mirror of
https://github.com/siteboon/claudecodeui.git
synced 2026-05-28 23:15:33 +08:00
* feat: add opencode support
* fix: stabilize opencode session startup
* fix: /models
* fix: improveUI for commands
* fix: format commands.js
* feat: load models through provider adapters
Provider model selection had outgrown a single hardcoded service.
The old service mixed shared caching with provider catalogs and CLI lookup details.
That made stale model lists more likely as providers changed on separate schedules.
Move model discovery behind each provider so lookup lives next to the integration.
The shared service now focuses on provider resolution, caching, persistence, and dedupe.
Return cache metadata and add bypassCache because model availability changes outside the app.
The UI and /models command can show freshness and let users force a provider refresh.
Surface model descriptions while keeping fallback catalogs for unavailable CLIs or SDKs.
* feat(models): resolve active session models through provider adapters
The model inventory command was showing a mix of catalog defaults and
composer-local state instead of the model that is actually active for a
real provider session. That made /models, /cost, and /status
misleading once a session had already started, especially for providers
whose effective runtime model can differ from the optimistic model value
held in the UI.
Introduce an explicit getCurrentActiveModel() contract on
IProviderModels so model resolution lives next to each provider's
catalog logic and uses the provider-native source of truth:
- Claude reads the init event from a resumed stream-json run
- Codex reads model from ~/.codex/config.toml
- Cursor reads lastUsedModel from the chat store.db
- OpenCode reads the persisted session model from opencode.db
- Gemini intentionally returns its default because the CLI does not
provide a reliable active-session lookup
Keep the returned shape intentionally minimal ({ model }). The goal is
to expose only what downstream command consumers need and avoid leaking
provider-specific metadata into a shared transport shape that would
create extra UI coupling and future cleanup cost.
Also make command behavior session-aware: when there is no concrete
session id, do not spawn provider processes or inspect provider session
storage just to answer /models, /cost, or /status. In a new-session
view the correct answer is simply the provider default, and doing more
work there adds latency and unnecessary side effects for no user value.
As part of this, centralize two supporting concerns:
- add a shared helper for building the default current-model result from
a provider catalog so fallbacks stay aligned with DEFAULT
- move leaf-directory validation into shared utils so Cursor session
readers and model lookup code enforce the same path-safety rule
Tests were expanded to cover both the new service delegation path and
the sessionless command behavior, while keeping cache-sensitive tests
isolated from persisted host cache state.
Why this change:
- command output should reflect the model actually driving a session
- new-session views should stay fast and side-effect free
- provider-specific active-model lookup should not be scattered across
routes or UI code
- fallback behavior should be explicit, consistent, and limited to the
provider default when no true active model can be resolved
* feat: support session-scoped model overrides
Model selection was acting like a provider-level preference.
That made resumed sessions drift back to a default or request-time model.
Users expect /models changes made inside a conversation to affect that session.
Store explicit session choices in app-owned ~/.cloudcli state.
This avoids editing provider transcripts or native provider config.
Resolve the effective model before launching each provider runtime.
Claude, Cursor, Codex, Gemini, and OpenCode now honor stored resume choices.
Expose a backend active-model change endpoint for existing sessions.
The models modal can now distinguish default changes from session overrides.
It also shows when a selected model will apply on the next response.
For Claude, stop probing active model state by resuming with a dummy prompt.
Read the indexed JSONL transcript from the end instead.
This preserves provider history while honoring /model stdout or model fields.
Add service tests for adapter delegation and resume-model precedence.
The tests keep cache state, override state, and requested fallback separate.
* feat: make command modal more compact
* fix: preserve opencode session creation events
OpenCode emits the real session id asynchronously on its first JSON output. The runner
registered that id from a helper that could not see the spawned process because
the process reference was scoped inside the model-resolution callback. That
ReferenceError was swallowed by the generic JSON parse fallback, so the client
never received session_created. Without that event, a new OpenCode chat stayed
on / and the assistant stream was not attached to the new session view.
Keep the process reference in the outer spawn scope so registration can update
the active-process map and websocket writer as soon as OpenCode announces the
session id. Split JSON parsing from event processing so malformed non-JSON
output can still stream as raw text, while registration or adapter failures are
surfaced as real errors instead of being hidden as assistant content.
Add a fake opencode executable regression test to lock in the expected lifecycle
ordering: session_created must be sent before live assistant messages, and the
same session id must carry through stream_end and complete.
* fix: clarify model refresh and onboarding providers
OpenCode is now a supported chat provider, but first-run onboarding still only offered
Claude, Cursor, Codex, and Gemini. That made OpenCode harder to discover and
forced users to finish setup before finding the provider in settings or chat.
Adding it to onboarding keeps first-run setup aligned with the providers the
application already supports elsewhere.
The model refresh control was also doing too much visual work. In the new chat
model picker, the previous Hard Refresh label looked like the dialog heading,
which made the primary task unclear. Users open that dialog to choose a model;
refreshing catalogs is only a secondary maintenance action for stale cached
provider model lists.
Rename and reposition the refresh affordance so the model picker reads as a
model picker first. The copy now explains why catalogs are cached, when a refresh
is useful, and that the refresh checks every provider. The /models modal gets the
same clarification so both model-selection surfaces describe the cache behavior
consistently.
* fix: format opencode model catalog labels
OpenCode returns provider-prefixed ids directly from the CLI. Passing those ids through as
labels made the model picker hard to scan: users saw values like
anthropic/claude-3-5-sonnet-20241022 or lowercased, hyphen-split text instead
of readable model names.
Keep the exact OpenCode id as the option value because that is what the CLI
expects, but derive a presentation label for the frontend. The formatter is
intentionally generic rather than a catalog of known providers. It handles common
identifier structure such as provider/model, hyphen-delimited words, v-prefixed
versions, adjacent numeric version tokens, and 8-digit date suffixes.
This keeps OpenCode usable as its model list expands across many upstream
providers without requiring code changes for every new provider or model family.
The description keeps the raw provider-prefixed id visible so users can still
confirm the precise model being selected.
* feat: add more fallback models for cursor
* docs: move model catalog out of shared
The model catalog is no longer a frontend/backend runtime contract.
Keeping it under shared made ownership misleading. It implied the catalog was
application code shared by runtime consumers, even though it now only supports
README links and public API documentation.
Move the catalog into public so it lives beside the docs surfaces that need it.
This gives the API docs a stable, served module and gives README readers a
linkable source without suggesting frontend or backend runtime dependency.
Render the API docs model list from the exported provider registry instead of a
hardcoded Claude/Cursor/Codex subset. That keeps Gemini and OpenCode visible and
makes future provider documentation changes flow through one docs-specific file.
Update README links, provider maintenance notes, and package files so published
artifacts include the standalone docs page and model catalog without relying on
the old shared path.
* fix: simplify empty-state model selector
Keep the provider empty state focused on the setup action users need there:
choosing a model.
The refresh control, cache timestamp, and refresh explanation made the dialog feel
like a cache-management surface.
That extra action is out of place in the empty state, where the goal is to start
a chat with the selected provider and model.
Remove the refresh-specific UI from ProviderSelectionEmptyState and drop the
now-unused refresh/cache props from the ChatMessagesPane pass-through.
Refresh behavior remains available in the dedicated command result flow.
493 lines
14 KiB
TypeScript
493 lines
14 KiB
TypeScript
import type { IncomingMessage } from 'node:http';
|
|
|
|
//----------------- HTTP RESPONSE SHAPES ------------
|
|
/**
|
|
* Canonical success envelope used by backend APIs that return a structured payload.
|
|
*
|
|
* Use this for route handlers that need a stable `success/data` shape so frontend
|
|
* consumers can parse responses consistently across endpoints.
|
|
*/
|
|
export type ApiSuccessShape<TData = unknown> = {
|
|
success: true;
|
|
data: TData;
|
|
};
|
|
|
|
/**
|
|
* Generic plain-object record used when parsing loosely typed JSON payloads.
|
|
*
|
|
* Use this only after runtime shape checks, not as a replacement for validated
|
|
* domain models.
|
|
*/
|
|
export type AnyRecord = Record<string, any>;
|
|
|
|
// ---------------------------
|
|
//----------------- WEBSOCKET TRANSPORT TYPES ------------
|
|
/**
|
|
* Minimal websocket client contract used by backend broadcaster services.
|
|
*
|
|
* Any transport object added to `connectedClients` must implement these two
|
|
* members so shared services can safely send JSON strings and check whether the
|
|
* socket is still open before broadcasting.
|
|
*/
|
|
export type RealtimeClientConnection = {
|
|
readyState: number;
|
|
send(data: string): void;
|
|
};
|
|
|
|
/**
|
|
* Authenticated user payload attached to websocket upgrade requests.
|
|
*
|
|
* Platform and OSS auth flows currently use either `id` or `userId`; both are
|
|
* represented here so websocket handlers can resolve a stable writer user id.
|
|
*/
|
|
export type AuthenticatedWebSocketUser = {
|
|
id?: string | number;
|
|
userId?: string | number;
|
|
username?: string;
|
|
[key: string]: unknown;
|
|
};
|
|
|
|
/**
|
|
* HTTP upgrade request shape after websocket authentication succeeds.
|
|
*
|
|
* `verifyClient` populates `request.user` with the authenticated payload, and
|
|
* downstream websocket handlers rely on this extended request type.
|
|
*/
|
|
export type AuthenticatedWebSocketRequest = IncomingMessage & {
|
|
user?: AuthenticatedWebSocketUser;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- PROVIDER MESSAGE MODEL ------------
|
|
/**
|
|
* Providers supported by the unified server runtime.
|
|
*
|
|
* Use this as the source of truth whenever a function or payload needs to identify
|
|
* a specific LLM integration.
|
|
*/
|
|
export type LLMProvider = 'claude' | 'codex' | 'gemini' | 'cursor' | 'opencode';
|
|
|
|
/**
|
|
* One selectable model row (matches the documentation `public/modelConstants.js` option shape).
|
|
*/
|
|
export type ProviderModelOption = {
|
|
value: string;
|
|
label: string;
|
|
description?: string;
|
|
};
|
|
|
|
/**
|
|
* Provider model catalog returned by `GET /api/providers/:provider/models`.
|
|
*/
|
|
export type ProviderModelsDefinition = {
|
|
OPTIONS: ProviderModelOption[];
|
|
DEFAULT: string;
|
|
};
|
|
|
|
/**
|
|
* Cache metadata returned alongside one provider model catalog.
|
|
*
|
|
* `updatedAt` is when the current cached snapshot was last refreshed from the
|
|
* provider itself. `expiresAt` is the backend cache expiry timestamp, and
|
|
* `source` tells callers whether the current response came from in-memory cache,
|
|
* persisted disk cache, or a fresh provider fetch.
|
|
*/
|
|
export type ProviderModelsCacheInfo = {
|
|
updatedAt: string;
|
|
expiresAt: string;
|
|
source: 'memory' | 'disk' | 'fresh';
|
|
};
|
|
|
|
/**
|
|
* Full provider model lookup result returned by the backend service layer.
|
|
*
|
|
* Use this shape when a caller needs both the selectable model catalog and the
|
|
* cache metadata that explains how current the catalog is.
|
|
*/
|
|
export type ProviderModelsResult = {
|
|
models: ProviderModelsDefinition;
|
|
cache: ProviderModelsCacheInfo;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- PROVIDER ACTIVE MODEL TYPES ------------
|
|
/**
|
|
* Provider-neutral result for the model that is actively driving a session or
|
|
* provider runtime at the time of lookup.
|
|
*
|
|
* `model` must always be populated. Provider adapters should use the
|
|
* provider-specific lookup method requested by the caller, and only fall back
|
|
* to the provider catalog `DEFAULT` value when the active model cannot be read.
|
|
*/
|
|
export type ProviderCurrentActiveModel = {
|
|
model: string;
|
|
};
|
|
|
|
/**
|
|
* Input payload used when one session needs to use a different model on its
|
|
* next resumed turn.
|
|
*
|
|
* This is a backend-owned session override, not a claim that the provider has
|
|
* already switched the currently running session in-place. Provider adapters
|
|
* persist this request so the next CLI/SDK resume can inject the chosen model
|
|
* using the provider-specific mechanism supported by that runtime.
|
|
*/
|
|
export type ProviderChangeActiveModelInput = {
|
|
sessionId: string;
|
|
model: string;
|
|
};
|
|
|
|
/**
|
|
* Provider-neutral session model-change state.
|
|
*
|
|
* `supported` indicates whether the provider adapter supports the app's
|
|
* session-scoped resume override flow. `changed` is the persisted boolean the
|
|
* resume layer checks before forcing a model on the next resumed turn. When
|
|
* `changed` is `false`, `model` is `null` and the runtime should use the
|
|
* normal request/default model selection path.
|
|
*/
|
|
export type ProviderSessionActiveModelChange = {
|
|
provider: LLMProvider;
|
|
sessionId: string;
|
|
supported: boolean;
|
|
changed: boolean;
|
|
model: string | null;
|
|
};
|
|
|
|
/**
|
|
* Message/event variants emitted by provider adapters and normalized transports.
|
|
*
|
|
* Keep this union in sync with event kinds produced by provider session adapters.
|
|
*/
|
|
export type MessageKind =
|
|
| 'text'
|
|
| 'tool_use'
|
|
| 'tool_result'
|
|
| 'thinking'
|
|
| 'stream_delta'
|
|
| 'stream_end'
|
|
| 'error'
|
|
| 'complete'
|
|
| 'status'
|
|
| 'permission_request'
|
|
| 'permission_cancelled'
|
|
| 'session_created'
|
|
| 'interactive_prompt'
|
|
| 'task_notification';
|
|
|
|
/**
|
|
* Provider-neutral message envelope used in REST responses and realtime channels.
|
|
*
|
|
* Every provider-specific message must be converted into this shape before being
|
|
* emitted outside provider-specific modules.
|
|
*/
|
|
export type NormalizedMessage = {
|
|
id: string;
|
|
sessionId: string;
|
|
timestamp: string;
|
|
provider: LLMProvider;
|
|
kind: MessageKind;
|
|
role?: 'user' | 'assistant';
|
|
content?: string;
|
|
/**
|
|
* Optional display-oriented metadata used by providers that need to expose
|
|
* richer transcript artifacts without introducing a brand-new message kind.
|
|
*
|
|
* Current Claude usage:
|
|
* - local slash commands expose parsed command fields
|
|
* - compact summaries are flagged so the UI can treat them differently later
|
|
*/
|
|
displayText?: string;
|
|
commandName?: string;
|
|
commandMessage?: string;
|
|
commandArgs?: string;
|
|
isLocalCommand?: boolean;
|
|
isLocalCommandStdout?: boolean;
|
|
isCompactSummary?: boolean;
|
|
images?: unknown;
|
|
toolName?: string;
|
|
toolInput?: unknown;
|
|
toolId?: string;
|
|
toolResult?: {
|
|
content?: string;
|
|
isError?: boolean;
|
|
toolUseResult?: unknown;
|
|
};
|
|
isError?: boolean;
|
|
text?: string;
|
|
tokens?: number;
|
|
canInterrupt?: boolean;
|
|
requestId?: string;
|
|
input?: unknown;
|
|
context?: unknown;
|
|
reason?: string;
|
|
newSessionId?: string;
|
|
status?: string;
|
|
summary?: string;
|
|
tokenBudget?: unknown;
|
|
subagentTools?: unknown;
|
|
toolUseResult?: unknown;
|
|
sequence?: number;
|
|
rowid?: number;
|
|
[key: string]: unknown;
|
|
};
|
|
|
|
/**
|
|
* Shared options used to fetch historical provider messages.
|
|
*
|
|
* Consumers should pass provider-specific lookup hints (`projectPath`) only
|
|
* when the selected provider requires them.
|
|
*/
|
|
export type FetchHistoryOptions = {
|
|
projectPath?: string;
|
|
limit?: number | null;
|
|
offset?: number;
|
|
};
|
|
|
|
/**
|
|
* Standardized response payload returned from provider history readers.
|
|
*
|
|
* Use this as the contract for APIs that return paginated conversation history.
|
|
*/
|
|
export type FetchHistoryResult = {
|
|
messages: NormalizedMessage[];
|
|
total: number;
|
|
hasMore: boolean;
|
|
offset: number;
|
|
limit: number | null;
|
|
tokenUsage?: unknown;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- PROVIDER SKILL TYPES ------------
|
|
/**
|
|
* Scope where a provider skill definition was discovered.
|
|
*
|
|
* Provider skill adapters should use this to describe the origin of each
|
|
* skill markdown file without leaking provider-specific folder names into route
|
|
* contracts. `repo` is used for Codex repository lookup locations, while
|
|
* `project` is used for providers that treat workspace-local skills as project
|
|
* scoped.
|
|
*/
|
|
export type ProviderSkillScope = 'user' | 'project' | 'plugin' | 'repo' | 'admin' | 'system';
|
|
|
|
/**
|
|
* Shared input accepted by provider skill listing operations.
|
|
*
|
|
* Routes pass `workspacePath` when a caller wants project/repository skills for
|
|
* a specific folder. Providers should fall back to the backend process cwd when
|
|
* this option is omitted.
|
|
*/
|
|
export type ProviderSkillListOptions = {
|
|
workspacePath?: string;
|
|
};
|
|
|
|
/**
|
|
* Normalized skill record returned by provider skill adapters.
|
|
*
|
|
* The `command` value is the exact invocation text the selected provider expects
|
|
* for this skill. Claude plugin skills use a namespaced command such as
|
|
* `/plugin-name:skill-name`, while Codex skills use the `$skill-name` form.
|
|
* `sourcePath` points to the skill markdown file that produced the record so
|
|
* callers can distinguish duplicate skill names across scopes.
|
|
*/
|
|
export type ProviderSkill = {
|
|
provider: LLMProvider;
|
|
name: string;
|
|
description: string;
|
|
command: string;
|
|
scope: ProviderSkillScope;
|
|
sourcePath: string;
|
|
pluginName?: string;
|
|
pluginId?: string;
|
|
};
|
|
|
|
/**
|
|
* Internal source descriptor consumed by shared provider skill discovery logic.
|
|
*
|
|
* Concrete provider adapters build these records from their native lookup rules.
|
|
* The shared skills provider then scans `rootDir` for child skill markdown files
|
|
* and uses `commandForSkill` or `commandPrefix` to produce the provider-specific
|
|
* invocation command. Set `recursive` only when a provider stores skills under
|
|
* arbitrary nested folders below the source root.
|
|
*/
|
|
export type ProviderSkillSource = {
|
|
scope: ProviderSkillScope;
|
|
rootDir: string;
|
|
recursive?: boolean;
|
|
commandPrefix?: '/' | '$';
|
|
commandForSkill?: (skillName: string) => string;
|
|
pluginName?: string;
|
|
pluginId?: string;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- SHARED ERROR TYPES ------------
|
|
/**
|
|
* Optional metadata used when constructing application-level errors.
|
|
*
|
|
* `statusCode` should reflect the HTTP response status, while `code` identifies
|
|
* the stable machine-readable error category.
|
|
*/
|
|
export type AppErrorOptions = {
|
|
code?: string;
|
|
statusCode?: number;
|
|
details?: unknown;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- MCP TYPES ------------
|
|
/**
|
|
* Scope where an MCP server definition is stored and resolved.
|
|
*
|
|
* `user` is global for a user account, `local` is provider-local, and `project`
|
|
* is tied to a specific project path.
|
|
*/
|
|
export type McpScope = 'user' | 'local' | 'project';
|
|
|
|
/**
|
|
* Transport protocol used by an MCP server definition.
|
|
*/
|
|
export type McpTransport = 'stdio' | 'http' | 'sse';
|
|
|
|
/**
|
|
* Normalized MCP server model exposed to frontend and route handlers.
|
|
*
|
|
* Provider adapters should map provider-native config to this structure before
|
|
* returning results.
|
|
*/
|
|
export type ProviderMcpServer = {
|
|
provider: LLMProvider;
|
|
name: string;
|
|
scope: McpScope;
|
|
transport: McpTransport;
|
|
command?: string;
|
|
args?: string[];
|
|
env?: Record<string, string>;
|
|
cwd?: string;
|
|
url?: string;
|
|
headers?: Record<string, string>;
|
|
envVars?: string[];
|
|
bearerTokenEnvVar?: string;
|
|
envHttpHeaders?: Record<string, string>;
|
|
};
|
|
|
|
/**
|
|
* Payload for create/update MCP server operations.
|
|
*
|
|
* Routes and services should accept this type, validate it, and then persist it
|
|
* through provider-specific MCP repositories.
|
|
*/
|
|
export type UpsertProviderMcpServerInput = {
|
|
name: string;
|
|
scope?: McpScope;
|
|
transport: McpTransport;
|
|
workspacePath?: string;
|
|
command?: string;
|
|
args?: string[];
|
|
env?: Record<string, string>;
|
|
cwd?: string;
|
|
url?: string;
|
|
headers?: Record<string, string>;
|
|
envVars?: string[];
|
|
bearerTokenEnvVar?: string;
|
|
envHttpHeaders?: Record<string, string>;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- PROVIDER AUTH TYPES ------------
|
|
/**
|
|
* Authentication status result returned by provider health checks.
|
|
*
|
|
* This shape is consumed by settings/status endpoints to report installation and
|
|
* credential state for each provider.
|
|
*/
|
|
export type ProviderAuthStatus = {
|
|
installed: boolean;
|
|
provider: LLMProvider;
|
|
authenticated: boolean;
|
|
email: string | null;
|
|
method: string | null;
|
|
error?: string;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- SHARED DATABASE CREDENTIAL TYPES ------------
|
|
/**
|
|
* Safe credential view returned by credential listing APIs.
|
|
*
|
|
* This intentionally excludes the raw credential secret while still exposing
|
|
* metadata needed for UI rendering and management operations.
|
|
*/
|
|
export type CredentialPublicRow = {
|
|
id: number;
|
|
credential_name: string;
|
|
credential_type: string;
|
|
description: string | null;
|
|
created_at: string;
|
|
is_active: number;
|
|
};
|
|
|
|
/**
|
|
* Result returned after creating a credential record.
|
|
*
|
|
* Use this return shape when callers need the created id and display metadata,
|
|
* but must never receive the stored secret value.
|
|
*/
|
|
export type CreateCredentialResult = {
|
|
id: number | bigint;
|
|
credentialName: string;
|
|
credentialType: string;
|
|
};
|
|
|
|
// ---------------------------
|
|
//----------------- PROJECT PERSISTENCE TYPES ------------
|
|
/**
|
|
* Canonical project row shape returned by the projects repository.
|
|
*
|
|
* Use this type whenever backend services need to pass around one database
|
|
* project record without leaking raw SQL row typing across modules.
|
|
*/
|
|
export type ProjectRepositoryRow = {
|
|
project_id: string;
|
|
project_path: string;
|
|
custom_project_name: string | null;
|
|
isStarred: number;
|
|
isArchived: number;
|
|
};
|
|
|
|
/**
|
|
* Result category returned by `projectsDb.createProjectPath`.
|
|
*
|
|
* `created` means a fresh row was inserted, `reactivated_archived` means an
|
|
* existing archived path was accepted and updated, and `active_conflict` means
|
|
* an already-active path blocked project creation.
|
|
*/
|
|
export type CreateProjectPathOutcome =
|
|
| 'created'
|
|
| 'reactivated_archived'
|
|
| 'active_conflict';
|
|
|
|
/**
|
|
* Structured result returned by project-path upsert operations.
|
|
*
|
|
* Services should use this result to decide whether a request succeeded,
|
|
* should return a conflict, or needs follow-up retrieval of row metadata.
|
|
*/
|
|
export type CreateProjectPathResult = {
|
|
outcome: CreateProjectPathOutcome;
|
|
project: ProjectRepositoryRow | null;
|
|
};
|
|
|
|
/**
|
|
* Validation result for user-supplied workspace/project paths.
|
|
*
|
|
* `resolvedPath` is present only when validation succeeds. `error` is present
|
|
* only when validation fails and is suitable for user-facing diagnostics.
|
|
*/
|
|
export type WorkspacePathValidationResult = {
|
|
valid: boolean;
|
|
resolvedPath?: string;
|
|
error?: string;
|
|
};
|