Files
buzz/scripts/run-corpus.mjs
Will Pfleger 4d47f48143 feat(agent): Phase 1 — model-capability manifest, generator, and test oracle (#3821)
## What this does

Introduces the model-capability manifest infrastructure (Phase 1 of the
Model-Capability Manifest plan v4, Thufir-approved 9/9/9). No consumer
cutover — `config.rs`, `llm.rs`, `catalog.rs`, and `buzzAgentConfig.ts`
are unchanged. Phase 2 wires them.

**Single source of truth** replaces hand-mirrored metadata across four
files:

```
scripts/model-capabilities.json         → hand-curated manifest
scripts/generate-model-capabilities.mjs → emits Rust + TS artifacts
crates/buzz-agent/src/generated_model_capabilities.rs
desktop/src/features/agents/ui/modelCapabilities.ts
```

## Resolver contract (plan v4 §Resolver contract)

Total function `resolve(provider, raw_model_id) → CapabilityResult`.
Three ordered steps:

1. Provider-qualified raw exact lookup — key is `(provider,
raw_model_id)`, matched before any prefix stripping. A prefixed alias
never inherits an exact record.
2. Provider-scoped ordered family rules — on normalized
(prefix-stripped) alias, by `match_priority` desc.
3. Per-axis provider fallback — `blank` vs `concrete_unknown`, per
provider.

Result is complete — every axis populated, runtime consumers never
compose fields.

## Boundaries the manifest does NOT own

- Transport for pure OpenAI, legacy Databricks, OpenRouter:
`OpenAiApi`/`openai_request()` remain authoritative.
`databricks_v2_wire_route` is DBv2-only (all other providers emit
`not-applicable`).
- Final display labels: `resolveModelLabel()` three-tier precedence
unchanged. `registry_label` feeds only the static registry tier.
- `llm.rs` scope: only `databricks_v2_route_for_model` (Phase 2).

## Test oracle (three independent layers)

1. Generated full-table coverage —
`scripts/generated-model-capabilities-coverage.json`: every manifest
entry + provider fallbacks.
2. Hand-authored normative corpus — `scripts/normative-corpus.json` (44
vectors): Anthropic manual-budget/adaptive families, OpenAI gpt-5
adversarial boundary cases, DBv2 segment-routing collision tests, P2-A
resolver-contract vectors, P2-B blank/concrete-unknown per provider.
Runs against JS resolver (`run-corpus.mjs`) and mirrored in Rust
(`generated_model_capabilities_tests.rs`).
3. Schema-negative tests — `scripts/test-manifest-validator.mjs` (17
tests): every validator rule has a failing-input test.

## Reconciliation table

`scripts/MODELS_DEV_RECONCILIATION.md` — all models.dev divergences
dispositioned. `databricks-gpt-5-4-mini` and `databricks-gpt-5-4-nano`
adopt models.dev `[low,medium,high]` (family rule adds `none+xhigh` the
endpoint doesn't advertise).

## CI

`.github/workflows/model-capability-regen-diff.yml`: triggers on
manifest/generator/artifact changes; regenerates and fails if stale;
runs JS corpus + schema-negative tests.

## Acceptance criteria (plan v4 Phase 1)

- Byte-clean regen: `node scripts/generate-model-capabilities.mjs
--check` passes
- Rust compiles: `cargo check -p buzz-agent`
- TS typechecks: `pnpm tsc --noEmit --strict`
- 44/44 normative corpus vectors pass (JS interpreter)
- 41/41 Rust corpus tests pass
- 17/17 schema-negative tests pass (every validator rule)
- Reconciliation table complete with doc citations
- No consumer changes (config.rs, llm.rs, catalog.rs, buzzAgentConfig.ts
untouched)

---------

Signed-off-by: Will Pfleger <pfleger.will@gmail.com>
Co-authored-by: npub1mn7jgtj4w2pd0g0zeuhxsa6jy6p0rewxz4kujt98my82ahfmp72sxjexk7 <dcfd242e557282d7a1e2cf2e6877522682f1e5c6156dc92ca7d90eaedd3b0f95@buzz.block.builderlab.xyz>
2026-07-31 12:39:37 -04:00

84 lines
2.9 KiB
JavaScript

#!/usr/bin/env node
/**
* Normative corpus runner — validates the generated TS interpreter against
* scripts/normative-corpus.json.
*
* This is the JS side of the two-interpreter corpus check. The runner imports
* resolveModelCapabilities() directly from the generated TypeScript module via
* Node's --experimental-strip-types flag. The Rust side lives in
* crates/buzz-agent/src/generated_model_capabilities.rs (shared corpus harness).
*
* Usage: node --experimental-strip-types scripts/run-corpus.mjs [--verbose]
* Exits 0 on all pass, 1 on any failure.
*/
import { readFileSync } from "node:fs";
import { join, dirname } from "node:path";
import { fileURLToPath } from "node:url";
const __dirname = dirname(fileURLToPath(import.meta.url));
const repoRoot = join(__dirname, "..");
const VERBOSE = process.argv.includes("--verbose");
// Import the generated TypeScript interpreter directly.
// Node 22+ --experimental-strip-types strips type annotations at load time; no build step needed.
const { resolveModelCapabilities } = await import(
join(repoRoot, "desktop", "src", "features", "agents", "ui", "modelCapabilities.ts")
);
const corpus = JSON.parse(
readFileSync(join(repoRoot, "scripts", "normative-corpus.json"), "utf8"),
);
// ----- Run corpus -----
let passed = 0;
let failed = 0;
for (const entry of corpus) {
// Skip group header entries
if (entry._group) continue;
if (!entry.expect) continue;
// resolveModelCapabilities returns camelCase keys (registryLabel, thinkingMode, etc.)
const result = resolveModelCapabilities(entry.provider, entry.raw_model_id);
const expect = entry.expect;
const failures = [];
for (const [key, expectedVal] of Object.entries(expect)) {
// Corpus uses snake_case; generated TS uses camelCase — convert for lookup.
const camelKey = key.replace(/_([a-z])/g, (_, c) => c.toUpperCase());
const actualVal = camelKey in result ? result[camelKey] : result[key];
if (Array.isArray(expectedVal)) {
// Order-sensitive comparison for effort arrays
const actualArr = Array.isArray(actualVal) ? actualVal : [];
if (JSON.stringify(actualArr) !== JSON.stringify(expectedVal)) {
failures.push(
` ${key}: expected [${expectedVal.join(", ")}] got [${actualArr.join(", ")}]`,
);
}
} else {
if (actualVal !== expectedVal) {
failures.push(` ${key}: expected ${JSON.stringify(expectedVal)} got ${JSON.stringify(actualVal)}`);
}
}
}
if (failures.length === 0) {
passed++;
if (VERBOSE) {
console.log(` PASS ${entry.id}`);
}
} else {
failed++;
console.error(` FAIL ${entry.id} (${entry.provider} / "${entry.raw_model_id}")`);
for (const f of failures) console.error(f);
if (entry._note) console.error(` note: ${entry._note}`);
}
}
console.log(`\nCorpus: ${passed} passed, ${failed} failed`);
process.exit(failed > 0 ? 1 : 0);