from __future__ import annotations
from dataclasses import dataclass
from typing import Hashable, List
import logging
import networkx as nx
logger = logging.getLogger(__name__)
[docs]
@dataclass(frozen=True)
class CRNGraphInfo:
species_nodes: List[Hashable]
rule_nodes: List[Hashable]
is_dag: bool
[docs]
def node_sort_key(graph: nx.DiGraph, node: Hashable):
data = graph.nodes[node]
kind = data.get("kind")
if kind == "rule":
return (
1,
data.get("step", 10**9),
data.get("rule_index", 10**9),
data.get("app_index", 10**9),
str(data.get("label", "")),
str(node),
)
return (
0,
data.get("step", -1),
str(data.get("label", "")),
str(data.get("smiles", "")),
str(node),
)
[docs]
def validate_crn_graph(graph: nx.DiGraph, *, strict: bool = True) -> CRNGraphInfo:
if not isinstance(graph, nx.DiGraph):
raise TypeError("CRNVis expects a networkx.DiGraph.")
valid_kinds = {"species", "rule"}
invalid_kind_nodes = [
n for n, d in graph.nodes(data=True) if d.get("kind") not in valid_kinds
]
if invalid_kind_nodes:
msg = (
"Found nodes with invalid or missing 'kind': "
f"{invalid_kind_nodes!r}. Expected only 'species' or 'rule'."
)
if strict:
raise ValueError(msg)
logger.warning(msg)
species_nodes = sorted(
[n for n, d in graph.nodes(data=True) if d.get("kind") == "species"],
key=lambda n: node_sort_key(graph, n),
)
rule_nodes = sorted(
[n for n, d in graph.nodes(data=True) if d.get("kind") == "rule"],
key=lambda n: node_sort_key(graph, n),
)
invalid_edges = []
unknown_roles = []
for u, v, d in graph.edges(data=True):
role = d.get("role")
u_kind = graph.nodes[u].get("kind")
v_kind = graph.nodes[v].get("kind")
if role == "reactant":
if not (u_kind == "species" and v_kind == "rule"):
invalid_edges.append((u, v, role))
elif role == "product":
if not (u_kind == "rule" and v_kind == "species"):
invalid_edges.append((u, v, role))
elif role is not None:
unknown_roles.append((u, v, role))
if invalid_edges:
msg = (
"Found edges whose direction is inconsistent with their role: "
f"{invalid_edges!r}."
)
if strict:
raise ValueError(msg)
logger.warning(msg)
if unknown_roles:
logger.warning(
"Found edges with unknown roles; they will be drawn with fallback style: %r",
unknown_roles,
)
return CRNGraphInfo(
species_nodes=species_nodes,
rule_nodes=rule_nodes,
is_dag=nx.is_directed_acyclic_graph(graph),
)
