root f9e72412c1 validatord: /v1/validate + /v1/iterate HTTP surface (port 3221)
Closes the last "Go primary" backlog item in
docs/ARCHITECTURE_COMPARISON.md. Go now owns the entire validator
path end-to-end — no Rust dep for staffing safety net.

Architecture: cmd/validatord on :3221 hosts both endpoints. Calls
chatd directly for the iterate loop's LLM hop (no gateway
self-loopback like the Rust shape). Gateway proxies /v1/validate +
/v1/iterate to validatord.

What's in:
- internal/validator/playbook.go — 3rd validator kind (PRD checks:
  fill: prefix, endorsed_names ≤ target_count×2, fingerprint required)
- internal/validator/lookup_jsonl.go — JSONL roster loader (Parquet
  deferred; producer one-liner documented in package comment)
- internal/validator/iterate.go — ExtractJSON helper + Iterate
  orchestrator with ChatCaller seam for unit tests
- cmd/validatord/main.go — HTTP routes, roster load, chat client
- internal/shared/config.go — ValidatordConfig + gateway URL field
- lakehouse.toml — [validatord] section
- cmd/gateway/main.go — proxy routes for /v1/validate + /v1/iterate

Smoke: 5/5 PASS through gateway :3110:
  ✓ playbook happy path
  ✓ playbook missing fingerprint → 422 schema/fingerprint
  ✓ phantom candidate W-PHANTOM → 422 consistency
  ✓ unknown kind → 400
  ✓ roster loaded with 3 records

go test ./... green across 33 packages.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-02 03:53:20 -05:00

190 lines
5.8 KiB
Go

package validator
import (
"context"
"errors"
"testing"
)
func TestExtractJSON_FromFencedBlock(t *testing.T) {
raw := "Here's my answer:\n```json\n{\"fills\": [{\"candidate_id\": \"W-1\"}]}\n```\nDone."
v := ExtractJSON(raw)
if v == nil {
t.Fatal("expected match in fenced block")
}
if _, ok := v["fills"]; !ok {
t.Errorf("missing fills key: %+v", v)
}
}
func TestExtractJSON_FromBareBraces(t *testing.T) {
raw := "Here you go: {\"fills\": [{\"candidate_id\": \"W-2\"}]}"
v := ExtractJSON(raw)
if v == nil {
t.Fatal("expected match in bare braces")
}
}
func TestExtractJSON_ReturnsNilOnNoObject(t *testing.T) {
if v := ExtractJSON("just prose, no json"); v != nil {
t.Errorf("expected nil, got %+v", v)
}
}
func TestExtractJSON_PicksFirstBalancedObject(t *testing.T) {
v := ExtractJSON(`{"a":1} then {"b":2}`)
if v == nil {
t.Fatal("expected match")
}
if v["a"].(float64) != 1 {
t.Errorf("expected first object, got %+v", v)
}
}
func TestExtractJSON_NestedBalancedObjects(t *testing.T) {
v := ExtractJSON(`prefix {"outer": {"inner": [1,2,3]}, "x": "y"} suffix`)
if v == nil {
t.Fatal("expected match on balanced nested object")
}
if outer, ok := v["outer"].(map[string]any); !ok || outer["inner"] == nil {
t.Errorf("nested structure lost: %+v", v)
}
}
func TestExtractJSON_TopLevelArrayReturnsFirstInnerObject(t *testing.T) {
// Both Rust and Go runtimes accept the first balanced {...} as a
// successful match — for `[{"a":1},{"b":2}]` that's the first
// inner object. Documenting this so the contract stays consistent
// across runtimes.
v := ExtractJSON(`[{"a":1},{"b":2}]`)
if v == nil {
t.Fatal("expected first inner object to be returned")
}
if v["a"].(float64) != 1 {
t.Errorf("expected first object {a:1}, got %+v", v)
}
}
// ─── Iterate orchestrator tests with scripted ChatCaller ────────────
func scriptedChat(responses ...string) (ChatCaller, *int) {
idx := 0
return func(_ context.Context, _, _ string, _, _ string, _ *float64, _ int) (string, error) {
if idx >= len(responses) {
return "", errors.New("scripted chat exhausted")
}
r := responses[idx]
idx++
return r, nil
}, &idx
}
func TestIterate_AcceptsFirstValidArtifact(t *testing.T) {
chat, calls := scriptedChat(`{"endorsed_names":["W-1"]}`)
validate := func(_ string, _ map[string]any) (Report, error) {
return Report{ElapsedMs: 1}, nil
}
resp, fail, err := Iterate(context.Background(),
IterateRequest{Kind: "playbook", Prompt: "produce X", Provider: "ollama", Model: "qwen3.5:latest"},
IterateConfig{}, chat, validate)
if err != nil || fail != nil {
t.Fatalf("expected success, got err=%v fail=%+v", err, fail)
}
if resp.Iterations != 1 {
t.Errorf("iterations = %d, want 1", resp.Iterations)
}
if len(resp.History) != 1 || resp.History[0].Status.Kind != "accepted" {
t.Errorf("history: %+v", resp.History)
}
if *calls != 1 {
t.Errorf("expected 1 chat call, got %d", *calls)
}
}
func TestIterate_RetriesOnNoJsonThenSucceeds(t *testing.T) {
chat, _ := scriptedChat(
"sorry I cannot do that",
`{"endorsed_names":["W-1"]}`,
)
validate := func(_ string, _ map[string]any) (Report, error) {
return Report{}, nil
}
resp, _, err := Iterate(context.Background(),
IterateRequest{Kind: "playbook", Prompt: "produce X", Provider: "ollama", Model: "x"},
IterateConfig{}, chat, validate)
if err != nil || resp == nil {
t.Fatalf("expected success, err=%v", err)
}
if resp.Iterations != 2 {
t.Errorf("iterations = %d, want 2", resp.Iterations)
}
if resp.History[0].Status.Kind != "no_json" {
t.Errorf("first history status: %+v", resp.History[0].Status)
}
}
func TestIterate_RetriesOnValidationFailureThenSucceeds(t *testing.T) {
chat, _ := scriptedChat(
`{"bad":"shape"}`,
`{"good":"shape"}`,
)
calls := 0
validate := func(_ string, body map[string]any) (Report, error) {
calls++
if _, ok := body["good"]; ok {
return Report{}, nil
}
return Report{}, &ValidationError{Kind: ErrSchema, Field: "x", Reason: "missing good"}
}
resp, _, err := Iterate(context.Background(),
IterateRequest{Kind: "playbook", Prompt: "produce X", Provider: "ollama", Model: "x"},
IterateConfig{}, chat, validate)
if err != nil || resp == nil {
t.Fatalf("expected success, err=%v", err)
}
if calls != 2 {
t.Errorf("validate calls = %d, want 2", calls)
}
if resp.History[0].Status.Kind != "validation_failed" {
t.Errorf("first history status: %+v", resp.History[0].Status)
}
if resp.History[0].Status.Error == "" {
t.Errorf("validation_failed entry must carry error string")
}
}
func TestIterate_MaxIterationsExhaustedReturnsFailure(t *testing.T) {
chat, _ := scriptedChat(`{}`, `{}`, `{}`)
validate := func(_ string, _ map[string]any) (Report, error) {
return Report{}, &ValidationError{Kind: ErrCompleteness, Reason: "always wrong"}
}
resp, fail, err := Iterate(context.Background(),
IterateRequest{Kind: "playbook", Prompt: "X", Provider: "ollama", Model: "x", MaxIterations: 3},
IterateConfig{}, chat, validate)
if err != nil {
t.Fatalf("infrastructure error unexpected: %v", err)
}
if resp != nil {
t.Fatalf("expected failure, got %+v", resp)
}
if fail.Iterations != 3 {
t.Errorf("iterations = %d, want 3", fail.Iterations)
}
if len(fail.History) != 3 {
t.Errorf("history length = %d, want 3", len(fail.History))
}
}
func TestIterate_PropagatesChatInfraError(t *testing.T) {
chat := func(_ context.Context, _, _ string, _, _ string, _ *float64, _ int) (string, error) {
return "", errors.New("connection refused")
}
validate := func(_ string, _ map[string]any) (Report, error) { return Report{}, nil }
_, _, err := Iterate(context.Background(),
IterateRequest{Kind: "playbook", Prompt: "X", Provider: "ollama", Model: "x"},
IterateConfig{}, chat, validate)
if err == nil {
t.Fatal("expected infrastructure error to surface")
}
}