LibreChat/api/server/services/Files/process.js
Danny Avila 35bf04b26c 🧰 refactor: Unify code-execution tools (#12767)
* 🛠️ feat: Add registerCodeExecutionTools helper

Idempotently registers `bash_tool` + `read_file` in the run's tool
registry and tool-definition list via a registry `.has()` dedupe. Sets
up the single code-execution tool path shared by:

- `initializeAgent` (when an agent has `execute_code` in its tools and
  the capability is enabled for the run)
- `injectSkillCatalog` (when skills are active; unconditional read_file,
  bash_tool follows `codeEnvAvailable`)

Both callers reach the helper in the same initialization sequence, so
the second call becomes a no-op and exactly one copy of each tool
reaches the LLM — no more double registration for agents that combine
`execute_code` capability with active skills.

Unit-tested on a fresh run, idempotence (second call, overlap with
prior tooldefs, partial overlap), and the no-registry variant.

* 🔀 refactor: Route injectSkillCatalog bash_tool + read_file through registerCodeExecutionTools

The `skill` tool is still registered inline (it's skill-path-specific),
but `bash_tool` + `read_file` now flow through the shared idempotent
helper so a prior registration from the execute_code path doesn't
produce a duplicate copy later in the same run. Behavior preserved:

- `read_file` always registers when any active skill is in scope —
  manually-primed `disable-model-invocation: true` skills still need it
  to load `references/*` from storage.
- `bash_tool` follows `codeEnvAvailable` exactly as before.

Adds a test pinning the cross-call dedupe: when `injectSkillCatalog`
runs AFTER `registerCodeExecutionTools` has already seeded the registry
+ tool definitions with bash_tool/read_file, the resulting
`toolDefinitions` still contains exactly one copy of each.

* 🪄 feat: Expand `execute_code` tool name into bash_tool + read_file at initialize-time

When an agent's `tools` include `execute_code` and the `execute_code`
capability is enabled for the run, `initializeAgent` now registers
`bash_tool` + `read_file` via `registerCodeExecutionTools` before
`injectSkillCatalog`. The legacy `execute_code` tool definition is no
longer handed to the LLM — `execute_code` remains on the agent
document as a capability-trigger marker, but the runtime expands it
into the skill-flavored tool pair.

Call ordering matters: the `execute_code` registration runs BEFORE
`injectSkillCatalog`, so the skill path's own `registerCodeExecutionTools`
call inside `injectSkillCatalog` becomes a no-op via the registry's
`.has()` check. Exactly one copy of each tool reaches the LLM whether
the agent has:

- only `execute_code` (legacy path)
- only skills
- both

No data migration needed — `agent.tools: ['execute_code']` stays in
the DB unchanged; the expansion is a runtime operation.

Three tests cover the matrix: execute_code + capability on →
bash_tool + read_file registered; execute_code + capability off →
neither registered; no execute_code + capability on → neither
registered.

* 🗑️ refactor: Drop CodeExecutionToolDefinition from the builtin registry

Removes the legacy `execute_code` entry from `agentToolDefinitions` and
the corresponding import. With the initialize-time expansion in place,
nothing consults `getToolDefinition('execute_code')` for a tool schema
any more — the capability gate still filters on the string
`execute_code`, but the actual tool definitions the LLM sees come from
`registerCodeExecutionTools` (i.e. `bash_tool` + `read_file`).

`loadToolDefinitions` in `packages/api/src/tools/definitions.ts`
silently drops `execute_code` when it no longer resolves in the
registry — that's the expected path and is now covered by an updated
test. No caller of `getToolDefinition('execute_code')` expects a
non-undefined result after this change.

* 🔌 refactor: Read CODE_API_KEY from env for primeCodeFiles + PTC

Finishes the Phase 4 server-env-keyed rollout on the two remaining
`loadAuthValues({ authFields: [EnvVar.CODE_API_KEY] })` sites in
`ToolService.js`:

- `primeCodeFiles` (user-attached file priming on execute_code agents)
- Programmatic Tool Calling (`createProgrammaticToolCallingTool`)

Both now read `process.env[EnvVar.CODE_API_KEY]` directly, matching
`bash_tool`'s pattern. The per-user plugin-auth path is no longer
consulted for code-env credentials anywhere in the hot path — the
agents library owns the actual tool-call execution and also reads the
env var internally.

Priming still fires for existing user-file workflows so the legacy
`toolContextMap[execute_code]` hint ("files available at /mnt/data/...")
stays in the prompt; only the key lookup changed.

* 🔧 fix: Type the pre-seeded dedupe-test tools as LCTool

CI TypeScript type checks caught `{ parameters: {} }` in the new
cross-call dedupe test: `LCTool.parameters` is a `JsonSchemaType`,
not `{}`. Use `{ type: 'object', properties: {} }` and type the
local registry Map through the parameter-derived shape so the
pre-seeded values match what `toolRegistry.set` expects.

* 🛡️ fix: Run execute_code expansion before GOOGLE_TOOL_CONFLICT gate

Codex review caught a latent regression: the original Phase 8 placement
ran `registerCodeExecutionTools` after `hasAgentTools` was computed,
so an execute-code-only agent on Google/Vertex with provider-specific
`options.tools` populated would no longer trip `GOOGLE_TOOL_CONFLICT`
— the legacy `CodeExecutionToolDefinition` used to populate
`toolDefinitions` before the guard, but after dropping it from the
registry, `toolDefinitions` stayed empty until my expansion ran
downstream of the guard. Mixed provider + agent tools would silently
flow through to the LLM.

Fix moves the `execute_code` expansion to BEFORE `hasAgentTools`
computation. `bash_tool` + `read_file` now contribute to the check
the same way the legacy `execute_code` def did. Covered by a new
test that pins the Google+execute_code+provider-tools scenario —
the `rejects.toThrow(/google_tool_conflict/)` path would have
silently passed on the prior placement.

* 🔗 fix: Thread codeEnvAvailable through handoff sub-agents

Round-2 codex review caught the other half of the execute_code
expansion gap: `discoverConnectedAgents` omitted `codeEnvAvailable`
from its forwarded `initializeAgent` params, so handoff sub-agents
with `agent.tools: ['execute_code']` lost the `bash_tool` + `read_file`
registration (pre-Phase 8 the legacy `CodeExecutionToolDefinition`
would have landed in their `toolDefinitions` via the registry).

- Add `codeEnvAvailable?` to `DiscoverConnectedAgentsParams` and
  forward it verbatim on every sub-agent `initializeAgent` call.
- Update the three JS call sites that construct the primary's
  `codeEnvAvailable` (`services/Endpoints/agents/initialize.js`,
  `controllers/agents/openai.js`, `controllers/agents/responses.js`)
  to pass the same flag into `discoverConnectedAgents` — one
  authoritative source per request.
- Two regression tests in `discovery.spec.ts` pin the true/false
  passthrough so a future refactor that drops the param-forwarding
  surfaces immediately.

Left intentionally unchanged: `packages/api/src/agents/openai/service.ts`
(public API helper with no in-repo caller). External consumers of
`createAgentChatCompletion` who want code execution should pass a
`codeEnvAvailable`-aware `initializeAgent` via `deps` — documenting
the full public-API surface is out of scope for this Phase 8 PR.

* 🔗 fix: Thread codeEnvAvailable through addedConvo + memory-agent paths

Round-3 codex review caught the last two production `initializeAgent`
callers missing the Phase-8 capability flag:

- `api/server/services/Endpoints/agents/addedConvo.js` (multi-convo
  parallel agent execution). Added `codeEnvAvailable` to
  `processAddedConvo`'s destructured params and forwarded it into
  the per-added-agent `initializeAgent` call. Caller in
  `api/server/services/Endpoints/agents/initialize.js` passes the
  same `codeEnvAvailable` it computed for the primary.
- `api/server/controllers/agents/client.js` (`useMemory` — memory
  extraction agent). Computes its own `codeEnvAvailable` from
  `appConfig?.endpoints?.[EModelEndpoint.agents]?.capabilities` and
  forwards into `initializeAgent`. Memory agents rarely list
  `execute_code`, but if one does, pre-Phase 8 they got the legacy
  `execute_code` tool registered unconditionally — the passthrough
  restores parity.

With this, every production caller of `initializeAgent` explicitly
resolves the capability: main chat flow (primary + handoff), OpenAI
chat completions (primary + handoff), Responses API (primary + handoff),
added convo parallel agents, and memory agents. The one remaining
caller, `packages/api/src/agents/openai/service.ts::createAgentChatCompletion`,
is a public API helper with no in-repo consumer (external callers
must pass a capability-aware `initializeAgent` via `deps`).

* 🪤 fix: Remove duplicate appConfig declaration causing TDZ ReferenceError

The Responses API controller had TWO `const appConfig = req.config;`
bindings inside `createResponse`: one at the top of the function
(added by the Phase 4 `bash_tool` decouple) and one inside the try
block (added by the polish PR #12760). Because `const` is block-scoped
with a temporal dead zone, the inner redeclaration put `appConfig` in
TDZ for the entire try block, so any earlier reference inside the
try — notably `appConfig?.endpoints?.[EModelEndpoint.agents]?.allowedProviders`
at line 348 — threw `ReferenceError: Cannot access 'appConfig'
before initialization`. The error was silently swallowed by the
outer try/catch, leaving `recordCollectedUsage` unreached and the
six `responses.unit.spec.js` token-usage tests failing.

Removing the inner redeclaration fixes the six failing tests
(verified: 11/11 pass locally post-fix, 0 regressions elsewhere).
The outer function-scoped binding already provides `appConfig` to
every downstream reference.

* 🔗 fix: Thread codeEnvAvailable through the OpenAI chat-completion public API

Round-4 codex review (legitimate on the type-safety angle, even though the
runtime concern was already covered): the `createAgentChatCompletion`
helper defines its own narrower `InitializeAgentParams` interface locally,
and the type was missing `codeEnvAvailable`. External consumers who
supply a capability-aware `deps.initializeAgent` couldn't route
`codeEnvAvailable` through without a type-cast workaround.

- Widen the local `InitializeAgentParams` interface to include
  `codeEnvAvailable?: boolean` (matches the real
  `packages/api/src/agents/initialize.ts` type).
- Derive `codeEnvAvailable` inside `createAgentChatCompletion` from
  `deps.appConfig?.endpoints?.agents?.capabilities` (the same source
  the in-repo controllers use) and forward to `deps.initializeAgent`.
  Uses a string literal `'execute_code'` lookup so this file stays free
  of a `librechat-data-provider` import — keeping the dependency surface
  of the public helper minimal.

With this, external consumers of `createAgentChatCompletion` who pass
`appConfig` with the agents capabilities get `bash_tool` + `read_file`
registration automatically; consumers who don't pass `appConfig` retain
the existing "explicit opt-in" semantics (the flag stays `undefined`,
expansion is skipped).

* 🧹 chore: Review-driven polish — observability log, JSDoc DRY, test gaps, no-op allocation

Addresses the comprehensive review of PR #12767:

- **Finding #1** (MINOR, observability): `initializeAgent` now emits a
  debug log when an agent lists `execute_code` in its tools but the
  runtime gate is off (`params.codeEnvAvailable` !== true). The
  event-driven `loadToolDefinitionsWrapper` path doesn't log
  capability-disabled warnings, so without this the tool silently
  vanishes from the LLM's definitions with zero trace. Operators
  debugging "why isn't code interpreter working?" now get a signal at
  the initialize layer.

- **Finding #5** (NIT, allocation): `registerCodeExecutionTools` now
  returns the input `toolDefinitions` array by reference on the no-op
  path (both tools already registered by a prior caller in the same
  run) instead of allocating a fresh spread array every time. The
  common dual-call scenario — `initializeAgent` then
  `injectSkillCatalog` — saves one O(n) copy per request.

- **Finding #4** (NIT, DRY): Collapsed the duplicated 6-line JSDoc
  comment in `openai.js`, `responses.js`, and `addedConvo.js` into
  either a one-line `@see DiscoverConnectedAgentsParams.codeEnvAvailable`
  pointer (the two JS call sites) or a compact 3-line block referring
  back to the canonical source (addedConvo's @param).

- **Finding #2** (MINOR, test gap): Added
  `api/server/services/Endpoints/agents/addedConvo.spec.js` with three
  cases covering `codeEnvAvailable=true`, `codeEnvAvailable=false`,
  and omitted (undefined) passthrough. A future refactor that drops
  the param from destructuring now surfaces here instead of silently
  regressing multi-convo parallel agents with `execute_code`.

- **Finding #3** (MINOR, test gap): Added
  `api/server/controllers/agents/__tests__/client.memory.spec.js`
  pinning the capability-flag derivation that `AgentClient::useMemory`
  uses — six cases covering present/absent/null/undefined config shapes
  plus an enum-literal pin (`'execute_code'` / `'agents'`). Catches
  enum renames or config-path shifts that would otherwise silently
  strip `bash_tool` + `read_file` from memory agents.

Finding #7 (jest.mock scoping, confidence 40) left as-is: the
reviewer's own risk assessment noted `buildToolSet` doesn't touch
the mocked exports, and restructuring a file-level `jest.mock` to
`jest.doMock` + dynamic `import()` introduces more complexity than
the speculative risk justifies. The existing mock is scoped to the
test file and contains the same stubs the adjacent
`skills.test.ts` already uses.

Finding #6 (PR description commit count) addressed out-of-band via
PR description update.

All existing tests pass, typecheck clean, lint clean across touched
files. New tests: 9 cases across 2 new spec files.

* 🧽 refactor: Replace hardcoded 'execute_code' string with AgentCapabilities enum in service.ts

Follow-up review (conf 55) caught that `openai/service.ts`'s Phase 8
`codeEnvAvailable` derivation used the literal `'execute_code'` while
every in-repo controller uses `AgentCapabilities.execute_code` from
`librechat-data-provider`. The file deliberately uses local type
interfaces to keep the public API helper's type surface small, but
that pattern was never a ban on single-value imports from the data
provider — `packages/api` already depends on it. Importing the enum
value means a future rename of `AgentCapabilities.execute_code`
propagates to this file automatically, matching the in-repo
controllers' behavior.

Other follow-up findings left as-is per the reviewer's own verdict:

- #2 (memory spec mirrors the production expression rather than
  calling `AgentClient::useMemory` directly): reviewer flagged as
  "not blocking" / "design-philosophy observation." The test file's
  JSDoc already explicitly documents the tradeoff and pins the enum
  literals to catch the most likely drift vector. Standing up
  `AgentClient` + all its mocks for a one-line regression guard is
  disproportionate.
- #3 (`addedConvo.spec.js` mock signature vs. underlying
  `loadAddedAgent` arity): reviewer's own confidence 25 noted the
  mock matches the wrapper's actual call pattern in the production
  file. Not a real gap.
- #4 was self-retracted as a false alarm.

* 🗑️ refactor: Fully deprecate CODE_API_KEY — remove all LibreChat-side references

The code-execution sandbox no longer authenticates via a per-run
`CODE_API_KEY` (frontend or backend). Auth moved server-side into the
agents library / sandbox service, so LibreChat drops every reference:

**Backend plumbing:**
- `api/server/services/Files/Code/crud.js`: `getCodeOutputDownloadStream`,
  `uploadCodeEnvFile`, `batchUploadCodeEnvFiles` no longer accept
  `apiKey` or send the `X-API-Key` header.
- `api/server/services/Files/Code/process.js`: `processCodeOutput`,
  `getSessionInfo`, `primeFiles` drop the `apiKey` param throughout.
- `api/server/services/ToolService.js`: stop reading
  `process.env[EnvVar.CODE_API_KEY]` for `primeCodeFiles` and PTC; the
  agents library handles auth internally. Remove the now-dead
  `loadAuthValues` + `EnvVar` imports. Drop the misleading
  "LIBRECHAT_CODE_API_KEY" hint from the bash_tool error log.
- `api/server/services/Files/process.js`: remove the `loadAuthValues`
  call around `uploadCodeEnvFile`.
- `api/server/routes/files/files.js`: code-env file download no longer
  fetches a per-user key.
- `api/server/controllers/tools.js`: `execute_code` is no longer a
  tool that needs verifyToolAuth with `[EnvVar.CODE_API_KEY]` — the
  endpoint always reports system-authenticated so the client skips
  the key-entry dialog. `processCodeOutput` called without `apiKey`.
- `api/server/controllers/agents/callbacks.js`: `processCodeOutput`
  invoked without the loadAuthValues round trip, for both LegacyHandler
  and Responses-API handlers.
- `api/app/clients/tools/util/handleTools.js`: `createCodeExecutionTool`
  called with just `user_id` + files.

**packages/api:**
- `packages/api/src/agents/skillFiles.ts`: `PrimeSkillFilesParams`,
  `PrimeInvokedSkillsDeps`, `primeSkillFiles`, `primeInvokedSkills` all
  drop the `apiKey` param; the gate is purely `codeEnvAvailable`.
- `packages/api/src/agents/handlers.ts`: `handleSkillToolCall` drops
  the `process.env[EnvVar.CODE_API_KEY]` read; skill-file priming is
  now gated solely on `codeEnvAvailable`. `ToolExecuteOptions`
  signatures drop apiKey from `batchUploadCodeEnvFiles` and
  `getSessionInfo`.
- `packages/api/src/agents/skillConfigurable.ts`: JSDoc no longer
  references the env var.
- `packages/api/src/tools/classification.ts`: PTC creation no longer
  gated on `loadAuthValues`; `buildToolClassification` drops the
  `loadAuthValues` dep entirely (no LibreChat-side callers need it for
  this path anymore).
- `packages/api/src/tools/definitions.ts`: `LoadToolDefinitionsDeps`
  drops the `loadAuthValues` field.

**Frontend:**
- Delete `client/src/hooks/Plugins/useAuthCodeTool.ts`,
  `useCodeApiKeyForm.ts`, and
  `client/src/components/SidePanel/Agents/Code/ApiKeyDialog.tsx` —
  the install/revoke dialogs for CODE_API_KEY are fully dead.
- `BadgeRowContext.tsx`: drop `codeApiKeyForm` from the context type and
  provider. `codeInterpreter` toggle treated as always authenticated
  (sandbox auth is server-side).
- `ToolsDropdown.tsx`, `ToolDialogs.tsx`, `CodeInterpreter.tsx`,
  `RunCode.tsx`, `SidePanel/Agents/Code/Action.tsx` +`Form.tsx`: all
  API-key dialog trigger refs, "Configure code interpreter" gear
  buttons, and auth-verification plumbing removed. The
  "Code Interpreter" toggle is now a plain `AgentCapabilities.execute_code`
  checkbox — no key-entry gate.
- `client/src/locales/en/translation.json`: drop the three
  `com_ui_librechat_code_api*` keys and `com_ui_add_code_interpreter_api_key`.
  Other locales are externally automated per CLAUDE.md.

**Config:**
- `.env.example`: remove the `# LIBRECHAT_CODE_API_KEY=your-key` section
  and its header.

**Tests:**
- `crud.spec.js`: assertions flipped to pin "no X-API-Key header" and
  "no apiKey param".
- `skillFiles.spec.ts`: removed env-var save/restore; tests now pin
  that the batch-upload path is gated solely on `codeEnvAvailable` and
  that no apiKey is threaded through.
- `handlers.spec.ts`: same — just the `codeEnvAvailable` gate pins
  remain.
- `classification.spec.ts`: remove the two tests that asserted
  `loadAuthValues` was (not) called for PTC.
- `definitions.spec.ts`: drop every `loadAuthValues: mockLoadAuthValues`
  entry from the deps shape.
- `process.spec.js`: strip the mock of `EnvVar.CODE_API_KEY`.

**Comment hygiene:**
- `tools.ts`, `initialize.ts`, `registry/definitions.ts`: shortened
  stale comment references to "legacy `execute_code` tool" without
  naming the retired env var.

Tests verified: 678 packages/api tests pass, 836 backend api tests
pass. Typecheck clean, lint clean. Only remaining CODE_API_KEY
mentions in the code are two regression-guard assertions:
- `crud.spec.js`: pins "no X-API-Key header" stays absent.
- `skillConfigurable.spec.ts`: pins `configurable` never grows a
  `codeApiKey` field.

* 🧹 chore: Remove the last two CODE_API_KEY name mentions in LibreChat

Follow-up to the prior full deprecation commit: two tests still named
the retired identifier in their regression-guard assertions.

- `packages/api/src/agents/skillConfigurable.spec.ts`: drop the
  "does not inject a codeApiKey key" test. The `codeApiKey` field is
  gone from the production configurable shape, so an absence-assertion
  naming it re-introduces the retired identifier in code.
- `api/server/services/Files/Code/crud.spec.js`: rename the
  "without an X-API-Key header" case back to "should request stream
  response from the correct URL" and drop the
  `expect(headers).not.toHaveProperty('X-API-Key')` assertion. The
  surrounding request-shape checks (URL, timeout, responseType) still
  pin the behavior; the explicit header-absence line was named-after
  the deprecated contract.

Result: `grep -rn "CODE_API_KEY\|codeApiKey\|LIBRECHAT_CODE_API_KEY"`
against the LibreChat source tree returns zero hits. The only
remaining `X-API-Key` strings in this repo are on unrelated OpenAPI
Action + MCP server auth configurations, where the string is
user-facing config, not a LibreChat-owned identifier.

Tests: 677 packages/api pass (2 pre-existing summarization e2e failures
unrelated); 126 api-workspace controller/service tests pass.
Typecheck and lint clean.

* 🎯 fix: Narrow codeEnvAvailable to per-agent (admin cap AND agent.tools)

Before this commit, `codeEnvAvailable` was computed in the three JS
controllers as the admin-level capability flag only
(`enabledCapabilities.has(AgentCapabilities.execute_code)`) and passed
through `initializeAgent` → `injectSkillCatalog` / `primeInvokedSkills` /
`enrichWithSkillConfigurable` unchanged. A skills-only agent whose
`tools` array didn't include `execute_code` still got `bash_tool`
registered (via `injectSkillCatalog`) and skill files re-primed to the
sandbox on every turn — wrong, because the agent never opted in to
code execution.

**Fix:** `initializeAgent` now computes the per-agent effective value
once as `params.codeEnvAvailable === true && agent.tools.includes(Tools.execute_code)`,
reuses the same boolean for:

1. The `execute_code` → `bash_tool + read_file` expansion gate
   (previously already consulted `agent.tools`; now shares the single
   `effectiveCodeEnvAvailable` binding).
2. The `injectSkillCatalog` call (previously got the raw admin flag).
3. The returned `InitializedAgent.codeEnvAvailable` field (new, typed as
   required boolean).

**Controllers (initialize.js, openai.js, responses.js):** store
`primaryConfig.codeEnvAvailable` in `agentToolContexts.set(primaryId, ...)`,
capture `config.codeEnvAvailable` in every handoff `onAgentInitialized`
callback, and read it from the per-agent ctx inside the
`toolExecuteOptions.loadTools` runtime closure. The hoisted
`const codeEnvAvailable = enabledCapabilities.has(...)` locals in the
two OpenAI-compat controllers are gone — they were shadowing the
narrowed per-agent value.

**primeInvokedSkills:** `handlePrimeInvokedSkills` in
`services/Endpoints/agents/initialize.js` now uses
`primaryConfig.codeEnvAvailable` (per-agent, narrowed) instead of the
raw admin flag. A skills-only primary agent won't re-prime historical
skill files to the sandbox even when the admin enabled the capability
globally.

**Efficiency:** one extra `&&` in `initializeAgent`. No runtime hot-path
cost — the `includes()` scan on `agent.tools` was already happening for
the `execute_code` expansion gate; it's now just bound to a local. Tool
execution closures read `ctx.codeEnvAvailable === true` (property
access + strict equality, O(1)).

**Ephemeral-agent note:** per-agent narrowing is authoritative for both
persisted and ephemeral flows. The ephemeral toggle
(`ephemeralAgent.execute_code`) is reconciled into `agent.tools`
upstream in `packages/api/src/agents/added.ts`, so
`agent.tools.includes('execute_code')` is the single source of truth
by the time `initializeAgent` runs.

**Tests:** two new regression tests pin the narrowing contract:

- `initialize.test.ts` — four-quadrant matrix on
  `InitializedAgent.codeEnvAvailable` (cap on × agent asks, cap on ×
  doesn't ask, cap off × asks, neither). Catches future refactors that
  drop either half of the AND.
- `skills.test.ts` — `injectSkillCatalog` with `codeEnvAvailable: false`
  against an active skill catalog must NOT register `bash_tool` even
  though it still registers `read_file` + `skill`. This is the state
  a skills-only agent gets post-narrowing.

All 191 affected packages/api tests pass + 836 backend api tests pass.
Typecheck clean, lint clean.

* 🧽 refactor: Comprehensive-review polish — hoist tool defs, pin verifyToolAuth contract, doc appConfig

Addresses the comprehensive review of Phase 8. Findings mapped:

**#1 (MINOR): `verifyToolAuth` unconditional auth for execute_code**
- Added doc comment explicitly stating the deployment contract
  (admin capability → reachable sandbox; no per-check health probe
  to keep UI-gate queries O(1)).
- New `api/server/controllers/__tests__/tools.verifyToolAuth.spec.js`
  with 4 regression tests pinning the contract:
  1. `authenticated: true` + `SYSTEM_DEFINED` for execute_code.
  2. 404 for unknown tool IDs.
  3. `loadAuthValues` is never consulted (catches a future revert
     that would resurface the per-user key-entry dialog).
  4. Response `message` is never `USER_PROVIDED`.

**#2 (MINOR): `openai/service.ts` undocumented `appConfig` dependency**
- Expanded the `ChatCompletionDependencies.appConfig` JSDoc to spell
  out that omitting it silently disables code execution for agents
  with `execute_code` in their tools. External consumers of
  `createAgentChatCompletion` now have the contract documented at
  the type boundary.

**#5 (NIT): `registerCodeExecutionTools` re-allocates tool defs**
- Hoisted `READ_FILE_DEF` and `BASH_TOOL_DEF` to module-level
  `Object.freeze`d constants. The shapes derive entirely from
  static `@librechat/agents` exports, so a single frozen object per
  tool is safe to share across every agent init. Eliminates the
  ~4-property allocations on every call (including the common
  second-call no-op path).

**#6 (NIT): Verbose history-priming comment in initialize.js**
- Trimmed the 16-line `handlePrimeInvokedSkills` block to a 5-line
  summary with `@see InitializedAgent.codeEnvAvailable` pointer.
  The canonical narrowing explanation lives on the type; the
  controller comment is just the ACL-vs-capability rationale.

**Skipped:**

- #3 (memory spec tests a mirror function): reviewer self-dismissed
  as a design tradeoff; the enum-literal pin already catches the
  highest-risk drift vector.
- #4 (cross-repo contract for `createCodeExecutionTool`): user will
  explicitly install the latest `@librechat/agents` dev version
  once the companion PR publishes, so the version pin will be
  authoritative.
- #7 (migration/deprecation note for self-hosters): out of scope
  per user direction — release notes handle this.

Tests verified: 679 packages/api + 840 backend api tests pass.
Typecheck + lint clean.

* 🔧 chore: Update @librechat/agents version to 3.1.68-dev.1 across package-lock and package.json files

This commit updates the version of the `@librechat/agents` package from `3.1.68-dev.0` to `3.1.68-dev.1` in the `package-lock.json` and relevant `package.json` files. This change ensures consistency across the project and incorporates any updates or fixes from the new version.
2026-04-25 04:02:01 -04:00

1063 lines
34 KiB
JavaScript

const fs = require('fs');
const path = require('path');
const mime = require('mime');
const { v4 } = require('uuid');
const {
isUUID,
megabyte,
FileContext,
FileSources,
imageExtRegex,
EModelEndpoint,
EToolResources,
mergeFileConfig,
AgentCapabilities,
checkOpenAIStorage,
removeNullishValues,
isAssistantsEndpoint,
getEndpointFileConfig,
documentParserMimeTypes,
} = require('librechat-data-provider');
const { logger } = require('@librechat/data-schemas');
const { sanitizeFilename, parseText, processAudioFile } = require('@librechat/api');
const {
convertImage,
resizeAndConvert,
resizeImageBuffer,
} = require('~/server/services/Files/images');
const { addResourceFileId, deleteResourceFileId } = require('~/server/controllers/assistants/v2');
const { getOpenAIClient } = require('~/server/controllers/assistants/helpers');
const { loadAuthValues } = require('~/server/services/Tools/credentials');
const { getFileStrategy } = require('~/server/utils/getFileStrategy');
const { checkCapability } = require('~/server/services/Config');
const { LB_QueueAsyncCall } = require('~/server/utils/queue');
const { getStrategyFunctions } = require('./strategies');
const { determineFileType } = require('~/server/utils');
const { STTService } = require('./Audio/STTService');
const db = require('~/models');
/**
* Creates a modular file upload wrapper that ensures filename sanitization
* across all storage strategies. This prevents storage-specific implementations
* from having to handle sanitization individually.
*
* @param {Function} uploadFunction - The storage strategy's upload function
* @returns {Function} - Wrapped upload function with sanitization
*/
const createSanitizedUploadWrapper = (uploadFunction) => {
return async (params) => {
const { req, file, file_id, ...restParams } = params;
// Create a modified file object with sanitized original name
// This ensures consistent filename handling across all storage strategies
const sanitizedFile = {
...file,
originalname: sanitizeFilename(file.originalname),
};
return uploadFunction({ req, file: sanitizedFile, file_id, ...restParams });
};
};
/**
* Enqueues the delete operation to the leaky bucket queue if necessary, or adds it directly to promises.
*
* @param {object} params - The passed parameters.
* @param {ServerRequest} params.req - The express request object.
* @param {MongoFile} params.file - The file object to delete.
* @param {Function} params.deleteFile - The delete file function.
* @param {Promise[]} params.promises - The array of promises to await.
* @param {string[]} params.resolvedFileIds - The array of promises to await.
* @param {OpenAI | undefined} [params.openai] - If an OpenAI file, the initialized OpenAI client.
*/
function enqueueDeleteOperation({ req, file, deleteFile, promises, resolvedFileIds, openai }) {
if (checkOpenAIStorage(file.source)) {
// Enqueue to leaky bucket
promises.push(
new Promise((resolve, reject) => {
LB_QueueAsyncCall(
() => deleteFile(req, file, openai),
[],
(err, result) => {
if (err) {
logger.error('Error deleting file from OpenAI source', err);
reject(err);
} else {
resolvedFileIds.push(file.file_id);
resolve(result);
}
},
);
}),
);
} else {
// Add directly to promises
promises.push(
deleteFile(req, file)
.then(() => resolvedFileIds.push(file.file_id))
.catch((err) => {
logger.error('Error deleting file', err);
return Promise.reject(err);
}),
);
}
}
// TODO: refactor as currently only image files can be deleted this way
// as other filetypes will not reside in public path
/**
* Deletes a list of files from the server filesystem and the database.
*
* @param {Object} params - The params object.
* @param {MongoFile[]} params.files - The file objects to delete.
* @param {ServerRequest} params.req - The express request object.
* @param {DeleteFilesBody} params.req.body - The request body.
* @param {string} [params.req.body.agent_id] - The agent ID if file uploaded is associated to an agent.
* @param {string} [params.req.body.assistant_id] - The assistant ID if file uploaded is associated to an assistant.
* @param {string} [params.req.body.tool_resource] - The tool resource if assistant file uploaded is associated to a tool resource.
*
* @returns {Promise<void>}
*/
const processDeleteRequest = async ({ req, files }) => {
const appConfig = req.config;
const resolvedFileIds = [];
const deletionMethods = {};
const promises = [];
/** @type {Record<string, OpenAI | undefined>} */
const client = { [FileSources.openai]: undefined, [FileSources.azure]: undefined };
const initializeClients = async () => {
if (appConfig.endpoints?.[EModelEndpoint.assistants]) {
const openAIClient = await getOpenAIClient({
req,
overrideEndpoint: EModelEndpoint.assistants,
});
client[FileSources.openai] = openAIClient.openai;
}
if (!appConfig.endpoints?.[EModelEndpoint.azureOpenAI]?.assistants) {
return;
}
const azureClient = await getOpenAIClient({
req,
overrideEndpoint: EModelEndpoint.azureAssistants,
});
client[FileSources.azure] = azureClient.openai;
};
if (req.body.assistant_id !== undefined) {
await initializeClients();
}
const agentFiles = [];
for (const file of files) {
const source = file.source ?? FileSources.local;
if (req.body.agent_id && req.body.tool_resource) {
agentFiles.push({
tool_resource: req.body.tool_resource,
file_id: file.file_id,
});
}
if (source === FileSources.text) {
resolvedFileIds.push(file.file_id);
continue;
}
if (checkOpenAIStorage(source) && !client[source]) {
await initializeClients();
}
const openai = client[source];
if (req.body.assistant_id && req.body.tool_resource) {
promises.push(
deleteResourceFileId({
req,
openai,
file_id: file.file_id,
assistant_id: req.body.assistant_id,
tool_resource: req.body.tool_resource,
}),
);
} else if (req.body.assistant_id) {
promises.push(openai.beta.assistants.files.del(req.body.assistant_id, file.file_id));
}
if (deletionMethods[source]) {
enqueueDeleteOperation({
req,
file,
deleteFile: deletionMethods[source],
promises,
resolvedFileIds,
openai,
});
continue;
}
const { deleteFile } = getStrategyFunctions(source);
if (!deleteFile) {
throw new Error(`Delete function not implemented for ${source}`);
}
deletionMethods[source] = deleteFile;
enqueueDeleteOperation({ req, file, deleteFile, promises, resolvedFileIds, openai });
}
if (agentFiles.length > 0) {
promises.push(
db.removeAgentResourceFiles({
agent_id: req.body.agent_id,
files: agentFiles,
}),
);
}
await Promise.allSettled(promises);
await db.deleteFiles(resolvedFileIds);
if (resolvedFileIds.length > 0) {
try {
await db.removeAgentResourceFilesFromAllAgents({ file_ids: resolvedFileIds });
} catch (error) {
logger.error('Error cleaning up orphaned agent file references', error);
}
}
};
/**
* Processes a file URL using a specified file handling strategy. This function accepts a strategy name,
* fetches the corresponding file processing functions (for saving and retrieving file URLs), and then
* executes these functions in sequence. It first saves the file using the provided URL and then retrieves
* the URL of the saved file. If any error occurs during this process, it logs the error and throws an
* exception with an appropriate message.
*
* @param {Object} params - The parameters object.
* @param {FileSources} params.fileStrategy - The file handling strategy to use.
* Must be a value from the `FileSources` enum, which defines different file
* handling strategies (like saving to Firebase, local storage, etc.).
* @param {string} params.userId - The user's unique identifier. Used for creating user-specific paths or
* references in the file handling process.
* @param {string} params.URL - The URL of the file to be processed.
* @param {string} params.fileName - The name that will be used to save the file (including extension)
* @param {string} params.basePath - The base path or directory where the file will be saved or retrieved from.
* @param {FileContext} params.context - The context of the file (e.g., 'avatar', 'image_generation', etc.)
* @returns {Promise<MongoFile>} A promise that resolves to the DB representation (MongoFile)
* of the processed file. It throws an error if the file processing fails at any stage.
*/
const processFileURL = async ({ fileStrategy, userId, URL, fileName, basePath, context }) => {
const { saveURL, getFileURL } = getStrategyFunctions(fileStrategy);
try {
const {
bytes = 0,
type = '',
dimensions = {},
} = (await saveURL({ userId, URL, fileName, basePath })) || {};
const filepath = await getFileURL({ fileName: `${userId}/${fileName}`, basePath });
return await db.createFile(
{
user: userId,
file_id: v4(),
bytes,
filepath,
filename: fileName,
source: fileStrategy,
type,
context,
width: dimensions.width,
height: dimensions.height,
},
true,
);
} catch (error) {
logger.error(`Error while processing the image with ${fileStrategy}:`, error);
throw new Error(`Failed to process the image with ${fileStrategy}. ${error.message}`);
}
};
/**
* Applies the current strategy for image uploads.
* Saves file metadata to the database with an expiry TTL.
*
* @param {Object} params - The parameters object.
* @param {ServerRequest} params.req - The Express request object.
* @param {Express.Response} [params.res] - The Express response object.
* @param {ImageMetadata} params.metadata - Additional metadata for the file.
* @param {boolean} params.returnFile - Whether to return the file metadata or return response as normal.
* @returns {Promise<void>}
*/
const processImageFile = async ({ req, res, metadata, returnFile = false }) => {
const { file } = req;
const appConfig = req.config;
const source = getFileStrategy(appConfig, { isImage: true });
const { handleImageUpload } = getStrategyFunctions(source);
const { file_id, temp_file_id, endpoint } = metadata;
const { filepath, bytes, width, height } = await handleImageUpload({
req,
file,
file_id,
endpoint,
});
const result = await db.createFile(
{
user: req.user.id,
file_id,
temp_file_id,
bytes,
filepath,
filename: file.originalname,
context: FileContext.message_attachment,
source,
type: `image/${appConfig.imageOutputType}`,
width,
height,
},
true,
);
if (returnFile) {
return result;
}
res.status(200).json({ message: 'File uploaded and processed successfully', ...result });
};
/**
* Applies the current strategy for image uploads and
* returns minimal file metadata, without saving to the database.
*
* @param {Object} params - The parameters object.
* @param {ServerRequest} params.req - The Express request object.
* @param {FileContext} params.context - The context of the file (e.g., 'avatar', 'image_generation', etc.)
* @param {boolean} [params.resize=true] - Whether to resize and convert the image to target format. Default is `true`.
* @param {{ buffer: Buffer, width: number, height: number, bytes: number, filename: string, type: string, file_id: string }} [params.metadata] - Required metadata for the file if resize is false.
* @returns {Promise<{ filepath: string, filename: string, source: string, type: string}>}
*/
const uploadImageBuffer = async ({ req, context, metadata = {}, resize = true }) => {
const appConfig = req.config;
const source = getFileStrategy(appConfig, { isImage: true });
const { saveBuffer } = getStrategyFunctions(source);
let { buffer, width, height, bytes, filename, file_id, type } = metadata;
if (resize) {
file_id = v4();
type = `image/${appConfig.imageOutputType}`;
({ buffer, width, height, bytes } = await resizeAndConvert({
inputBuffer: buffer,
desiredFormat: appConfig.imageOutputType,
}));
filename = `${path.basename(req.file.originalname, path.extname(req.file.originalname))}.${
appConfig.imageOutputType
}`;
}
const fileName = `${file_id}-${filename}`;
const filepath = await saveBuffer({ userId: req.user.id, fileName, buffer });
return await db.createFile(
{
user: req.user.id,
file_id,
bytes,
filepath,
filename,
context,
source,
type,
width,
height,
},
true,
);
};
/**
* Applies the current strategy for file uploads.
* Saves file metadata to the database with an expiry TTL.
* Files must be deleted from the server filesystem manually.
*
* @param {Object} params - The parameters object.
* @param {ServerRequest} params.req - The Express request object.
* @param {Express.Response} params.res - The Express response object.
* @param {FileMetadata} params.metadata - Additional metadata for the file.
* @returns {Promise<void>}
*/
const processFileUpload = async ({ req, res, metadata }) => {
const appConfig = req.config;
const isAssistantUpload = isAssistantsEndpoint(metadata.endpoint);
const assistantSource =
metadata.endpoint === EModelEndpoint.azureAssistants ? FileSources.azure : FileSources.openai;
// Use the configured file strategy for regular file uploads (not vectordb)
const source = isAssistantUpload ? assistantSource : appConfig.fileStrategy;
const { handleFileUpload } = getStrategyFunctions(source);
const { file_id, temp_file_id = null } = metadata;
/** @type {OpenAI | undefined} */
let openai;
if (checkOpenAIStorage(source)) {
({ openai } = await getOpenAIClient({ req }));
}
const { file } = req;
const sanitizedUploadFn = createSanitizedUploadWrapper(handleFileUpload);
const {
id,
bytes,
filename,
filepath: _filepath,
embedded,
height,
width,
} = await sanitizedUploadFn({
req,
file,
file_id,
openai,
});
if (isAssistantUpload && !metadata.message_file && !metadata.tool_resource) {
await openai.beta.assistants.files.create(metadata.assistant_id, {
file_id: id,
});
} else if (isAssistantUpload && !metadata.message_file) {
await addResourceFileId({
req,
openai,
file_id: id,
assistant_id: metadata.assistant_id,
tool_resource: metadata.tool_resource,
});
}
let filepath = isAssistantUpload ? `${openai.baseURL}/files/${id}` : _filepath;
if (isAssistantUpload && file.mimetype.startsWith('image')) {
const result = await processImageFile({
req,
file,
metadata: { file_id: v4() },
returnFile: true,
});
filepath = result.filepath;
}
const result = await db.createFile(
{
user: req.user.id,
file_id: id ?? file_id,
temp_file_id,
bytes,
filepath,
filename: filename ?? sanitizeFilename(file.originalname),
context: isAssistantUpload ? FileContext.assistants : FileContext.message_attachment,
model: isAssistantUpload ? req.body.model : undefined,
type: file.mimetype,
embedded,
source,
height,
width,
},
true,
);
res.status(200).json({ message: 'File uploaded and processed successfully', ...result });
};
/**
* Applies the current strategy for file uploads.
* Saves file metadata to the database with an expiry TTL.
* Files must be deleted from the server filesystem manually.
*
* @param {Object} params - The parameters object.
* @param {ServerRequest} params.req - The Express request object.
* @param {Express.Response} params.res - The Express response object.
* @param {FileMetadata} params.metadata - Additional metadata for the file.
* @returns {Promise<void>}
*/
const processAgentFileUpload = async ({ req, res, metadata }) => {
const { file } = req;
const appConfig = req.config;
const { agent_id, tool_resource, file_id, temp_file_id = null } = metadata;
let messageAttachment = !!metadata.message_file;
if (agent_id && !tool_resource && !messageAttachment) {
throw new Error('No tool resource provided for agent file upload');
}
if (tool_resource === EToolResources.file_search && file.mimetype.startsWith('image')) {
throw new Error('Image uploads are not supported for file search tool resources');
}
if (!messageAttachment && !agent_id) {
throw new Error('No agent ID provided for agent file upload');
}
const isImage = file.mimetype.startsWith('image');
let fileInfoMetadata;
const entity_id = messageAttachment === true ? undefined : agent_id;
const basePath = mime.getType(file.originalname)?.startsWith('image') ? 'images' : 'uploads';
if (tool_resource === EToolResources.execute_code) {
const isCodeEnabled = await checkCapability(req, AgentCapabilities.execute_code);
if (!isCodeEnabled) {
throw new Error('Code execution is not enabled for Agents');
}
const { handleFileUpload: uploadCodeEnvFile } = getStrategyFunctions(FileSources.execute_code);
const stream = fs.createReadStream(file.path);
const fileIdentifier = await uploadCodeEnvFile({
req,
stream,
filename: file.originalname,
entity_id,
});
fileInfoMetadata = { fileIdentifier };
} else if (tool_resource === EToolResources.file_search) {
const isFileSearchEnabled = await checkCapability(req, AgentCapabilities.file_search);
if (!isFileSearchEnabled) {
throw new Error('File search is not enabled for Agents');
}
// Note: File search processing continues to dual storage logic below
} else if (tool_resource === EToolResources.context) {
const { file_id, temp_file_id = null } = metadata;
/**
* @param {object} params
* @param {string} params.text
* @param {number} params.bytes
* @param {string} params.filepath
* @param {string} params.type
* @return {Promise<void>}
*/
const createTextFile = async ({ text, bytes, filepath, type = 'text/plain' }) => {
const textBytes = Buffer.byteLength(text, 'utf8');
if (textBytes > 15 * megabyte) {
throw new Error(
`Extracted text from "${file.originalname}" exceeds the 15MB storage limit (${Math.round(textBytes / megabyte)}MB). Try a shorter document.`,
);
}
const fileInfo = removeNullishValues({
text,
bytes,
file_id,
temp_file_id,
user: req.user.id,
type,
filepath: filepath ?? file.path,
source: FileSources.text,
filename: file.originalname,
model: messageAttachment ? undefined : req.body.model,
context: messageAttachment ? FileContext.message_attachment : FileContext.agents,
});
if (!messageAttachment && tool_resource) {
await db.addAgentResourceFile({
file_id,
agent_id,
tool_resource,
updatingUserId: req?.user?.id,
});
}
const result = await db.createFile(fileInfo, true);
return res
.status(200)
.json({ message: 'Agent file uploaded and processed successfully', ...result });
};
const fileConfig = mergeFileConfig(appConfig.fileConfig);
const shouldUseConfiguredOCR =
appConfig?.ocr != null &&
fileConfig.checkType(file.mimetype, fileConfig.ocr?.supportedMimeTypes || []);
const shouldUseDocumentParser =
!shouldUseConfiguredOCR && documentParserMimeTypes.some((regex) => regex.test(file.mimetype));
const shouldUseOCR = shouldUseConfiguredOCR || shouldUseDocumentParser;
const resolveDocumentText = async () => {
if (shouldUseConfiguredOCR) {
try {
const ocrStrategy = appConfig?.ocr?.strategy ?? FileSources.document_parser;
const { handleFileUpload } = getStrategyFunctions(ocrStrategy);
return await handleFileUpload({ req, file, loadAuthValues });
} catch (err) {
logger.error(
`[processAgentFileUpload] Configured OCR failed for "${file.originalname}", falling back to document_parser:`,
err,
);
}
}
try {
const { handleFileUpload } = getStrategyFunctions(FileSources.document_parser);
return await handleFileUpload({ req, file, loadAuthValues });
} catch (err) {
logger.error(
`[processAgentFileUpload] Document parser failed for "${file.originalname}":`,
err,
);
}
};
if (shouldUseConfiguredOCR && !(await checkCapability(req, AgentCapabilities.ocr))) {
throw new Error('OCR capability is not enabled for Agents');
}
if (shouldUseOCR) {
const ocrResult = await resolveDocumentText();
if (ocrResult) {
const { text, bytes, filepath: ocrFileURL } = ocrResult;
return await createTextFile({ text, bytes, filepath: ocrFileURL });
}
throw new Error(
`Unable to extract text from "${file.originalname}". The document may be image-based and requires an OCR service to process.`,
);
}
const shouldUseSTT = fileConfig.checkType(
file.mimetype,
fileConfig.stt?.supportedMimeTypes || [],
);
if (shouldUseSTT) {
const sttService = await STTService.getInstance();
const { text, bytes } = await processAudioFile({ req, file, sttService });
return await createTextFile({ text, bytes });
}
const shouldUseText = fileConfig.checkType(
file.mimetype,
fileConfig.text?.supportedMimeTypes || [],
);
if (!shouldUseText) {
throw new Error(`File type ${file.mimetype} is not supported for text parsing.`);
}
const { text, bytes } = await parseText({ req, file, file_id });
return await createTextFile({ text, bytes, type: file.mimetype });
}
// Dual storage pattern for RAG files: Storage + Vector DB
let storageResult, embeddingResult;
const isImageFile = file.mimetype.startsWith('image');
const source = getFileStrategy(appConfig, { isImage: isImageFile });
if (tool_resource === EToolResources.file_search) {
// FIRST: Upload to Storage for permanent backup (S3/local/etc.)
const { handleFileUpload } = getStrategyFunctions(source);
const sanitizedUploadFn = createSanitizedUploadWrapper(handleFileUpload);
storageResult = await sanitizedUploadFn({
req,
file,
file_id,
basePath,
entity_id,
});
// SECOND: Upload to Vector DB
const { uploadVectors } = require('./VectorDB/crud');
embeddingResult = await uploadVectors({
req,
file,
file_id,
entity_id,
});
// Vector status will be stored at root level, no need for metadata
fileInfoMetadata = {};
} else {
// Standard single storage for non-RAG files
const { handleFileUpload } = getStrategyFunctions(source);
const sanitizedUploadFn = createSanitizedUploadWrapper(handleFileUpload);
storageResult = await sanitizedUploadFn({
req,
file,
file_id,
basePath,
entity_id,
});
}
let { bytes, filename, filepath: _filepath, height, width } = storageResult;
// For RAG files, use embedding result; for others, use storage result
let embedded = storageResult.embedded;
if (tool_resource === EToolResources.file_search) {
embedded = embeddingResult?.embedded;
filename = embeddingResult?.filename || filename;
}
let filepath = _filepath;
if (!messageAttachment && tool_resource) {
await db.addAgentResourceFile({
file_id,
agent_id,
tool_resource,
updatingUserId: req?.user?.id,
});
}
if (isImage) {
const result = await processImageFile({
req,
file,
metadata: { file_id: v4() },
returnFile: true,
});
filepath = result.filepath;
}
const fileInfo = removeNullishValues({
user: req.user.id,
file_id,
temp_file_id,
bytes,
filepath,
filename: filename ?? sanitizeFilename(file.originalname),
context: messageAttachment ? FileContext.message_attachment : FileContext.agents,
model: messageAttachment ? undefined : req.body.model,
metadata: fileInfoMetadata,
type: file.mimetype,
embedded,
source,
height,
width,
});
const result = await db.createFile(fileInfo, true);
res.status(200).json({ message: 'Agent file uploaded and processed successfully', ...result });
};
/**
* @param {object} params - The params object.
* @param {OpenAI} params.openai - The OpenAI client instance.
* @param {string} params.file_id - The ID of the file to retrieve.
* @param {string} params.userId - The user ID.
* @param {string} [params.filename] - The name of the file. `undefined` for `file_citation` annotations.
* @param {boolean} [params.saveFile=false] - Whether to save the file metadata to the database.
* @param {boolean} [params.updateUsage=false] - Whether to update file usage in database.
*/
const processOpenAIFile = async ({
openai,
file_id,
userId,
filename,
saveFile = false,
updateUsage = false,
}) => {
const _file = await openai.files.retrieve(file_id);
const originalName = filename ?? (_file.filename ? path.basename(_file.filename) : undefined);
const filepath = `${openai.baseURL}/files/${userId}/${file_id}${
originalName ? `/${originalName}` : ''
}`;
const type = mime.getType(originalName ?? file_id);
const source =
openai.req.body.endpoint === EModelEndpoint.azureAssistants
? FileSources.azure
: FileSources.openai;
const file = {
..._file,
type,
file_id,
filepath,
usage: 1,
user: userId,
context: _file.purpose,
source,
model: openai.req.body.model,
filename: originalName ?? file_id,
};
if (saveFile) {
await db.createFile(file, true);
} else if (updateUsage) {
try {
await db.updateFileUsage({ file_id });
} catch (error) {
logger.error('Error updating file usage', error);
}
}
return file;
};
/**
* Process OpenAI image files, convert to target format, save and return file metadata.
* @param {object} params - The params object.
* @param {ServerRequest} params.req - The Express request object.
* @param {Buffer} params.buffer - The image buffer.
* @param {string} params.file_id - The file ID.
* @param {string} params.filename - The filename.
* @param {string} params.fileExt - The file extension.
* @returns {Promise<MongoFile>} The file metadata.
*/
const processOpenAIImageOutput = async ({ req, buffer, file_id, filename, fileExt }) => {
const currentDate = new Date();
const formattedDate = currentDate.toISOString();
const appConfig = req.config;
const _file = await convertImage(req, buffer, undefined, `${file_id}${fileExt}`);
// Create only one file record with the correct information
const file = {
..._file,
usage: 1,
user: req.user.id,
type: mime.getType(fileExt),
createdAt: formattedDate,
updatedAt: formattedDate,
source: getFileStrategy(appConfig, { isImage: true }),
context: FileContext.assistants_output,
file_id,
filename,
};
db.createFile(file, true);
return file;
};
/**
* Retrieves and processes an OpenAI file based on its type.
*
* @param {Object} params - The params passed to the function.
* @param {OpenAIClient} params.openai - The OpenAI client instance.
* @param {RunClient} params.client - The LibreChat client instance: either refers to `openai` or `streamRunManager`.
* @param {string} params.file_id - The ID of the file to retrieve.
* @param {string} [params.basename] - The basename of the file (if image); e.g., 'image.jpg'. `undefined` for `file_citation` annotations.
* @param {boolean} [params.unknownType] - Whether the file type is unknown.
* @returns {Promise<{file_id: string, filepath: string, source: string, bytes?: number, width?: number, height?: number} | null>}
* - Returns null if `file_id` is not defined; else, the file metadata if successfully retrieved and processed.
*/
async function retrieveAndProcessFile({
openai,
client,
file_id,
basename: _basename,
unknownType,
}) {
if (!file_id) {
return null;
}
let basename = _basename;
const processArgs = { openai, file_id, filename: basename, userId: client.req.user.id };
// If no basename provided, return only the file metadata
if (!basename) {
return await processOpenAIFile({ ...processArgs, saveFile: true });
}
const fileExt = path.extname(basename);
if (client.attachedFileIds?.has(file_id) || client.processedFileIds?.has(file_id)) {
return processOpenAIFile({ ...processArgs, updateUsage: true });
}
/**
* @returns {Promise<Buffer>} The file data buffer.
*/
const getDataBuffer = async () => {
const response = await openai.files.content(file_id);
const arrayBuffer = await response.arrayBuffer();
return Buffer.from(arrayBuffer);
};
let dataBuffer;
if (unknownType || !fileExt || imageExtRegex.test(basename)) {
try {
dataBuffer = await getDataBuffer();
} catch (error) {
logger.error('Error downloading file from OpenAI:', error);
dataBuffer = null;
}
}
if (!dataBuffer) {
return await processOpenAIFile({ ...processArgs, saveFile: true });
}
// If the filetype is unknown, inspect the file
if (dataBuffer && (unknownType || !fileExt)) {
const detectedExt = await determineFileType(dataBuffer);
const isImageOutput = detectedExt && imageExtRegex.test('.' + detectedExt);
if (!isImageOutput) {
return await processOpenAIFile({ ...processArgs, saveFile: true });
}
return await processOpenAIImageOutput({
file_id,
req: client.req,
buffer: dataBuffer,
filename: basename,
fileExt: detectedExt,
});
} else if (dataBuffer && imageExtRegex.test(basename)) {
return await processOpenAIImageOutput({
file_id,
req: client.req,
buffer: dataBuffer,
filename: basename,
fileExt,
});
} else {
logger.debug(`[retrieveAndProcessFile] Non-image file type detected: ${basename}`);
return await processOpenAIFile({ ...processArgs, saveFile: true });
}
}
/**
* Converts a base64 string to a buffer.
* @param {string} base64String
* @returns {Buffer<ArrayBufferLike>}
*/
function base64ToBuffer(base64String) {
try {
const typeMatch = base64String.match(/^data:([A-Za-z-+/]+);base64,/);
const type = typeMatch ? typeMatch[1] : '';
const base64Data = base64String.replace(/^data:([A-Za-z-+/]+);base64,/, '');
if (!base64Data) {
throw new Error('Invalid base64 string');
}
return {
buffer: Buffer.from(base64Data, 'base64'),
type,
};
} catch (error) {
throw new Error(`Failed to convert base64 to buffer: ${error.message}`);
}
}
async function saveBase64Image(
url,
{ req, file_id: _file_id, filename: _filename, endpoint, context, resolution },
) {
const appConfig = req.config;
const effectiveResolution = resolution ?? appConfig.fileConfig?.imageGeneration ?? 'high';
const file_id = _file_id ?? v4();
let filename = `${file_id}-${_filename}`;
const { buffer: inputBuffer, type } = base64ToBuffer(url);
if (!path.extname(_filename)) {
const extension = mime.getExtension(type);
if (extension) {
filename += `.${extension}`;
} else {
throw new Error(`Could not determine file extension from MIME type: ${type}`);
}
}
const image = await resizeImageBuffer(inputBuffer, effectiveResolution, endpoint);
const source = getFileStrategy(appConfig, { isImage: true });
const { saveBuffer } = getStrategyFunctions(source);
const filepath = await saveBuffer({
userId: req.user.id,
fileName: filename,
buffer: image.buffer,
});
return await db.createFile(
{
type,
source,
context,
file_id,
filepath,
filename,
user: req.user.id,
bytes: image.bytes,
width: image.width,
height: image.height,
},
true,
);
}
/**
* Filters a file based on its size and the endpoint origin.
*
* @param {Object} params - The parameters for the function.
* @param {ServerRequest} params.req - The request object from Express.
* @param {string} [params.req.endpoint]
* @param {string} [params.req.file_id]
* @param {number} [params.req.width]
* @param {number} [params.req.height]
* @param {number} [params.req.version]
* @param {boolean} [params.image] - Whether the file expected is an image.
* @param {boolean} [params.isAvatar] - Whether the file expected is a user or entity avatar.
* @returns {void}
*
* @throws {Error} If a file exception is caught (invalid file size or type, lack of metadata).
*/
function filterFile({ req, image, isAvatar }) {
const { file } = req;
const { endpoint, endpointType, file_id, width, height } = req.body;
if (!file_id && !isAvatar) {
throw new Error('No file_id provided');
}
if (file.size === 0) {
throw new Error('Empty file uploaded');
}
/* parse to validate api call, throws error on fail */
if (!isAvatar) {
isUUID.parse(file_id);
}
if (!endpoint && !isAvatar) {
throw new Error('No endpoint provided');
}
const appConfig = req.config;
const fileConfig = mergeFileConfig(appConfig.fileConfig);
const endpointFileConfig = getEndpointFileConfig({
endpoint,
fileConfig,
endpointType,
});
const fileSizeLimit =
isAvatar === true ? fileConfig.avatarSizeLimit : endpointFileConfig.fileSizeLimit;
if (file.size > fileSizeLimit) {
throw new Error(
`File size limit of ${fileSizeLimit / megabyte} MB exceeded for ${
isAvatar ? 'avatar upload' : `${endpoint} endpoint`
}`,
);
}
const isSupportedMimeType = fileConfig.checkType(
file.mimetype,
endpointFileConfig.supportedMimeTypes,
);
if (!isSupportedMimeType) {
throw new Error('Unsupported file type');
}
if (!image || isAvatar === true) {
return;
}
if (!width) {
throw new Error('No width provided');
}
if (!height) {
throw new Error('No height provided');
}
}
module.exports = {
filterFile,
processFileURL,
saveBase64Image,
processImageFile,
uploadImageBuffer,
processFileUpload,
processDeleteRequest,
processAgentFileUpload,
retrieveAndProcessFile,
};