Files
lda-wf/src/wf_core/validation/context_paths.py
T

315 lines
12 KiB
Python

from __future__ import annotations
from collections.abc import Iterator, Mapping
from typing import Any
from wf_core.analysis.context_scopes import (
ContextSchema,
root_context_schema,
)
from wf_core.analysis.control_regions import ControlRegionAnalysis
from wf_core.context_contracts import RESERVED_CONTEXT_KEYS
from wf_core.models.conditions import (
BinaryCondition,
Condition,
ExistsCondition,
LiteralOperand,
NotCondition,
PathOperand,
VariadicCondition,
)
from wf_core.models.input_bindings import (
ArrayExpression,
InputExpression,
InputExpressionBinding,
InputPathBinding,
LiteralExpression,
ObjectExpression,
PathExpression,
)
from wf_core.models.steps import (
ConditionNode,
ForeachNode,
InterruptNode,
NodeUse,
SubgraphNode,
)
from wf_core.models.workflow import Workflow
from wf_core.paths import GraphSourcePath
from wf_core.schema_navigation import SchemaNavigator
from wf_core.validation.issues import ValidationIssueCode, ValidationReport
def validate_context_paths(
workflow: Workflow,
*,
context_schemas: Mapping[str, ContextSchema],
report: ValidationReport,
control_regions: ControlRegionAnalysis | None = None,
) -> None:
"""Validate every ``context.*`` path against its consuming location schema.
Ordinary input/state validation stays where it is; this pass owns the
stronger program-location-aware meaning of ``context.*``. A path is valid
only if every literal segment is a declared object property in the
consuming node's generated schema. The whole ``context`` object and the
``context.foreach`` map remain readable; unknown dynamic keys do not.
The shared control-region analysis is threaded through so validation runs
it once; alias ownership never triggers a second traversal.
"""
nodes_by_index = list(workflow.nodes)
node_index_by_id = {node.id: idx for idx, node in enumerate(nodes_by_index)}
_validate_alias_ownership(
workflow, node_index_by_id, report, control_regions=control_regions
)
for idx, node in enumerate(nodes_by_index):
schema = context_schemas.get(node.id)
if isinstance(node, NodeUse):
_validate_step_input_bindings(
node.input, f"nodes[{idx}].input", node.id, schema, report
)
elif isinstance(node, SubgraphNode):
_validate_step_input_bindings(
node.input, f"nodes[{idx}].input", node.id, schema, report
)
elif isinstance(node, ConditionNode):
for location, path in _condition_paths(node.check, f"nodes[{idx}].check"):
_validate_one_context_path(path, location, node.id, schema, report)
elif isinstance(node, ForeachNode):
# Context-rooted `over` paths reach this pass; the old
# input/state-only check stays permissive for them.
if node.over.root == "context":
_validate_context_foreach_source(
node.over, f"nodes[{idx}].over", node.id, schema, report
)
elif isinstance(node, InterruptNode):
_validate_step_input_bindings(
node.request, f"nodes[{idx}].request", node.id, schema, report
)
_validate_workflow_output(workflow, report)
def _validate_step_input_bindings(
bindings: list[Any],
base: str,
node_id: str,
schema: ContextSchema | None,
report: ValidationReport,
) -> None:
"""Validate context paths in one input/request binding list.
`base` is the list location such as `nodes[3].input` or
`nodes[1].request`; each binding contributes `base[i]` and each path
field contributes a further suffix like `.path` or
`.expression.items[0].path`.
"""
for binding_index, binding in enumerate(bindings):
binding_location = f"{base}[{binding_index}]"
if isinstance(binding, InputPathBinding):
if binding.path.root == "context":
_validate_one_context_path(
binding.path, f"{binding_location}.path", node_id, schema, report
)
elif isinstance(binding, InputExpressionBinding):
for location, path in _expression_paths(
binding.expression, f"{binding_location}.expression"
):
if path.root == "context":
_validate_one_context_path(path, location, node_id, schema, report)
def _expression_paths(
expression: InputExpression,
location: str,
) -> Iterator[tuple[str, GraphSourcePath]]:
"""Yield ``(model path, graph path)`` for every path leaf in an expression.
Finite recursion mirrors the input-expression model: paths, arrays, and
objects. Literals contribute no paths.
"""
match expression:
case PathExpression(path=path):
yield location + ".path", path
case ArrayExpression(items=items):
for index, item in enumerate(items):
yield from _expression_paths(item, f"{location}.items[{index}]")
case ObjectExpression(fields=fields):
for name, item in fields.items():
yield from _expression_paths(item, f"{location}.fields.{name}")
case LiteralExpression():
return
def _condition_paths(
condition: Condition,
location: str,
) -> Iterator[tuple[str, GraphSourcePath]]:
"""Yield context-candidate paths from a condition tree with model locations."""
if isinstance(condition, ExistsCondition):
yield location + ".path", condition.path
return
if isinstance(condition, NotCondition):
yield from _condition_paths(condition.arg, f"{location}.arg")
return
if isinstance(condition, VariadicCondition):
for index, arg in enumerate(condition.args):
yield from _condition_paths(arg, f"{location}.args[{index}]")
return
if isinstance(condition, BinaryCondition):
yield from _operand_paths(condition.left, f"{location}.left")
yield from _operand_paths(condition.right, f"{location}.right")
return
def _operand_paths(
operand: PathOperand | LiteralOperand, location: str
) -> Iterator[tuple[str, GraphSourcePath]]:
if isinstance(operand, LiteralOperand):
return
yield location + ".path", operand.path
def _validate_one_context_path(
path: GraphSourcePath,
location: str,
node_id: str | None,
schema: ContextSchema | None,
report: ValidationReport,
) -> None:
if path.root != "context":
return
if schema is None:
report.add(
ValidationIssueCode.INVALID_CONTEXT_PATH,
location,
f"invalid context path {str(path)!r} at {node_id or location!r}: "
"no context schema for this program location",
)
return
failing, available = _failing_segment(schema, path.parts)
if failing is not None:
listed = f" (available: {available})" if available else ""
report.add(
ValidationIssueCode.INVALID_CONTEXT_PATH,
location,
f"invalid context path {str(path)!r} at {node_id or location!r}: "
f"unknown segment {failing!r}{listed}",
)
def _validate_context_foreach_source(
path: GraphSourcePath,
location: str,
node_id: str,
schema: ContextSchema | None,
report: ValidationReport,
) -> None:
"""Require an existing context path to declare an array source.
Missing paths keep the more precise ``INVALID_CONTEXT_PATH`` diagnostic;
this adds the foreach-specific error only after the path resolves.
"""
_validate_one_context_path(path, location, node_id, schema, report)
if schema is None:
return
source = SchemaNavigator(schema).at_path(path.parts)
if source is not None and not source.is_array():
report.add(
ValidationIssueCode.INVALID_FOREACH_SOURCE,
location,
f"foreach source {str(path)!r} must resolve to an array",
)
def _failing_segment(
schema: Mapping[str, Any], parts: tuple[str, ...]
) -> tuple[str | None, str]:
"""Return the first unknown segment plus the keys available there.
Returns ``(None, "")`` when the path walks declared properties (or
permissive unconstrained schemas). Reference resolution is delegated to
the shared Draft 2020-12 navigator; unresolvable refs fail closed.
Composition is a union for path availability: a path is readable when
some structural branch declares it.
"""
return SchemaNavigator(schema).first_unknown(parts)
def _validate_workflow_output(workflow: Workflow, report: ValidationReport) -> None:
schema = root_context_schema()
for output_index, binding in enumerate(workflow.output):
if isinstance(binding, InputPathBinding):
if binding.path.root == "context":
_validate_one_context_path(
binding.path,
f"output[{output_index}].path",
"workflow_output",
schema,
report,
)
elif isinstance(binding, InputExpressionBinding):
for location, path in _expression_paths(
binding.expression, f"output[{output_index}].expression"
):
if path.root == "context":
_validate_one_context_path(
path, location, "workflow_output", schema, report
)
def _validate_alias_ownership(
workflow: Workflow,
node_index_by_id: dict[str, int],
report: ValidationReport,
*,
control_regions: ControlRegionAnalysis | None = None,
) -> None:
"""Reject reserved or colliding active foreach aliases.
Reserved names are every standard context field plus ``foreach``,
``loop_item``, and ``loop_index`` (that is, ``RESERVED_CONTEXT_KEYS``).
Siblings in separate control regions may reuse an alias because they are
never active together; only aliases active in the same owner stack
collide. Failures point at the inner foreach's ``as`` field. The shared
control-region analysis is reused; this helper never traverses alone.
"""
if control_regions is None:
from wf_core.analysis.control_regions import analyze_control_regions
control_regions = analyze_control_regions(workflow)
analysis = control_regions
foreach_by_id = {
node.id: node for node in workflow.nodes if isinstance(node, ForeachNode)
}
reported: set[str] = set()
for stack in analysis.owner_stack_by_node.values():
seen_aliases: dict[str, str] = {}
for owner_id in stack:
foreach = foreach_by_id.get(owner_id)
if foreach is None:
continue
alias = foreach.as_
idx = node_index_by_id.get(owner_id)
location = f"nodes[{idx}].as" if idx is not None else f"nodes[{owner_id}]"
if not alias or alias in RESERVED_CONTEXT_KEYS:
if owner_id not in reported:
reported.add(owner_id)
report.add(
ValidationIssueCode.FOREACH_CONTEXT_ALIAS_CONFLICT,
location,
f"foreach alias {alias!r} for node {owner_id!r} "
"collides with reserved context keys",
)
continue
if alias in seen_aliases:
if owner_id not in reported:
reported.add(owner_id)
report.add(
ValidationIssueCode.FOREACH_CONTEXT_ALIAS_CONFLICT,
location,
f"foreach alias {alias!r} for node {owner_id!r} "
f"collides with active alias from {seen_aliases[alias]!r}",
)
continue
seen_aliases[alias] = owner_id