"""Validation Demo - Shows how the solver helps LLMs self-correct.
This example demonstrates the Phase 2 validation features:
- Pre-solve model validation
- Actionable error messages
- Smart typo detection with suggestions
- Three-level severity (ERROR, WARNING, INFO)
Run: python examples/validation_demo.py
"""
import asyncio
from chuk_mcp_solver.models import (
Constraint,
ConstraintKind,
LinearConstraintParams,
LinearTerm,
Objective,
ObjectiveSense,
SolveConstraintModelRequest,
SolverMode,
Variable,
VariableDomain,
VariableDomainType,
)
from chuk_mcp_solver.solver import get_solver
async def example_1_undefined_variable():
"""Example 1: Undefined variable with typo detection."""
print("=" * 80)
print("EXAMPLE 1: Undefined Variable with Typo Detection")
print("=" * 80)
print()
solver = get_solver("ortools")
# Model with typo: references 'y' but variable is named 'x'
request = SolveConstraintModelRequest(
mode=SolverMode.OPTIMIZE,
variables=[
Variable(
id="x",
domain=VariableDomain(type=VariableDomainType.INTEGER, lower=0, upper=10),
)
],
constraints=[
Constraint(
id="c1",
kind=ConstraintKind.LINEAR,
params=LinearConstraintParams(
terms=[LinearTerm(var="y", coef=1)], # Typo: 'y' instead of 'x'
sense="<=",
rhs=5,
),
)
],
objective=Objective(sense=ObjectiveSense.MAXIMIZE, terms=[LinearTerm(var="x", coef=1)]),
)
response = await solver.solve_constraint_model(request)
print(f"Status: {response.status}")
print()
print("Validation Error Message:")
print("-" * 80)
if response.explanation:
print(response.explanation.summary)
print()
print("✅ Notice how the error message suggests 'Did you mean x?'")
print()
async def example_2_invalid_bounds():
"""Example 2: Invalid domain bounds."""
print("=" * 80)
print("EXAMPLE 2: Invalid Domain Bounds")
print("=" * 80)
print()
solver = get_solver("ortools")
# Model with invalid bounds: lower > upper
request = SolveConstraintModelRequest(
mode=SolverMode.SATISFY,
variables=[
Variable(
id="x",
domain=VariableDomain(
type=VariableDomainType.INTEGER,
lower=10,
upper=5, # Invalid!
),
)
],
constraints=[
Constraint(
id="c1",
kind=ConstraintKind.LINEAR,
params=LinearConstraintParams(
terms=[LinearTerm(var="x", coef=1)], sense="<=", rhs=20
),
)
],
)
response = await solver.solve_constraint_model(request)
print(f"Status: {response.status}")
print()
print("Validation Error Message:")
print("-" * 80)
if response.explanation:
print(response.explanation.summary)
print()
print("✅ Clear error message explaining the problem")
print()
async def example_3_duplicate_ids():
"""Example 3: Duplicate variable IDs."""
print("=" * 80)
print("EXAMPLE 3: Duplicate Variable IDs")
print("=" * 80)
print()
solver = get_solver("ortools")
# Model with duplicate variable IDs
request = SolveConstraintModelRequest(
mode=SolverMode.SATISFY,
variables=[
Variable(
id="x",
domain=VariableDomain(type=VariableDomainType.INTEGER, lower=0, upper=10),
),
Variable(
id="x", # Duplicate ID!
domain=VariableDomain(type=VariableDomainType.INTEGER, lower=0, upper=20),
),
],
constraints=[
Constraint(
id="c1",
kind=ConstraintKind.LINEAR,
params=LinearConstraintParams(
terms=[LinearTerm(var="x", coef=1)], sense="<=", rhs=5
),
)
],
)
response = await solver.solve_constraint_model(request)
print(f"Status: {response.status}")
print()
print("Validation Error Message:")
print("-" * 80)
if response.explanation:
print(response.explanation.summary)
print()
print("✅ Detects duplicate IDs across variables")
print()
async def example_4_valid_model():
"""Example 4: Valid model passes validation."""
print("=" * 80)
print("EXAMPLE 4: Valid Model (No Validation Errors)")
print("=" * 80)
print()
solver = get_solver("ortools")
# Valid model
request = SolveConstraintModelRequest(
mode=SolverMode.OPTIMIZE,
variables=[
Variable(
id="x",
domain=VariableDomain(type=VariableDomainType.INTEGER, lower=0, upper=10),
),
Variable(
id="y",
domain=VariableDomain(type=VariableDomainType.INTEGER, lower=0, upper=10),
),
],
constraints=[
Constraint(
id="budget",
kind=ConstraintKind.LINEAR,
params=LinearConstraintParams(
terms=[LinearTerm(var="x", coef=2), LinearTerm(var="y", coef=3)],
sense="<=",
rhs=20,
),
)
],
objective=Objective(
sense=ObjectiveSense.MAXIMIZE,
terms=[LinearTerm(var="x", coef=5), LinearTerm(var="y", coef=4)],
),
)
response = await solver.solve_constraint_model(request)
print(f"Status: {response.status}")
print(f"Objective Value: {response.objective_value}")
if response.solutions:
print("\nSolution:")
for var in response.solutions[0].variables:
print(f" {var.id} = {var.value}")
print()
print("✅ Valid model passes validation and solves successfully")
print()
async def main():
"""Run all validation examples."""
print()
print("╔" + "=" * 78 + "╗")
print("║" + " " * 20 + "VALIDATION DEMO - PHASE 2 FEATURES" + " " * 24 + "║")
print("╚" + "=" * 78 + "╝")
print()
print("This demo shows how the solver validates models and provides")
print("actionable error messages to help LLMs self-correct.")
print()
await example_1_undefined_variable()
await example_2_invalid_bounds()
await example_3_duplicate_ids()
await example_4_valid_model()
print("=" * 80)
print("SUMMARY")
print("=" * 80)
print()
print("The validation framework helps LLMs by:")
print(" • Detecting errors before expensive solve operations")
print(" • Providing specific locations of errors")
print(" • Suggesting corrections (e.g., 'Did you mean x?')")
print(" • Using three severity levels (ERROR, WARNING, INFO)")
print()
print("This enables LLMs to self-correct and iterate toward valid models.")
print()
if __name__ == "__main__":
asyncio.run(main())
