 
     def description_tree_visualizer_branch(self, child: "ITreeNode[StateT]") -> str:

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/node_factory/algorithm_node_factory.py

@@ -37,7 +37,9 @@ class AlgorithmNodeFactory[StateT: State = State]:
     """
 
     tree_node_factory: TreeNodeFactory[AlgorithmNode[StateT], StateT]
-    state_representation_factory: RepresentationFactory | None
+    state_representation_factory: (
+        RepresentationFactory[StateT, ContentRepresentation] | None
+    )
     node_tree_evaluation_factory: NodeTreeEvaluationFactory[StateT]
     exploration_index_data_create: node_indices.ExplorationIndexDataFactory[
         AlgorithmNode[StateT], StateT

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/nodes/algorithm_node/algorithm_node.py

@@ -57,7 +57,7 @@ class AlgorithmNode[StateT: State = State]:
             tree_node (TreeNode): The tree node that is wrapped.
             tree_evaluation (NodeTreeEvaluation): The object computing the value.
             exploration_index_data (NodeExplorationData | None): The object storing the information to help the algorithm decide the next nodes to explore.
-            state_representation (StateRepresentation | None): The board representation.
+            state_representation (ContentRepresentation | None): The state representation used for evaluation.
         """
         self.tree_node = tree_node
         self.tree_evaluation = tree_evaluation
@@ -145,9 +145,9 @@ class AlgorithmNode[StateT: State = State]:
         )
 
     @property
-    def all_branches_keys(self) -> BranchKeyGeneratorP:
+    def all_branches_keys(self) -> BranchKeyGeneratorP[BranchKey]:
         """
-        Returns a generator that yields the branch keys for the current board state.
+        Returns a generator that yields the branch keys for the current state.
 
         Returns:
             BranchKeyGenerator: A generator that yields the branch keys.

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/nodes/itree_node.py

@@ -108,7 +108,7 @@ class ITreeNode[StateT: State = State](Protocol):
         """
 
     @property
-    def all_branches_keys(self) -> BranchKeyGeneratorP:
+    def all_branches_keys(self) -> BranchKeyGeneratorP[BranchKey]:
         """
         Get the available branch keys of the node.
 

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/nodes/tree_node.py

@@ -152,7 +152,7 @@ class TreeNode[
         return not self.parent_nodes
 
     @property
-    def all_branches_keys(self) -> BranchKeyGeneratorP:
+    def all_branches_keys(self) -> BranchKeyGeneratorP[BranchKey]:
         """
         Returns a generator that yields the branch keys for the current state.
 

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/recommender_rule/recommender_rule.py

@@ -16,9 +16,10 @@ Example usage:
 from dataclasses import dataclass
 from enum import Enum
 from random import Random
-from typing import Literal, Mapping, Protocol
+from typing import Literal, Protocol
 
 from valanga import BranchKey, State
+from valanga.policy import BranchPolicy
 
 from anemone.nodes.algorithm_node.algorithm_node import (
     AlgorithmNode,
@@ -26,15 +27,6 @@ from anemone.nodes.algorithm_node.algorithm_node import (
 from anemone.utils.small_tools import softmax
 
 
-@dataclass(frozen=True, slots=True)
-class BranchPolicy:
-    """
-    Represents a probability distribution over branches.
-    """
-
-    probs: Mapping[BranchKey, float]  # should sum to ~1.0
-
-
 def sample_from_policy(policy: BranchPolicy, rng: Random) -> BranchKey:
     """Sample a branch key from a probability policy using a RNG."""
     branches = list(policy.probs.keys())

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/tree_and_value_branch_selector.py

@@ -5,8 +5,9 @@ from queue import Queue
 from random import Random
 
 from valanga import TurnState
+from valanga.policy import Recommendation
 
-from anemone.basics import BranchRecommendation, Seed
+from anemone.basics import Seed
 from anemone.progress_monitor.progress_monitor import (
     AllStoppingCriterionArgs,
 )
@@ -43,14 +44,12 @@ class TreeAndValueBranchSelector[StateT: TurnState = TurnState]:
     recommend_branch_after_exploration: recommender_rule.AllRecommendFunctionsArgs
     queue_progress_player: Queue[IsDataclass] | None
 
-    def select_branch(
-        self, state: StateT, selection_seed: Seed
-    ) -> BranchRecommendation:
+    def select_branch(self, state: StateT, selection_seed: Seed) -> Recommendation:
         """
         Selects the best branch based on the tree and value strategy.
 
         Args:
-        - board: The current board state.
+        - state: The current state to explore.
         - selection_seed: The seed used for randomization during branch selection.
 
         Returns:
@@ -59,7 +58,7 @@ class TreeAndValueBranchSelector[StateT: TurnState = TurnState]:
         tree_exploration: TreeExploration = self.create_tree_exploration(state=state)
         self.random_generator.seed(selection_seed)
 
-        branch_recommendation: BranchRecommendation = tree_exploration.explore(
+        branch_recommendation: Recommendation = tree_exploration.explore(
             random_generator=self.random_generator
         ).branch_recommendation
 

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/tree_exploration.py

@@ -20,9 +20,10 @@ from queue import Queue
 from random import Random
 from typing import TYPE_CHECKING, Any, Callable
 
-from valanga import BoardEvaluation, BranchKey, PlayerProgressMessage, State, TurnState
+from valanga import BranchKey, PlayerProgressMessage, State, StateEvaluation, TurnState
+from valanga.game import BranchName
+from valanga.policy import Recommendation
 
-from anemone.basics import BranchRecommendation
 from anemone.nodes.algorithm_node.algorithm_node import AlgorithmNode
 from anemone.progress_monitor.progress_monitor import (
     AllStoppingCriterionArgs,
@@ -39,7 +40,7 @@ from . import tree_manager as tree_man
 from .trees.factory import ValueTreeFactory
 
 if TYPE_CHECKING:
-    from anemone.recommender_rule.recommender_rule import BranchPolicy
+    from valanga.policy import BranchPolicy
 
 
 @dataclass
@@ -48,21 +49,22 @@ class TreeExplorationResult[NodeT: AlgorithmNode[Any] = AlgorithmNode[Any]]:
     Tree Exploration Result holds the result of a tree exploration.
     """
 
-    branch_recommendation: BranchRecommendation
+    branch_recommendation: Recommendation
     tree: trees.Tree[NodeT]
 
 
 def compute_child_evals[StateT: State](
     root: AlgorithmNode[StateT],
-) -> dict[BranchKey, BoardEvaluation]:
+) -> dict[BranchName, StateEvaluation]:
     """Compute evaluations for each existing child branch."""
-    evals: dict[BranchKey, BoardEvaluation] = {}
+    evals: dict[BranchName, StateEvaluation] = {}
     for bk, child in root.branches_children.items():
         if child is None:
             continue
 
         # Use whatever your canonical per-node evaluation is:
-        evals[bk] = child.tree_evaluation.evaluate()
+        bk_name = root.state.branch_name_from_key(bk)
+        evals[bk_name] = child.tree_evaluation.evaluate()
     return evals
 
 
@@ -188,15 +190,16 @@ class TreeExploration[NodeT: AlgorithmNode[Any] = AlgorithmNode[Any]]:
             policy, random_generator
         )
 
+        best_branch_name = self.tree.root_node.state.branch_name_from_key(best_branch)
         self.tree_manager.print_best_line(
             tree=self.tree
         )  # todo maybe almost best chosen line no?
 
-        branch_recommendation = BranchRecommendation(
-            branch_key=best_branch,
+        branch_recommendation = Recommendation(
+            recommended_name=best_branch_name,
             evaluation=self.tree.root_node.tree_evaluation.evaluate(),
             policy=policy,
-            child_evals=compute_child_evals(self.tree.root_node),
+            branch_evals=compute_child_evals(self.tree.root_node),
         )
 
         tree_exploration_result: TreeExplorationResult[NodeT] = TreeExplorationResult(

{algorhino_anemone-0.1.2 → algorhino_anemone-0.1.4}/src/anemone/tree_manager/tree_manager.py

@@ -70,7 +70,7 @@ class TreeManager[
         Returns:
             The tree expansion object.
         """
-        # The parent board is copied, we only copy the stack (history of previous board) if the depth is smaller than 2
+        # The parent state is copied; we only copy the stack (history of previous states) if the depth is smaller than 2.
         # Having the stack information allows checking for draw by repetition.
         # To limit computation we limit copying it all the time. The resulting policy will only be aware of immediate
         # risk of draw by repetition
@@ -117,7 +117,7 @@ class TreeManager[
             The tree expansion object.
         """
 
-        # Creation of the child node. If the board already exited in another node, that node is returned as child_node.
+        # Creation of the child node. If the state already existed in another node, that node is returned as child_node.
         tree_depth: int = parent_node.tree_depth + 1
         state_tag: StateTag = state.tag
 

algorhino_anemone-0.1.2/src/anemone/basics.py

@@ -1,36 +0,0 @@
-"""Basic types and protocols for Anemone."""
-
-from dataclasses import dataclass
-from typing import Annotated, Mapping, Protocol
-
-from valanga import BoardEvaluation, BranchKey, Color, HasTurn, State
-
-from anemone.recommender_rule.recommender_rule import BranchPolicy
-
-type Seed = Annotated[int, "seed"]
-type TreeDepth = Annotated[int, "Depth level of a node in a tree structure"]
-
-
-class StateWithTurn(State, HasTurn, Protocol):
-    """A `valanga.State` that also exposes turn information."""
-
-    ...
-
-
-@dataclass(frozen=True, slots=True)
-class BranchRecommendation:
-    """A recommendation for a specific branch in a tree node."""
-
-    branch_key: BranchKey
-    evaluation: BoardEvaluation | None = None
-    policy: BranchPolicy | None = None
-    child_evals: Mapping[BranchKey, BoardEvaluation] | None = None
-
-
-class HasBlackAndWhiteTurn(Protocol):
-    """Protocol for state that has black and white turns."""
-
-    @property
-    def turn(self) -> Color:
-        """Return the current player's turn color."""
-        ...