src/dooable/graph.py

"""Finite executable DAGs with one terminal node per canonical outcome.

The graph is the declared support of an experiment. A root-to-terminal path
records every action, including parent selection and termination. Construction
budgets are enforced before edges enter the graph.
"""

from dataclasses import dataclass, field, asdict
from collections import deque
from pathlib import Path
import json, math


@dataclass
class Node:
    """A resource-augmented state or unique canonical terminal outcome."""

    id: str
    outcome: str | None = None
    features: list[float] = field(default_factory=list)
    metadata: dict = field(default_factory=dict)


@dataclass
class Edge:
    """One admissible action with an additive nonnegative execution cost."""

    id: str
    source: str
    target: str
    cost: float = 0.0
    action: dict = field(default_factory=dict)


@dataclass
class Graph:
    """Finite rooted DAG with validated topology and indexed adjacency lists."""

    nodes: list[Node]
    edges: list[Edge]
    root: str = "root"
    metadata: dict = field(default_factory=dict)

    def __post_init__(self):
        self.node_index = {n.id: i for i, n in enumerate(self.nodes)}
        if len(self.node_index) != len(self.nodes):
            raise ValueError("Duplicate node ID")
        if self.root not in self.node_index:
            raise ValueError("Root is missing")
        self.root_index = self.node_index[self.root]
        self.incoming = [[] for _ in self.nodes]
        self.outgoing = [[] for _ in self.nodes]
        if len({e.id for e in self.edges}) != len(self.edges):
            raise ValueError("Duplicate edge ID")
        for i, e in enumerate(self.edges):
            if not math.isfinite(e.cost) or e.cost < 0:
                raise ValueError("Costs must be finite and nonnegative")
            if e.source not in self.node_index or e.target not in self.node_index:
                raise ValueError("Unknown edge endpoint")
            self.outgoing[self.node_index[e.source]].append(i)
            self.incoming[self.node_index[e.target]].append(i)
        if self.incoming[self.root_index]:
            raise ValueError("Root has incoming edges")
        degree = [len(x) for x in self.incoming]
        queue = deque(i for i, d in enumerate(degree) if d == 0)
        order = []
        while queue:
            u = queue.popleft()
            order.append(u)
            for ei in self.outgoing[u]:
                v = self.node_index[self.edges[ei].target]
                degree[v] -= 1
                if degree[v] == 0:
                    queue.append(v)
        if len(order) != len(self.nodes):
            raise ValueError("Executable graph contains a cycle")
        self.order = order
        self.terminals = {
            n.outcome: i for i, n in enumerate(self.nodes) if n.outcome is not None
        }
        if len(self.terminals) != sum(n.outcome is not None for n in self.nodes):
            raise ValueError("Each outcome must have one terminal node")
        if not self.terminals:
            raise ValueError("No terminal outcomes")
        reached = {self.root_index}
        for u in order:
            if u not in reached:
                raise ValueError(f"Unreachable node {self.nodes[u].id}")
            for ei in self.outgoing[u]:
                reached.add(self.node_index[self.edges[ei].target])
            if self.nodes[u].outcome is not None and self.outgoing[u]:
                raise ValueError("Terminal outcome has outgoing edges")
            if not self.outgoing[u] and self.nodes[u].outcome is None:
                raise ValueError("Nonterminal dead end")

    def save(self, path):
        """Write graph structure, features, and metadata to a JSON file."""
        Path(path).parent.mkdir(parents=True, exist_ok=True)
        Path(path).write_text(json.dumps(asdict(self), indent=2))

    @classmethod
    def load(cls, path):
        """Read graph JSON and reconstruct all topology checks and indices."""
        d = json.loads(Path(path).read_text())
        return cls(
            [Node(**n) for n in d["nodes"]],
            [Edge(**e) for e in d["edges"]],
            d["root"],
            d.get("metadata", {}),
        )


def toy_graph(multiplicity=8):
    """Two equally rewarded outcomes, with unequal route counts and costs."""
    nodes = [Node("root", features=[0, 0, 0])]
    edges = []
    for j in range(multiplicity):
        nodes.append(Node(f"a{j}", features=[1, j / max(multiplicity, 1), 0]))
        edges.extend(
            [
                Edge(f"ra{j}", "root", f"a{j}", 0.2),
                Edge(f"at{j}", f"a{j}", "A", 0.3 + j * 0.25),
            ]
        )
    nodes.extend(
        [
            Node("b", features=[1, 0, 1]),
            Node("A", "A", [2, 0, 0]),
            Node("B", "B", [2, 0, 1]),
        ]
    )
    edges.extend([Edge("rb", "root", "b", 0.2), Edge("bt", "b", "B", 0.3)])
    return Graph(
        nodes,
        edges,
        metadata={"kind": "multiplicity", "multiplicity": multiplicity, "max_edges": 2},
    )


def grid_graph(width=8, budget=12):
    """A budgeted lattice with canonical endpoint merging across path lengths."""
    nodes = [Node("root", features=[0, 0, 0, 1, 0])]
    edges = []
    states = {(0, 0, 0): "s0_0_0"}
    nodes.append(Node("s0_0_0", features=[0, 0, 0, 0, 0]))
    edges.append(Edge("start", "root", "s0_0_0", 0, {"kind": "start"}))
    outcomes = set()
    for k in range(budget + 1):
        current = [(s, n) for s, n in states.items() if s[2] == k]
        for (x, y, _), sid in current:
            out = f"{x},{y}"
            tid = "t" + out
            if out not in outcomes:
                nodes.append(Node(tid, out, [x / width, y / width, 0, 0, 1]))
                outcomes.add(out)
            edges.append(Edge("e" + str(len(edges)), sid, tid, 0, {"kind": "stop"}))
            if k == budget:
                continue
            for dx, dy in [(1, 0), (0, 1), (-1, 0), (0, -1)]:
                xx, yy = x + dx, y + dy
                if not (0 <= xx < width and 0 <= yy < width):
                    continue
                # A narrow passage adds an interpretable path cost.
                cost = 1.0 + 0.7 * (xx == width // 2 and yy != width // 2)
                key = (xx, yy, k + 1)
                nid = f"s{xx}_{yy}_{k+1}"
                if key not in states:
                    states[key] = nid
                    nodes.append(
                        Node(
                            nid,
                            features=[
                                xx / width,
                                yy / width,
                                (k + 1) / max(budget, 1),
                                0,
                                0,
                            ],
                        )
                    )
                edges.append(
                    Edge(
                        "e" + str(len(edges)),
                        sid,
                        nid,
                        cost,
                        {"kind": "move", "dx": dx, "dy": dy},
                    )
                )
    return Graph(
        nodes, edges, metadata={"kind": "grid", "width": width, "budget": budget}
    )


def string_graph(length=4, budget=2):
    """Binary-string editing with two parents and canonical terminal merging."""
    if length < 1 or budget < 0:
        raise ValueError("String length must be positive and budget nonnegative")

    def feat(string, depth=0, terminal=False):
        return [float(x) for x in string] + [
            depth / max(1, budget),
            float(terminal),
            0.0,
        ]

    nodes = [Node("root", features=[0.0] * (length + 2) + [1.0])]
    edges = []
    states = {}
    terminals = set()

    def state(string, depth):
        key = (string, depth)
        if key not in states:
            states[key] = f"s{string}_{depth}"
            nodes.append(Node(states[key], features=feat(string, depth)))
        return states[key]

    for parent in ["0" * length, "1" * length]:
        edges.append(
            Edge(
                f"e{len(edges)}",
                "root",
                state(parent, 0),
                0.0,
                {"kind": "start", "string": parent},
            )
        )
    for depth in range(budget + 1):
        for (string, k), sid in list(states.items()):
            if k != depth:
                continue
            if string not in terminals:
                nodes.append(Node("t" + string, string, feat(string, terminal=True)))
                terminals.add(string)
            edges.append(
                Edge(f"e{len(edges)}", sid, "t" + string, 0.0, {"kind": "stop"})
            )
            if depth == budget:
                continue
            for position in range(length):
                changed = (
                    string[:position]
                    + str(1 - int(string[position]))
                    + string[position + 1 :]
                )
                destination = state(changed, depth + 1)
                edges.append(
                    Edge(
                        f"e{len(edges)}",
                        sid,
                        destination,
                        1.0 + 0.1 * position,
                        {
                            "kind": "substitute",
                            "position": position,
                            "product": changed,
                        },
                    )
                )
    return Graph(
        nodes, edges, metadata={"kind": "strings", "length": length, "budget": budget}
    )