Files
Bohdan Triapitsyn 28f0736d69 feat: small-model utility calls on existing OpenCode providers (#2049)
Adds a server-side "small model" capability: direct, cheap LLM calls that
reuse the user's existing OpenCode provider logins — the mechanism OpenCode
uses internally for titles and summaries but does not expose through the
SDK or plugins. Zero new dependencies; plain fetch with per-provider wire
formats, credentials never leave the server.

Core (packages/web/server/lib/small-model):
- Resolution mirrors OpenCode's session scoping: explicit settings override
  → small_model from the OpenCode config → family scan within the session's
  provider → the session's own model. The global provider scan only serves
  callers without a session context, and background callers forbid it
  entirely (restrictToPreferredProvider), so conversation content never
  reaches a provider the user didn't pick — explicit choices excepted.
- Per-provider auth replicating OpenCode's plugin loaders: GitHub Copilot
  (device token as bearer, no exchange), ChatGPT plan via the codex
  Responses API (single-flight OAuth refresh written back to auth.json),
  Anthropic messages, Google generateContent, generic OpenAI-compatible.
- OpenCode's free models (opencode/big-pickle, *-free) are never called
  directly; unauthenticated providers are skipped by design.
- Prompt clamping to the model's catalog context limit; thinking disabled
  where a wire switch exists (Z.AI/GLM, MiniMax-M3, Gemini Flash); robust
  content parsing with a clear error when a thinking model spends its whol
  budget on reasoning.
- Settings → Sessions gains a Small Model group: use-default checkbox plus
  an override picker limited to authenticated providers, persisted with
  web/desktop/VS Code sanitization parity.

Consumers:
- Session assist: a server-side watcher on the global SSE hub generates a
  short recap and one suggested follow-up after a session idles quietly fo
  a minute, stored on session metadata (openchamber.assist). Freshness is
  keyed to the last assistant message id, so new activity invalidates the
  payload everywhere with no extra writes. The chat shows the recap under
  the last message after five quiet minutes and the suggestion as a
  dismissible chip above the composer (tap fills the input, never sends).
  Gated by a new Chat setting (default on) that is a hard generation
  switch. Language is anchored to the conversation itself, with a
  script-mismatch guard against model/backend language hallucination.
- TTS: a third input mode, summarized — long replies are condensed to
  spoken prose before playback on any TTS engine.
- Git: commit-message and PR generation moved off the active chat session
  onto the small model fed with real diffs and the commit list (bodies
  included), with a session-transport fallback for free-model-only setups.
- Notes: Add to notes distills long selections into 1-3 dense sentences
  preserving exact identifiers, with verbatim fallback on failure.

Fixes along the way:
- The global event watcher now starts unconditionally; it was gated behind
  the desktop-notify env, leaving the server-side event hub dead in
  packaged apps.
- OpenCode re-emits message.updated for old user messages after idle; the
  watcher no longer mistakes those for new activity.
- Session metadata merges from a fresh read right before the PATCH, so
  writes made during the generation window (suggestion dismissals, review
  links) are preserved; the assist runtime stops during graceful shutdown.
2026-07-05 23:19:10 +03:00

132 lines
5.2 KiB
JavaScript

import { getCatalogProvider } from './catalog.js';
// Mirrors OpenCode's getSmallModel fallback chain:
// 1. `small_model` from the merged config layers ("provider/model").
// 2. GitHub Copilot's hidden utility models when Copilot is logged in.
// 3. Family-priority scan of the authenticated providers' catalog models.
const FAMILY_PRIORITY = ['gemini-flash', 'gpt-nano', 'claude-haiku'];
const COPILOT_UTILITY_MODELS = ['gpt-5.4-nano', 'gpt-4.1', 'gpt-4o', 'gpt-4o-mini'];
// The ChatGPT-plan codex backend only accepts a small allowlist of models
// (nano/API-key models are rejected with 400) — this is its cheapest one.
const OPENAI_OAUTH_SMALL_MODEL = 'gpt-5.4-mini';
const AUTH_PROVIDER_ALIASES = {
'github-copilot': ['github-copilot', 'copilot'],
};
export function getAuthEntryForProvider(auth, providerID) {
const aliases = AUTH_PROVIDER_ALIASES[providerID] || [providerID];
for (const alias of aliases) {
const entry = auth?.[alias];
if (entry && typeof entry === 'object') {
return entry;
}
}
return null;
}
export function isUsableAuthEntry(entry) {
if (!entry || typeof entry !== 'object') return false;
if (entry.type === 'api') return typeof entry.key === 'string' && entry.key.length > 0;
if (entry.type === 'oauth') {
return (typeof entry.access === 'string' && entry.access.length > 0)
|| (typeof entry.refresh === 'string' && entry.refresh.length > 0);
}
if (entry.type === 'wellknown') return typeof entry.token === 'string' && entry.token.length > 0;
return false;
}
export function parseModelRef(value) {
if (typeof value !== 'string') return null;
const trimmed = value.trim();
const slash = trimmed.indexOf('/');
if (slash <= 0 || slash === trimmed.length - 1) return null;
return {
providerID: trimmed.slice(0, slash),
modelID: trimmed.slice(slash + 1),
};
}
const pickByFamily = (models, family) => {
const matches = Object.values(models)
.filter((model) => model && typeof model === 'object' && model.family === family);
if (matches.length === 0) return null;
matches.sort((a, b) => String(b.release_date || '').localeCompare(String(a.release_date || '')));
return matches[0];
};
// Small-model candidates within ONE provider, by family priority. Copilot and
// ChatGPT-plan OpenAI have fixed small models that never appear in the
// catalog; everyone else is scanned through the catalog families.
const pickWithinProvider = (providerID, auth, catalog, family) => {
if (providerID === 'openai' && auth.openai?.type === 'oauth') {
return family === 'gpt-nano'
? { providerID, modelID: OPENAI_OAUTH_SMALL_MODEL, source: 'codex-small' }
: null;
}
if (providerID === 'github-copilot') {
return family === 'gpt-nano'
? { providerID, modelID: COPILOT_UTILITY_MODELS[0], source: 'copilot-utility' }
: null;
}
const provider = getCatalogProvider(catalog, providerID);
if (!provider || !provider.models || typeof provider.models !== 'object') return null;
const model = pickByFamily(provider.models, family);
return model?.id ? { providerID, modelID: model.id, source: 'family-scan' } : null;
};
export function resolveSmallModel({ auth, catalog, settingsSmallModel, configSmallModel, preferredProviderID, preferredModelID }) {
// OpenChamber's own setting (Settings → Sessions → Small Model override)
// outranks everything, including the OpenCode config.
const fromSettings = parseModelRef(settingsSmallModel);
if (fromSettings) {
return { ...fromSettings, source: 'settings' };
}
const explicit = parseModelRef(configSmallModel);
if (explicit) {
return { ...explicit, source: 'config' };
}
// Like OpenCode: when the caller has a session context, the utility call
// stays on the session's provider. Scan its families for a small model,
// otherwise run on the session's own model — never silently switch to a
// different provider's subscription.
const preferred = typeof preferredProviderID === 'string' && preferredProviderID
? preferredProviderID
: null;
if (preferred && isUsableAuthEntry(getAuthEntryForProvider(auth, preferred))) {
for (const family of FAMILY_PRIORITY) {
const match = pickWithinProvider(preferred, auth, catalog, family);
if (match) return match;
}
if (typeof preferredModelID === 'string' && preferredModelID) {
return { providerID: preferred, modelID: preferredModelID, source: 'session-model' };
}
}
// No session context (or its provider has no usable login): scan all
// authenticated providers by family priority.
const authedProviders = Object.keys(auth || {}).filter((providerID) =>
providerID !== preferred && isUsableAuthEntry(auth[providerID]));
for (const family of FAMILY_PRIORITY) {
for (const providerID of authedProviders) {
const match = pickWithinProvider(providerID, auth, catalog, family);
if (match) return match;
}
}
// Copilot's utility fallback for legacy auth aliases the loop above missed.
const copilotEntry = getAuthEntryForProvider(auth, 'github-copilot');
if (isUsableAuthEntry(copilotEntry)) {
return {
providerID: 'github-copilot',
modelID: COPILOT_UTILITY_MODELS[0],
source: 'copilot-utility',
};
}
return null;
}