Files
gbrain/src/core/embed-preflight.ts
T
a25209bbb2 v0.42.58.0 fix(ai): provider-agnostic gateway — env clobber, base-URL /v1, embedding dims (#1249 #1250 #1292 #2271 #2209) (#2627)
* fix(ai): drop empty-string env values before merge so they can't clobber config keys (#1249)

Claude Code injects ANTHROPIC_API_KEY='' to neuter subprocess LLM calls; an
unconditional process.env spread let that empty string override a valid
config.json key, breaking every gateway op with NO_ANTHROPIC_API_KEY. Filter
'' / undefined before the merge; '0' and 'false' are preserved.

* fix(ai): normalize native provider base URLs + replace embedding guard with a dims-presence check (#1250, #1292)

#1250: createAnthropic/createOpenAI were called with no baseURL, so an
env-injected bare host (e.g. ANTHROPIC_BASE_URL without /v1) 404'd. Add a
shared resolveNativeBaseUrl and pass a normalized baseURL at all anthropic +
openai native sites (google deferred until its suffix is verified).

#1292/D6: the user_provided_model_unset guard was structurally unreachable as
a no-model check (parseModelId throws on a bare provider) and only ever
false-positived for litellm:<model>, silently disabling vector search. Replace
it with a real dims-presence check for user-provided/zero-default recipes and
delete the dead branch in both consumers. Also stop configureGateway from
fabricating a default embedding_dimensions, so 'no dims set' stays honest.

* fix(ai): trust user-declared embedding dims for local recipes + litellm /v1 hint (#2271, #2209)

#2271: a new trust_custom_dims flag adds a passthrough tier so ollama /
llama-server / litellm accept a user-supplied --embedding-dimensions instead of
being hard-rejected. Fail-closed for fixed-dim providers (openai/voyage/
zeroentropy) and excludes openrouter (declares dims_options). Register modern
ollama embed model names.

#2209: litellm setup_hint now states the /v1 path convention and the docs
pointer is corrected to docs/integrations/embedding-providers.md.

* docs+test(ai): KEY_FILES current-state for provider-agnostic gateway + embed-preflight dims-unset test (#1249, #1250, #1292)

* fix(ai): point user_provided_dims_unset remediation at 'gbrain init' (config set rejects it) + coverage

Pre-landing adversarial review (P1): the new dims-unset guard told users to run
'gbrain config set embedding_dimensions <N>', which config.ts hard-rejects (it's a
schema-sizing field). Both consumer messages now point at 'gbrain init
--embedding-dimensions'. Adds: pgvector-cap-still-fires regression for the
trust_custom_dims passthrough, and a configureGateway backfill-invariant test.

* chore: bump version and changelog (v0.42.57.0)

Provider-agnostic plumbing wave: #1249 empty-env clobber, #1250 native baseURL
normalization, #1292 embedding dims-presence guard, #2271 trust_custom_dims
passthrough, #2209 litellm /v1 hint.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* docs: sync embedding-providers guide for provider-agnostic gateway wave (v0.42.57.0)

Post-ship doc drift fix for the v0.42.57.0 AI-gateway wave:
- LiteLLM section now names the /v1 base-URL convention (#2209).
- Ollama section lists the newly-registered modern embedders qwen3-embed-8b
  + snowflake-arctic-embed-l-v2, and notes dims-trust for local recipes (#2271).
- llama-server section notes gbrain trusts the user-declared dimension (#2271).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* docs: post-ship doc sweep for v0.42.57.0 provider-agnostic gateway wave

- KEY_FILES.md types.ts entry: document EmbeddingTouchpoint.trust_custom_dims
  (#2271 passthrough tier, runs after dims_options + Matryoshka allowlists)
- ENGINES.md: embedding design-choice note now names the provider-agnostic
  gateway delegation instead of the stale OpenAI-only parenthetical
- embedding-providers.md: drop an exact-duplicate doctor-8c paragraph
- llms-full.txt regenerated (ENGINES.md is inlined in the bundle)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: apply codex doc-review findings for v0.42.57.0 (base-URL env note, litellm multimodal)

- embedding-providers.md OpenAI section: document OPENAI_BASE_URL /
  ANTHROPIC_BASE_URL bare-host /v1 normalization (#1250 user-facing surface)
- TL;DR table: litellm multimodal is backend-permitting (recipe declares
  supports_multimodal: true, routed via the openai-compat multimodal path),
  not "no"

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: pin engine-find-trajectory schema to 1536 + stop gateway-state leaks across shard files

CI shard 5 failed 7 findTrajectory tests with 'expected 1280 dimensions, not
1536': engine-find-trajectory hardcodes 1536-d vectors but sizes its schema
from AMBIENT gateway state in beforeAll — which runs before the
legacy-embedding-preload's per-test 1536 restore. A preceding file that ends
with a dimensionless configureGateway (facts-extract-silent-no-op) or a bare
resetGateway poisons the next fresh initSchema down to 1280-d columns. The
new test files in this PR reshuffled shard bin-packing and exposed the trap.

- engine-find-trajectory: pin OpenAI/1536 explicitly before initSchema (the
  pattern bunfig's preload documents) — deterministic regardless of neighbors
- facts-extract-silent-no-op, diagnose-embedding-dims, embed-preflight:
  restore the legacy 1536 pin in afterAll instead of ending reset/dimensionless

Reproduced: synthetic dimensionless-gateway file + old victim = the exact 7
CI failures; with the pin = 0. Verified in-process pair runs both orders.

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-10 10:05:23 +09:00

126 lines
4.8 KiB
TypeScript

/**
* Embedding credential preflight.
*
* v0.41.6.0 D1 — fail fast at sync/embed/import entry when the configured
* embedding provider can't be reached. Without this, gbrain proceeds into
* the import phase, hits 565 per-file embed errors, writes 565 identical
* "OpenAI embedding requires OPENAI_API_KEY." rows to `~/.gbrain/sync-failures.jsonl`,
* and blocks the sync bookmark from advancing.
*
* Routes through `gateway.diagnoseEmbedding()` so the structured reason
* (missing_env / no_touchpoint / unknown_provider / etc.) drives a precise
* user-facing error message.
*
* Skip protocol: callers SHOULD NOT call validateEmbeddingCreds when the
* user explicitly passed `--no-embed` (the canonical opt-out). The
* deferred-setup sentinel is owned by `assertEmbeddingEnabled` in
* `embedding-dim-check.ts`; this preflight runs AFTER that check fires.
*/
import { diagnoseEmbedding, type EmbeddingDiagnosis } from './ai/gateway.ts';
/**
* Tagged error thrown by validateEmbeddingCreds. CLI catch sites format
* `.userMessage` to stderr and exit non-zero. The structured fields
* (`provider`, `model`, `missingEnvVars`, `reason`) enable programmatic
* consumers (`gbrain doctor --json`, future autopilot health checks) to
* read state without parsing the human message.
*/
export class EmbeddingCredentialError extends Error {
readonly diagnosis: EmbeddingDiagnosis;
readonly userMessage: string;
constructor(diagnosis: EmbeddingDiagnosis, userMessage: string) {
super(userMessage);
this.name = 'EmbeddingCredentialError';
this.diagnosis = diagnosis;
this.userMessage = userMessage;
}
}
/**
* Run the preflight. Throws EmbeddingCredentialError when the gateway
* can't serve embeddings. Returns silently when ok.
*
* Pure function: reads nothing except what `gateway.diagnoseEmbedding()`
* already had at gateway configure-time.
*/
export function validateEmbeddingCreds(): void {
const d = diagnoseEmbedding();
if (d.ok) return;
throw new EmbeddingCredentialError(d, formatEmbeddingCredsError(d));
}
/**
* Format a paste-ready, multi-line error message from a non-ok diagnosis.
* Exported for tests and for the doctor JSON output.
*/
export function formatEmbeddingCredsError(d: EmbeddingDiagnosis): string {
if (d.ok) return '';
switch (d.reason) {
case 'no_gateway_config':
return [
'Embedding gateway is not configured.',
'This is usually a startup-order bug. Re-run with --no-embed to import',
'without embedding, then file an issue at https://github.com/garrytan/gbrain/issues',
].join('\n');
case 'no_model_configured':
return [
'No embedding model is configured for this brain.',
'',
' Set one: gbrain config set embedding_model openai:text-embedding-3-small',
' Or skip embedding now: re-run with --no-embed',
].join('\n');
case 'unknown_provider':
return [
`Embedding model "${d.model}" uses an unknown provider "${d.provider}".`,
'',
` ${d.message}`,
'',
' Pick a known provider: gbrain config set embedding_model openai:text-embedding-3-small',
].join('\n');
case 'no_touchpoint':
return [
`Provider "${d.provider}" does not offer an embedding touchpoint.`,
'',
' Switch providers: gbrain config set embedding_model openai:text-embedding-3-small',
' Or run with --no-embed to import-only and embed later.',
].join('\n');
case 'user_provided_dims_unset':
return [
`Provider "${d.provider}" ships no default embedding dimension; set one explicitly.`,
'',
` Re-init with the dimension: gbrain init --embedding-dimensions <N>`,
' (embedding_dimensions is a schema-sizing field — `config set` rejects it on purpose)',
' Or run with --no-embed to import-only and embed later.',
].join('\n');
case 'missing_env': {
const envs = d.missingEnvVars.join(', ');
const primaryEnv = d.missingEnvVars[0];
const lines = [
`Embedding model "${d.model}" requires ${envs}.`,
'',
`Set it in your shell, or:`,
` • Re-run with --no-embed to import-only and embed later once the key is set.`,
];
// Only offer a provider-switch hint when the current provider isn't openai
// (otherwise we'd be suggesting they switch to the thing they already have).
if (d.provider !== 'openai') {
lines.push(` • Switch providers: gbrain config set embedding_model openai:text-embedding-3-small`);
} else {
lines.push(` • Switch providers: gbrain config set embedding_model voyage:voyage-3-large`);
}
lines.push('');
lines.push(`Example shell setup:`);
lines.push(` export ${primaryEnv}=...`);
return lines.join('\n');
}
}
}