Files
ccs/tests/unit/cliproxy/thinking-validator.test.ts
T
Sergey GaluzaandOnSteroids 45fe7ab086 feat(cliproxy): add 'max' thinking level for Claude Opus 4.7
Anthropic exposes `max` as a distinct adaptive-thinking effort above
`xhigh` on Opus 4.7. Previously the validator aliased user input `max`
down to `xhigh`, so users couldn't reach the real top-tier effort
through CCS.

Extend the validator:
- Add `max` to VALID_THINKING_LEVELS and THINKING_LEVEL_RANK (rank 6,
  above xhigh)
- Add `max` to THINKING_LEVEL_BUDGETS (65536, CCS-internal numeric
  mapping for closest-level lookups)
- Widen ThinkingSupport.maxLevel union to include 'max'

The existing `max: 'xhigh'` alias in findClosestLevel is kept as a
graceful fallback for models whose levels list does not include `max`
(e.g. Codex `gpt-5.4`), because exact-match on validLevels takes
priority — so Opus 4.7 returns `max` directly while Codex still maps
`max` -> `xhigh`.

Update the claude.claude-opus-4-7 catalog entry to expose
`['low', 'medium', 'high', 'xhigh', 'max']` with `maxLevel: 'max'`.

Built [OnSteroids](https://onsteroids.ai)

Co-Authored-By: OnSteroids <built@onsteroids.ai>
2026-04-22 21:32:14 +02:00

198 lines
7.4 KiB
TypeScript

/**
* Thinking Validator Unit Tests
*
* Tests for thinking budget validation logic
*/
import { describe, it, expect } from 'bun:test';
import {
validateThinking,
THINKING_LEVEL_BUDGETS,
VALID_THINKING_LEVELS,
THINKING_OFF_VALUES,
THINKING_BUDGET_MIN,
THINKING_BUDGET_MAX,
THINKING_BUDGET_DEFAULT_MIN,
} from '../../../src/cliproxy/thinking-validator';
describe('Thinking Validator', () => {
describe('Constants', () => {
it('should export valid budget bounds', () => {
expect(THINKING_BUDGET_MIN).toBe(0);
expect(THINKING_BUDGET_MAX).toBe(100000);
expect(THINKING_BUDGET_DEFAULT_MIN).toBe(512);
});
it('should export valid thinking levels', () => {
expect(VALID_THINKING_LEVELS).toContain('minimal');
expect(VALID_THINKING_LEVELS).toContain('low');
expect(VALID_THINKING_LEVELS).toContain('medium');
expect(VALID_THINKING_LEVELS).toContain('high');
expect(VALID_THINKING_LEVELS).toContain('xhigh');
expect(VALID_THINKING_LEVELS).toContain('max');
expect(VALID_THINKING_LEVELS).toContain('auto');
});
it('should export level budget mappings', () => {
expect(THINKING_LEVEL_BUDGETS.minimal).toBe(512);
expect(THINKING_LEVEL_BUDGETS.low).toBe(1024);
expect(THINKING_LEVEL_BUDGETS.medium).toBe(8192);
expect(THINKING_LEVEL_BUDGETS.high).toBe(24576);
expect(THINKING_LEVEL_BUDGETS.xhigh).toBe(32768);
// `max` sits above xhigh — unconstrained thinking on Opus 4.7 / Mythos
expect(THINKING_LEVEL_BUDGETS.max).toBeGreaterThan(THINKING_LEVEL_BUDGETS.xhigh);
});
it('should treat max as a distinct top tier on models that list it (Opus 4.7)', () => {
// Opus 4.7 exposes both xhigh and max; max must not collapse into xhigh.
const result = validateThinking('claude', 'claude-opus-4-7', 'max');
expect(result.valid).toBe(true);
expect(result.value).toBe('max');
expect(result.warning).toBeUndefined();
});
it('should still alias max -> xhigh for models without a max level (backcompat)', () => {
// Codex catalog uses ['low','medium','high','xhigh'] with maxLevel 'xhigh'.
// User input "max" should map down to xhigh rather than be rejected.
const result = validateThinking('codex', 'gpt-5.4', 'max');
expect(result.valid).toBe(true);
expect(result.value).toBe('xhigh');
});
it('should export off values', () => {
expect(THINKING_OFF_VALUES).toContain('off');
expect(THINKING_OFF_VALUES).toContain('none');
expect(THINKING_OFF_VALUES).toContain('disabled');
expect(THINKING_OFF_VALUES).toContain('0');
});
});
describe('Off values', () => {
it('should handle "off" value', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', 'off');
expect(result.valid).toBe(true);
expect(result.value).toBe('off');
expect(result.warning).toBeUndefined();
});
it('should handle "none" value', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', 'none');
expect(result.valid).toBe(true);
expect(result.value).toBe('off');
});
it('should handle "disabled" value', () => {
const result = validateThinking('agy', 'claude-sonnet-4-20250514', 'disabled');
expect(result.valid).toBe(true);
expect(result.value).toBe('off');
});
it('should handle "0" string as off', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', '0');
expect(result.valid).toBe(true);
expect(result.value).toBe('off');
});
it('should be case-insensitive for off values', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', 'OFF');
expect(result.valid).toBe(true);
expect(result.value).toBe('off');
});
});
describe('Named levels (levels-type models like Gemini)', () => {
it('should accept valid level names', () => {
const levels = ['low', 'medium', 'high'];
for (const level of levels) {
const result = validateThinking('gemini', 'gemini-3-pro-preview', level);
expect(result.valid).toBe(true);
}
});
it('should be case-insensitive for level names', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', 'HIGH');
expect(result.valid).toBe(true);
});
it('should map numeric budgets to closest level for level-type models', () => {
// Level-type models convert budget to closest named level
const result = validateThinking('gemini', 'gemini-3-pro-preview', 8192);
expect(result.valid).toBe(true);
expect(typeof result.value).toBe('string'); // Should be a level name
expect(result.warning).toContain('Mapped');
});
it('caps unsupported higher standard levels to the highest supported model level', () => {
const result = validateThinking('codex', 'gpt-5-mini', 'xhigh');
expect(result.valid).toBe(true);
expect(result.value).toBe('high');
expect(result.warning).toContain('mapped to "high"');
});
});
describe('Budget-type models (like Claude via agy)', () => {
// Claude models via agy use budget-type thinking
const budgetModel = 'claude-sonnet-4-5-thinking';
it('should accept valid numeric budget', () => {
const result = validateThinking('agy', budgetModel, 8192);
expect(result.valid).toBe(true);
expect(result.value).toBe(8192);
});
it('should accept numeric string budget', () => {
const result = validateThinking('agy', budgetModel, '8192');
expect(result.valid).toBe(true);
expect(result.value).toBe(8192);
});
it('should reject negative budgets', () => {
const result = validateThinking('agy', budgetModel, -100);
expect(result.valid).toBe(false);
expect(result.warning).toContain('Negative');
});
it('should clamp excessively high budgets', () => {
const result = validateThinking('agy', budgetModel, 999999999);
expect(result.valid).toBe(true);
expect(result.warning).toContain('Clamped');
});
it('should reject partial numeric parses like "123abc"', () => {
const result = validateThinking('agy', budgetModel, '123abc');
// Should either fail or find closest level match, not parse as 123
expect(result.value).not.toBe(123);
});
});
describe('Auto value', () => {
it('should accept auto value', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', 'auto');
expect(result.valid).toBe(true);
expect(result.value).toBe('auto');
});
});
describe('Unknown models', () => {
it('should pass through value for unknown models with warning', () => {
const result = validateThinking('gemini', 'unknown-model-xyz', 'high');
expect(result.valid).toBe(true);
expect(result.warning).toContain('unknown');
});
});
describe('Edge cases', () => {
it('should handle whitespace in input', () => {
const result = validateThinking('gemini', 'gemini-3-pro-preview', ' high ');
expect(result.valid).toBe(true);
});
it('should handle empty string by passing through for unknown', () => {
// Empty string on known model goes to level/budget validation
const result = validateThinking('gemini', 'gemini-3-pro-preview', '');
// For level-type models, empty string will try to match levels
expect(result.valid).toBeDefined(); // Just check it doesn't crash
});
});
});