mirror of
https://github.com/BerriAI/litellm.git
synced 2026-11-04 07:27:51 +00:00
79 commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Kerry Lu
|
6f86b117b5
|
fix(moderations): stop the router image check from 500ing on a list input (#49454)
* fix(moderations): stop the router image check from 500ing on a list input * fix(moderations): give the list cast its cast-ok reason * fix(moderations): type the responses input guard and keep SDK compaction items * fix(moderations): import the compaction item through litellm.types and mark the TypeIs import * fix(moderations): sort the compaction item import --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
paveln83
|
27fbcd9895
|
fix(anthropic): map pause_turn and return refusal explanations on chat and responses (#49192)
* fix(anthropic): map pause_turn and return refusal explanations on chat and responses pause_turn now maps to finish_reason length (Python and Rust) and keeps the raw stop_reason in provider_specific_fields. Responses reports a paused turn as incomplete with max_output_tokens and shows the unrun server search as an in_progress web_search_call instead of a client function_call. A refusal's stop_details.explanation now fills message.refusal and the final stream delta.refusal on chat, and becomes a refusal content part on Responses, streamed or not. * fix(anthropic): let a chat client resend a paused turn so Claude continues it * fix(anthropic): replay a paused chat turn without its thinking blocks, and drop its unrun search before a new message Chat cannot say where Claude wrote each thinking block in a paused web search turn, and Anthropic answers 400 when one moves but accepts the turn without them. A new message after a paused turn also 400s on a server_tool_use with no result, so that unrun search is dropped. * fix(responses): send the streamed refusal part's type and text in content_part.added * fix(anthropic): keep pause_turn and Claude refusal explanations out of the Messages adapter's text A paused Claude turn now ends on stop_reason pause_turn instead of max_tokens, streamed or not. A Claude refusal keeps its explanation in stop_details with Claude's category and no longer repeats it as a text block, so a guardrail mask on a streamed refusal is delivered. A paused turn sent back with list content drops its thinking blocks like string content does * fix(anthropic): replay a paused turn with one unrun search and no thinking on chat and Responses * fix(responses): report a paused Claude turn as completed so Codex and the Agents SDK keep going * fix(responses): mark a paused turn's message item completed * fix(responses): mark every item of a paused turn completed * fix(responses): keep a refusal's explanation when a refused turn is sent back A refusal part in a Responses input message now goes to the model as text, so the assistant turn is no longer dropped from the history. The integration test now expects a paused Claude turn on Responses to report completed, and the paused-turn test helper is typed * fix(responses): scan refusal parts in Responses guardrails --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Vineeth Sai Varikuntla
|
a7b66dd8b9
|
fix(openai): stop stripping cache_control from the caller's own messages and tools (#35602)
* fix(openai): stop stripping cache_control from the caller's messages Copy before removing the field, so the caller's own message objects are left alone. * fix(fireworks_ai): strip cache_control on a copy so a fallback keeps the caller's hints * fix: keep the caller's request intact on more providers and trim extra Claude cache hints under drop_params * fix: drop only the oldest extra cache hints inside one message when trimming to the Claude limit * fix: keep replayed session turns yielding their hints as whole turns while a request's own message keeps its newest hints * fix(cache_control): count only hints that reach Claude and cap hints on OpenAI-compatible Claude routes under drop_params * fix(cache_control): keep the capped Claude messages a list in OpenAI-compatible request bodies * fix(caching): cap Claude hints for string drop_params, a Claude base_model, and Converse tool messages The 4-hint cap now reads drop_params through drops_unsupported_params on chat, Converse, and Anthropic routes, names Claude by base_model on OpenAI-compatible routes, and counts a Converse tool message's hints as the one cache point it sends. * fix(caching): trim Claude cache hints in one wire-order pass and reserve a top-level cache_control on OpenAI-compatible routes Dropping the oldest hint one at a time re-walked the whole request per hint, which took seconds on long tool histories. The cap now drops each message's oldest hints in the order the provider counts them, so a string assistant turn loses its text hint before its tool call hints. OpenAI-compatible Claude routes also keep room for a request-level or extra_body cache_control under drop_params. * test: drop narrating comments and stacked generator clauses from cache_control tests * fix(vertex_ai): put every system message in the Gemini context cache so generateContent never sends it beside cachedContent --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
mika-lt
|
d1d6d55063
|
fix(hosted_vllm): accept the vLLM production-stack router's file and batch objects (#49068)
* fix(files): default a missing file object type to "file" so OpenAI-compatible servers that omit it stop 500ing * fix(batches): accept the vLLM production-stack router's batch object The router leaves out the batch object's "object" field and reports "pending" and "running" statuses, so batch create and retrieve through a hosted_vllm deployment failed validation with a 500. A missing or null object now reads as "batch", pending maps to validating and running to in_progress, and any other unknown value is still rejected * fix(batches): return OpenAI batch objects from batch list behind the vLLM router * fix(batches): type the batch list conversion without a TypeVar and refuse an async client on the sync path * fix(batches): keep batch list items as OpenAI's Batch so an odd item never fails a list * refactor(types): reuse the shared JSON value alias on the batch type * fix: return the provider page for an async client in sync batch list and pass through non-string list statuses * test: inject the SDK's own transport in the sync batch list test so it passes on OpenAI SDK 3.x --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Nate Armstrong
|
87d91d78b7
|
refactor(router): route the router family through RoutingContext and close the outside surface (#49118)
Strategies, pre-call checks, cooldown and fallback helpers, and the auto, complexity and quality sub-routers now take a RoutingContext instead of the Router, and read the router only through registry(), settings(), health() and the context's forwarding methods. Router-free helpers move from router_utils to litellm_core_utils, outside callers of router state switch to entry points, and the surface allowlist shrinks to the endpoint methods plus the entry points |
||
|
Ahmet Selim Fedakar
|
f7f87e6193
|
fix(responses): normalize empty function_call_output to an empty tool result (#48845)
* fix(responses): normalize empty function_call_output to an empty tool result * test(responses): preserve empty tool outputs through guardrails --------- Co-authored-by: Tin <tin@berri.ai> |
||
|
tessellark
|
5d5e3685c1
|
fix(logging): warn when drop_params drops a requested reasoning level on the remaining paths (#48933)
* fix(logging): warn when drop_params drops a requested reasoning level on the remaining paths
* fix(logging): keep guardrail and token-count reads of a request from logging drop warnings
Only the real Anthropic-messages adapter request reports a dropped thinking budget. Guardrail,
shadow-eval and token-count translations of the same request now drop it silently, so one request
logs one warning. The 422 retry also no longer hashes each loc entry, so a list inside loc cannot
raise TypeError.
* fix(bedrock): keep the DeepSeek R1 Converse reasoning_effort drop at debug level
Only allowed_openai_params reaches that branch, and it puts reasoning_effort back after the provider mapping, so a warning there claimed a drop that never happened
* fix(logging): warn on Gemini non-integer budget drops and keep the auto-router baseline silent
Gemini 2.x now logs the drop of a non-integer thinking budget. The auto-router
savings baseline builds a body it never sends, so it no longer warns. Logged
values are capped with reprlib, Azure AI marks only its drop_params retry as
gated, and Bedrock Chat Completions names the model without its route prefix.
* fix(logging): warn on the Gemini 3, Magistral, Ollama, Fireworks and DeepSeek ARN reasoning drops
Gemini 3 and later now log a dropped thinking budget, Magistral logs the
level it swaps for a reasoning prompt, Ollama logs a level it maps to
think off, Fireworks logs chat_template_kwargs effort keys dropped for a
model the map does not list as reasoning, and DeepSeek on Bedrock logs a
drop when the model is addressed by an ARN that advertises the param.
* fix(logging): warn on the remaining Ollama, DeepSeek and Tencent reasoning drops
The ollama/ completion route now shares the chat route's think mapping, so
a level it turns off logs the same warning. DeepSeek warns when it drops
thinking.budget_tokens, and Tencent warns when an adaptive-only model gets a
non-integer budget. Magistral no longer warns on low, medium, high or
enabled thinking, the values its reasoning prompt stands in for. The Gemini 3
budget warning no longer claims the model uses its default level.
* fix(reasoning): warn when Claude drops a non-string reasoning_effort
Claude chat and native /v1/messages dropped a reasoning_effort like 5 or
{"effort": 5} with no log line. Both now log one warn-only WARNING. A
Magistral thinking.type sent as a list no longer raises TypeError.
---------
Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com>
|
||
|
mika-lt
|
f32ac94ad1
|
fix(openai): keep the upstream error body on e.response for OpenAI and Azure errors (#49079)
* fix(openai): keep the upstream error response on mapped exceptions so e.response.text has the wire body * fix: keep the upstream error body on 429 and 5xx errors RateLimitError, InternalServerError, BadGatewayError, and ServiceUnavailableError rebuilt their response and dropped the upstream body. They now keep the body and exactly the upstream headers, and an unread streaming response no longer raises. * fix: keep a gzip-encoded 429 or 5xx error readable The kept body is already decoded, so the rebuilt response drops content-encoding instead of decoding the body a second time and raising DecodingError. * fix: keep the upstream headers and drop credentials on every mapped provider error The copied error response now keeps every upstream header and carries a request with only the method and the URL, so no provider key, query secret, or prompt reaches e.request. Timeout and APIError get the response too, and azure/ chat, streaming, embedding, and image errors attach it * fix: keep the upstream error body on gateway timeouts and re-raised gateway errors * test(azure): cover the 404, 500, 502, and 504 error bodies and the missing-key placeholder * fix: keep no e.response on timeouts and errors litellm raises itself * fix: keep the upstream elapsed time on error responses and the upstream body on azure_text errors * fix: keep the upstream body without credentials on Azure content filter errors and message-matched OpenAI errors --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Alon Sokolovski
|
2b4cb92c8b
|
fix(context_management)!: run clear_tool_uses in the gateway on the Responses API path (#47861)
* fix(context_management): run clear_tool_uses in the gateway on the Responses API path /v1/messages to openai/ models goes through the Responses API, which translated only compact_20260112 (to its native compaction) and dropped every other edit without a warning, so clear_tool_uses_20250919 never cleared anything. Edits without a Responses API form now run in the same polyfill as the chat completions path, before the call, and their applied_edits come back on the response and the final message_delta. compact stays native. * fix(context_management): check the sync polyfill result instead of casting it The strict ruff gate (TID251) bans typing.cast. * test(context_management): cover the Responses path's polyfill error handling A polyfill crash is a 500 before the call, a rejected edit under drop_params is skipped, and the sync stream reports its edits. normalize_spec never raises for a dict spec, so its guard is gone. * fix(context_management): count unreadable blocks and check tool_result text before clearing clear_tool_uses now maps blocks the token counter cannot read (server tool, compaction) to JSON text, so an input_tokens trigger no longer 500s on that history. The Responses API path checks malformed tool_result text on the original messages before clearing, as the chat completions path does. Also removes comments that repeated the code and makes the test fixtures immutable. * fix(context_management): count only blocks the model receives and keep list specs unsent clear_tool_uses now leaves blocks token_counter cannot read, like server tool results and compaction, out of its count instead of counting them as JSON text. No adapter sends them to the model, so an input_tokens trigger no longer fires on tokens the model never sees. On the Responses API path, an OpenAI-style list context_management stays unsent, as it was before, instead of reaching OpenAI's native compaction, where a threshold under 1000 is a 400. On the chat completions path, a clear edit listed before compact_20260112 now runs on the turns after the latest compaction block, as Anthropic does, instead of counting and clearing turns the compaction block drops. The crash test that replaced apply_context_management with a mock is removed. * fix(context_management): apply clear edits in OpenAI count_tokens and skip nested off-spec blocks when counting * fix(context_management): run edits on the history drop_params sends, so count_tokens and /v1/messages clear alike * fix(context_management): count clear_at_least freed tokens over sent blocks only * fix(context_management): count tools with JSON Schema shorthands and reject a non-object trigger or keep A tool whose schema has an array without items, a boolean subschema, or a search_result block without a title made the token counter raise, so a clear edit answered 500. A string trigger or keep also answered 500, even under drop_params. Both now count or get the API's 400, and drop_params drops the malformed field. * fix(token_counter): count a malformed tool schema keyword as absent and move shorthand framed-estimate cases * fix(count_tokens): pass the system prompt into clearing so count_tokens matches inference * fix(context_management): keep a malformed system prompt from failing clear_tool_uses on the Responses API A number, an object, or a list holding non-object items as system made the polyfill answer 500. The polyfill now keeps a list's object items and treats any other shape as no system, on both /v1/messages and count_tokens, so the two paths still count the same text --------- Co-authored-by: AlonSoko <25619863+AlonSoko@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
paveln83
|
4ce724e3df
|
fix(caching): key cached async clients by a per-loop token instead of id(loop) (#48877)
* fix(caching): key cached async clients by a per-loop token instead of id(loop) Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(caching): drop deferred closes queued on a collected event loop Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(caching): pin that a close queued on a live loop waits for that loop * test(caching): run loop-reuse tests until a loop address repeats The collected-loop tests ran a fixed 50 loops and passed with the fix reverted whenever no loop reused an earlier address. They now stop at the first reuse and skip if none happens, through a shared helper. * test(caching): cover the cache key fallback for loops without weak references Also corrects the evicted client closer docstring: a dropped bucket's clients are left to the garbage collector. * fix(http): free short-lived event loops whose handler finalizer close finished as the loop stopped * fix(caching): share sync clients across event loops and catch a failed finalizer close * fix(agentops): stop closing the shared sync http client after the auth call --------- Co-authored-by: mateo <mateo@berri.ai> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
paveln83
|
06cdee91da
|
fix(guardrails): scan and rewrite streamed /v1/completions under post_call guardrails (#49035)
* feat(guardrails): scan and rewrite streamed /v1/completions in post_call pipelines Gives OpenAITextCompletionHandler a process_output_streaming_response that assembles the buffered TextCompletionResponse chunks per choice index at end of stream, runs the guardrail over the assembled text, and writes text rewrites back into the first text-carrying chunk of each choice. Tests and live proof still pending. * fix(guardrails): scan and rewrite streamed /v1/completions under post_call guardrails Assemble each choice's streamed text once the stream ends, run the guardrail on it, write a mask back into the buffered chunks, and end the stream with the guardrail's error on a block. User key metadata now goes through the shared metadata bucket so applied guardrails reach spend logs. * test(guardrails): cover streamed /v1/completions under post_call guardrails * fix(guardrails): keep Gray Swan key metadata, scan a cut-short text stream, and pass non-chunk frames through * fix(guardrails): replace blocked text completion streams like chat, keep Gray Swan's vendor payload, scan text streams in the content filter * fix(guardrails): release an unfinished choice's held text and omit usage the upstream never sent The content filter's streaming hook held back the end of a choice until its finish_reason, which litellm sends only for the first choice of an n>1 stream, so choice 1 lost its last characters on text and chat streams. A closing chunk now releases every unfinished choice when the stream ends. A blocked text stream's closing chunk reported 0/0 usage when the upstream sent none; it now carries usage only when the upstream reported it. * fix(guardrails): send no usage on the content filter's end-of-stream chunk * fix(guardrails): keep the usage chunk last and log the masked text of streamed text completions * fix(guardrails): stream Responses API events through the content filter as they arrive * fix(guardrails): treat a null text completion choice index as choice zero when assembling streams --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Mateo Wang
|
0cf2bb7296
|
fix(openai): warn when the GPT-5 reasoning effort gate drops a level under drop_params (#47761)
* fix(openai): warn when the GPT-5 reasoning effort gate drops a level under drop_params A GPT-5 request whose reasoning_effort the model map does not enable for that model lost the effort with no log line under drop_params. An Azure deployment name missing from the cost map lost reasoning_effort none this way and ran at the model's default effort. The gate now logs a warning naming the model, the level, the map entry it checked, and the model_info.base_model and model_info.reasoning_effort_levels fixes. * fix(openai): warn when the Responses route or a deployment's declared levels drop a reasoning effort The /v1/responses gate dropped a row-disabled GPT-5 effort, and reasoning for an OpenAI model without supports_reasoning, with no log line, and the router logged a declared-level drop only at INFO. Both Responses drops now log a WARNING and the router logs its drop at WARNING. The warning now leads with model_info.reasoning_effort_levels, since base_model only helps a deployment named differently from its underlying model. * fix: warn only when a reasoning level is actually dropped, and on generic reasoning_effort drops * fix: warn when thinking or a Bedrock Converse reasoning level is dropped, and name the router drop a reasoning level * fix: skip the drop warning when a non-reasoning model is asked for no reasoning, and honor declared levels on Bedrock Responses * fix: stay silent when reasoning is turned off on /v1/responses and warn on the Bedrock Responses fallback drop * fix(reasoning): warn when any reasoning-off request is dropped and give each route only the advice it honors A request to turn reasoning off on a model name the catalog lacks now warns on every route, since it may be a reasoning model. Chat suggests model_info.reasoning_effort_levels and, for a name with no catalog entry, model_info.base_model. OpenAI Responses suggests declared levels only where the config honors them, and Bedrock suggests neither. The Bedrock and Mantle Responses forwarding of declared levels is reverted until it can be proven live. * fix(reasoning): suggest reasoning_effort_levels and base_model only where the route reads them The chat gate suggested model_info.reasoning_effort_levels on routes that never pass declared levels (github_copilot, chatgpt and other OpenAIConfig delegates) and model_info.base_model on OpenAI and Azure AI chat, whose gate looks up the deployment name rather than base_model. /v1/responses/compact suggested levels it never reads either. Declared levels are now None when a caller passes none, so the advice appears only when a caller fed them, and the base_model advice is limited to the Azure chat gate, the one that looks up base_model. The openai row of the chat warning test now expects no base_model advice, the Responses warning test goes through get_optional_params_responses_api the way /v1/responses and compact call it, and the Azure direct-config test passes declared levels the way get_optional_params does. Long test lines are wrapped. * fix: read only the built-in catalog when choosing the reasoning drop warning |
||
|
mika-lt
|
71babd931b
|
fix(llms): move later developer messages for models declared system-first (#48897)
* fix(transformation): merge consecutive system and developer messages for non-OpenAI providers and Responses API
* fix(base_utils): keep later cache_control and billing metadata boundaries when merging system messages
* perf(base_utils): fold consecutive system messages in a single linear pass
The tuple accumulator passed to reduce copied the whole prefix on every step,
making map_developer_role_to_system_role O(n^2) in the number of messages.
Message count is not bounded by default, so a caller could tie up a proxy
worker with thousands of small messages before any provider call. Use a list
and merge into its last element in place instead.
* fix(base_utils): hoist later developer messages into the leading system message for non-OpenAI providers
A developer message that arrives after the first user turn (Codex CLI's
per-turn permissions item, a re-injected instruction after compaction, a
mid-conversation policy update) was translated to a system message in
place, and system-first chat templates such as Fireworks' qwen3.8 reject
any system message that is not at the beginning with 400. The fold that
merges consecutive system messages never saw it because a user turn sat
in between.
Move every developer message that follows the leading system block to
the end of that block before the fold, so it merges into the single
leading system message; client-authored mid-conversation system messages
stay where the client put them.
* fix(openai): hoist later developer messages into one leading system message on OpenAI-compatible backends
Developer role translation moves onto the provider config. OpenAI's own endpoint,
Azure OpenAI, and the o-series keep their current behavior. Every other config
inheriting OpenAIGPTConfig (Fireworks, DeepSeek, Together, vLLM, Groq, openai/
with a custom api_base) hoists developer messages that arrive after the leading
system block into that block and folds the block into a single system message,
which is what system-first chat templates require. Providers with their own
message format are untouched.
Messages carrying cache_control or Anthropic billing metadata fold as their own
text blocks, so breakpoints and provider-side stripping survive the merge. The
Responses bridge goes back to prepending instructions as a separate system
message and lets the hoist fold it.
* refactor(base_utils): fold each run of system messages in one pass
The pairwise fold recopied the accumulated system text or block list on every
merge, so a request carrying many consecutive system or developer messages cost
quadratic time in preprocessing. Group consecutive system messages once and
build the merged message from the whole run: plain strings join with a blank
line, and any cached, billing, or block content keeps one text block per
message. The name of the last named message in the run survives, as before.
* fix(base_utils): keep a closing developer message in place and hoist for OpenAIConfig backends
A developer message that closes the conversation stays where the client put it, translated
to system in place, so the request never ends on the assistant's turn. OpenAIConfig and its
subclasses (Azure AI Foundry, GitHub Copilot, ChatGPT, RAGFlow) now share the OpenAI-compatible
hoist, litellm_proxy passes developer messages through to the downstream proxy, and the Anthropic
billing header literals use the shared constant
* fix(base_utils): keep a closing developer message in place only after an assistant turn
A developer message that closes the conversation after a user or tool message hoists like
any other, so the request shape the ticket reports keeps working on system-first templates;
only a closing run that follows an assistant turn stays in place
* refactor(openai): reuse is_openai_backed_api_base for the hosted endpoint check
* test(base_utils): cover an empty system message inside a block-content run
Codecov flagged the empty-string branch of _text_blocks as the one line the
PR left uncovered. The existing empty-string test only hits the plain-text
join; this one forces the block path by merging an empty system message with
a developer message that carries list content.
* fix(openai): hoist developer messages for litellm_proxy and harden the system fold
- drop the LiteLLMProxyChatConfig passthrough so litellm_proxy/ hoists later developer messages like every other OpenAI-compatible endpoint, whichever backend or LiteLLM version sits downstream
- a system message with null or missing content inside a folded run is dropped instead of raising KeyError('content')
- a message-level cache_control on a list-content member lands on its last block so the breakpoint survives the fold
* fix(litellm_proxy): keep the AllMessageValues import the transform signatures still name
* fix(fireworks_ai): hoist developer items after the first turn on the native Responses path
The native Fireworks Responses path landed on staging mapping developer input items to
system in place, so a developer item after the first input item still hit the same
"System message must be at the beginning" 400 on system-first chat templates. The
ordering rules now live in one shared helper over item roles, and the native path
reorders through it before mapping, keeping a closing developer item after an assistant
turn in place exactly like the chat path does
* fix(transformation): carry the developer-message hoist onto main
|
||
|
devin-ai-integration[bot]
|
b93d26bc85
|
fix(responses): drop a malformed prompt_cache_breakpoint under drop_params on the direct /v1/responses path (#45330)
* fix(responses): drop a malformed prompt_cache_breakpoint under drop_params on the direct /v1/responses path * refactor(responses): reuse the shared prompt_cache_breakpoint_for_wire helper in the chat bridge * fix(responses): read a non-flag deployment drop_params as unset in the marker walk, like the param mapper * fix(responses): drop every prompt_cache_breakpoint under drop_params when the model has no breakpoints Also drop a marker set on the message item itself, sanitize before the cache_control conversion so a client typo no longer erases that breakpoint, and skip items that carry no marker so unmarked history costs no extra walk * fix(responses): keep a valid prompt_cache_breakpoint unless the model is known to reject it A deployment alias the cost map does not know (an Azure deployment name, a fine-tune, a litellm_proxy alias) kept its valid markers before this PR and must keep them under drop_params. base_model is read from litellm_params first, then model_info, as chat completions does. * fix(responses): drop a valid marker on known non-breakpoint models and keep the tool schema gate unchanged * fix(responses): judge breakpoint support by model name so cost-map aliases keep valid markers * fix(responses): keep the caller's request stripped of cache_control as before when a marker is copied * fix(responses): keep valid prompt cache markers in place and let cache_control fill a dropped one Under drop_params the marker step now hands back the caller's own objects when every marker is already wire-safe, so a request with valid markers is no longer copied and walked a second time. A deployment that turns cache_control into markers keeps the caller's own markers too, whatever its base_model says, so one request never gets two answers about breakpoint support. On /v1/messages a malformed raw marker next to cache_control is dropped before the conversion, so the block gets the cache hint cache_control alone would give, as on /v1/responses. * test(responses): keep a null marker and the original key order on a model that reads breakpoints --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
rj-cairn
|
5eb1042bbf
|
fix(token_counter): count OpenAI models newer than tiktoken's table with o200k_base (#48728)
* fix(token_counter): count OpenAI models newer than tiktoken's table with o200k_base Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(token_counter): keep tiktoken's encoding for suffixed variants of models it knows Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(token_counter): move the o200k provenance note into the assert message Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(token_counter): strip openai/ and azure/ prefixes before picking the encoding Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(rust_bridge): azure/gpt-4o now resolves to o200k_base Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(token_counter): check the provider prefix through _fix_model_name to keep the type gate flat Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(token_counter): check o200k on the gpt-6 models tiktoken 0.14 still does not know * fix(token_counter): count Azure deployment names and OpenAI alias groups with their OpenAI tokenizer * fix(token_counter): count mixed, alias, and base-model deployments with every served model's tokenizer A group that serves OpenAI and non-OpenAI models now reserves at least each served model's count, model_group_alias names count as their target group, custom Azure deployment names count with their base model, and azure_ai/, chatgpt/ and similar prefixes resolve to the listed OpenAI model. * fix(token_counter): keep mixed groups on OpenAI tokenizers, cache served models, and count context checks, failed rows and stream fallback with the served model * fix(token-counter): keep mixed groups at their old reservation and count new OpenAI release names with o200k A group that serves an OpenAI model next to another vendor now counts its own name the way tiktoken reads it as written, plus its OpenAI models, so it reserves what main did instead of the lower o200k count base_model picks the tokenizer for estimated usage only when it names an OpenAI model, now also in the Cursor stream counter and the output-limit response Unlisted gpt-5+ and o-series names count with o200k_base, and wildcard routes run the context-window check with the matched OpenAI model * fix(token-counter): count an alias to a group no OpenAI model serves by the alias as written Keeps the count main gives such an alias, so its budget reservation and rate limit match main and no other vendor's tokenizer loads for it. Also sorts the import the main merge left unsorted. * test: type the new token-counting tests' parameters * fix(token-counter): count a release-named group no OpenAI model serves as main does, and snapshot the OpenAI model set before iterating it * fix(token-counter): read served models through the router's allowlisted deployments_named, and coordinate the registration race test * perf(token-counter): classify each distinct served model once when walking a model group Budgeted requests walk every deployment in their model group, since the router's cached copy is outside the router surface the proxy may use. Classifying each distinct served model once instead of once per deployment cuts a 200-deployment group from 2.3-4.5 ms to about 1.4 ms with the same result. --------- Co-authored-by: mateo <mateo@berri.ai> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
tessellark
|
dbaae521f9
|
fix(responses): drop the null created_by OpenAI rejects when output items are replayed (#48783)
* fix(responses): drop the null created_by OpenAI rejects when output items are replayed * fix(responses): drop a null created_by from replayed input items too A conversation saved from an older proxy still carries created_by: null on its compaction, apply_patch and shell items, so OpenAI keeps rejecting the replay. Drop the null from input items on the HTTP request transform and the WebSocket response.create rewrite, keeping a created_by that is not null. * fix(responses): drop a null created_by from counted input items too * test(responses): cover saved reasoning nulls and re-compaction of a saved conversation * fix(responses): drop a set created_by from input under drop_params * fix(types): type the Responses output serializer's parameters * fix(responses): drop a set created_by on reasoning items and Perplexity presets under drop_params --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
mika-lt
|
4e5f585c2a
|
fix(guardrails): scan Responses tool results in texts-mode guardrails (#48827)
* fix(guardrails): scan Responses tool results in texts-mode guardrails * fix(guardrails): keep guardrail-added tools and non-text parts on Responses tool scans Tool result texts now come from the same conversion that builds the chat tool message, so items without a call_id, dict outputs, and outputs with no text parts stay aligned for Bedrock and PANW. A changed joined text replaces the first text part and keeps the other parts. scan_only_tool_results appends tools the guardrail returns with new names, like chat, and tool_result and computer_call_output items are scanned through the shared type list. * fix(guardrails): skip empty and screenshot tool outputs and keep text part types on Responses scans * fix(guardrails): keep Bedrock latest-turn masking aligned when tool outputs send no text * fix(guardrails): scan each Responses tool output text part on its own * fix(guardrails): split a Responses tool output's structured row when only textless outputs share its call_id * fix(guardrails): pair each Responses tool output with its own structured row --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
simonp9
|
83ae1bd63e
|
fix(responses): drop params compact rejects and send Azure compaction to the v1 route (#48876)
* fix(responses): drop metadata from compact requests to OpenAI and Azure Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses): drop user and drop_params-rejected fields on compact and route Azure compaction to v1 * fix(responses): keep stream on compact and keep non-dated Azure api-versions on the v1 compact route * fix(responses): drop compact params only for providers verified to reject them --------- Co-authored-by: mateo <mateo@berri.ai> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Kerry Lu
|
e1580b7b83
|
refactor(images): remove dall-e-2 and dall-e-3 code paths (#48703)
OpenAI shut down dall-e-2 and dall-e-3 on 2026-05-12. Drop their image generation and edit configs, the Azure dall-e-2 create-and-poll flow, and the silent dall-e-2 fallback when no model is passed. image_generation without a model or provider now fails locally with "model is required". image_edit and image_variation leave model out of the provider request instead of sending dall-e-2, so providers that ignore model (Topaz) keep working and OpenAI returns its own 400. |
||
|
tin-berri
|
cd01c73dd9
|
fix(responses): preserve typed parts when logprobs is omitted (#48686) | ||
|
tin-berri
|
b9e4298506
|
fix: preserve media across Responses and Gemini conversions (#48484) | ||
|
Mateo Wang
|
14193435b8
|
refactor(anthropic): rename llms/anthropic/pass_through to llms/anthropic/messages (#48510)
* refactor(anthropic): rename llms/anthropic/pass_through to llms/anthropic/messages Move the unified /v1/messages code out of a directory called pass_through, since it is not the passthrough feature. The inner messages sub-package becomes native, and the Vertex AI partner model experimental_pass_through package becomes messages. The old packages stay as import shims that warn once and alias every submodule in sys.modules, so old deep imports still resolve Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(anthropic): keep Final variables out of loops in the rename shim Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(anthropic): match the warning text in the rename shim test Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(anthropic): resolve old pass_through imports lazily through a meta path finder Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(anthropic): mark the alias loader spec restore for the type discipline gate Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * style(anthropic): format the renamed package shim Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * style(anthropic): keep the rebind-ok marker on the spec restore line Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(anthropic): carry the new package search path on alias specs Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(anthropic): declare the shim test locals Final Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
ishaan-j-berri
|
ebe0871a28
|
fix(decisions): preserve omitted probabilities (#48449) (#48524) | ||
|
Mateo Wang
|
7e14638001
|
refactor(llms): rename model-version transformation files to family names (#48509)
gpt_5_transformation.py becomes gpt_reasoning_transformation.py (OpenAI and Azure), the Bedrock anthropic_claude3_transformation.py files become anthropic_claude_transformation.py, anthropic_claude2_transformation.py becomes anthropic_legacy_text_completions_transformation.py, and vertex_ai_partner_models/llama3 becomes vertex_ai_partner_models/llama. Class and symbol names are unchanged. Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
464932b1d1
|
refactor(typing): clear 1181 basedpyright Any errors across 431 files (#47789)
* refactor(typing): narrow Any sources across providers, proxy and guardrails (wave 1) * refactor(typing): narrow Any sources across hubs, providers and proxy (wave 2) * refactor(typing): validate RunwayML task JSON and let dotprompt loaders accept exported prompts * fix: restore pass-through behavior on paths the typing pass tightened MCP client IP extraction trusts X-Forwarded-For only when use_x_forwarded_for is True, dotprompt skips a malformed JSON prompt entry instead of raising, RunwayML task payloads pass through unvalidated, and a malformed provider body raises the same AttributeError as before at the CompactifAI, Gemini, and mock transport parse sites * fix(runwayml): keep a malformed 2xx task body on the 500 path * fix(edenai,cli): keep a non-object body in the error and fail on a malformed migration report EdenAI's embedding and image generation parse dropped a non-object body from the error they raised, so the message no longer named what the provider sent. A helper now raises with the body in the message, and both paths have a regression test. `lite encryption migrate` exited 0 with no message when the proxy's report was not an object or its residual_legacy was not a number, where base crashed. It now fails with a message naming the value. The CompactifAI and Gemini image-edit non-object body tests accept TypeError or AttributeError, which both base and this branch raise, so they guard against a ValidationError instead of pinning one class. RunwayML's async parse gets the same non-object body test as the sync one. * fix: keep base's error classes, duck-typed checks and fail-closed guardrails the typing pass changed * test: annotate the new regression tests' locals with Final * test: annotate the remaining new test locals with Final |
||
|
Mateo Wang
|
a8113b0dd7
|
fix(responses)!: round-trip Anthropic compaction blocks as compaction items (#46754)
* fix(responses): round-trip Anthropic compaction blocks as compaction items /v1/responses on Anthropic routes now returns Claude's compaction block as a compaction output item (and streams it as output_item.added/done), and a compaction input item is replayed to Claude as its native compaction block, so the next turn reuses the summary instead of compacting again. Other routes receive the summary as assistant text. The Anthropic stream handler now keeps the summary text that arrives in compaction_delta. Fixes #41456 * fix(responses): replay compaction blocks natively only when the request compacts * fix(responses): carry compaction through dict-form edits, SDK replay, affinity, and guardrails Read Anthropic's dict-form context_management, accept SDK ResponseCompactionItem objects on replay, strip the encrypted_content affinity wrap before decoding, scan and rewrite compaction summaries in Responses guardrails, drop the history a summary replay replaces, send the summary as a user message so thinking ordering holds, and send the streamed compaction .done with the newest block. * fix(responses): count, route, guard, and redact compaction items consistently A compaction item now reads as its summary everywhere the proxy counts or routes a Responses request: pre-call context checks, the TPM limiter, the input token endpoint, the stream usage estimate, and the previous_response_id session rebuild all drop the history the item replaces. A Claude compaction item sent to a non-Anthropic model (including a fallback) becomes a user summary message instead of an encrypted blob the provider rejects, and Claude on Bedrock Claude Platform takes the summary path. Output guardrails see the summary of a returned compaction item, and logging redaction hides it. * fix(responses): start previous_response_id replay at a stored compaction and count reservations from it * fix(responses): keep a stored compaction summary whole in spend logs * fix(responses): keep stored compaction items summary-only, redact compaction_delta, and finish compaction before reasoning A spend-log row now stores a top-level compaction input item as its summary alone, so extra keys in a client-built item no longer reach the database whole. Message redaction also nulls the streamed compaction_delta, and the streaming bridge sends the compaction item's done event before the reasoning item that follows it. Adds the audit's integration tests for compaction items on the wire. * test(responses): cover stored compaction items, compaction_delta redaction, and compaction done order Adds integration cells for the summary-only spend row of a caller-sent compaction item and its previous_response_id replay, a redacted chat stream row that holds no streamed compaction_delta, and a compaction then thinking stream whose done events follow output order. * fix(spend_logs): honor the redaction header on failure rows for litellm_metadata routes A failed /v1/responses or /v1/messages call keeps its proxy request headers in litellm_metadata, but the failure row read redaction headers from metadata only, so x-litellm-enable-message-redaction never redacted those rows. * fix(spend_logs): lift only the redaction opt-in header into failure rows The failure hook on litellm_metadata routes copied every real request header into the caller's own metadata dict, so the stored request body gained the headers, a caller's metadata.headers was replaced, and a disable header loosened global redaction on failed calls. It now copies only the enable headers into a new litellm_params metadata dict, and only when the caller sent no metadata.headers, so no failure row is less redacted than on main * test(responses): cover failure-row redaction headers and caller metadata on litellm_metadata routes * fix(responses): keep earlier-turn instructions after compaction and count compaction items by their summary A previous_response_id rebuild that starts at a compaction now keeps the system and developer messages of every earlier turn. Token estimates for raw Responses input count a compaction item as its summary text instead of the serialized item. * fix(responses): drop the stored history a caller-sent compaction item replaces on previous_response_id * fix(responses): count only message items with a string role as compaction instructions * test(responses): flatten the last two-for comprehension in the compaction wire test * fix(responses): replay from a compaction in stored Responses output and SDK compaction items A previous_response_id whose stored row is a Responses output holding a compaction item now starts its replay at that compaction, and the handler dumps SDK compaction objects before checking for a summary. Adds tests for the uncovered compaction paths. * fix(responses): keep the stored tool call a function_call_output answers when a compaction item replaces history * fix(responses): keep a stored tool call on every turn after a compaction item and bound the stored summary A previous_response_id rebuild that starts at a turn holding a compaction item now sends again each stored call a replayed tool output answers, so a retry or a later turn no longer loses the turn-1 call. The start is picked only from rows up to the referenced response. A spend row keeps a compaction summary up to the larger of 64K chars and MAX_STRING_LENGTH_PROMPT_IN_DB instead of with no bound. * fix(responses): replay a compacted session only through the referenced response A later sibling turn that sent a compaction item put its summary between a stored tool call and its tool output, so Claude answered 400 on a turn that continued the first of two retried turns, and on a retry sent without the item * fix(spend-logs): keep main's stored value for a prompt length limit under 2 At a MAX_STRING_LENGTH_PROMPT_IN_DB of 0 or 1 the shared truncation helper stored only the truncation marker, where main stores the marker followed by the whole value * fix(spend-logs): keep main's stored tail for a negative prompt length limit At a negative MAX_STRING_LENGTH_PROMPT_IN_DB the shared truncation helper stored one more character of each value than main, since it dropped main's clamp of the end slice to the limit minus the start * fix(responses): keep a compaction item when a guardrail rewrites its summary through structured_messages * test(responses): run the compaction wire tests on the worker database under isolated workers * fix(router): keep the encrypted content affinity pin out of provider requests * fix(proxy): inspect compaction summaries in guardrails, strip a client-sent affinity pin, and price failures from the newest compaction * fix(proxy): count a failed chat or /v1/messages call on its whole history, not from a compaction-typed message * fix(proxy): trim a failed call's input at the newest compaction only when the client called /v1/responses * fix(proxy): trim a failed call's input at the newest compaction when the proxy dispatched it as a Responses call * test(proxy): run the dispatched-call-type failure test through the real request setup * fix(proxy): count a failed Cursor stream's input from the newest compaction item The Cursor route maps to no call type, so the stream generator's failure hook fell back to the route and counted the whole history. Pass the processor's route_type. * fix(proxy): count a broken Cursor stream's prompt from the newest compaction item * fix(responses): keep a tool call a later output answers when trimming before a compaction item * fix(responses): move a compaction-split tool call's turn to right before its output on both compaction routes * perf(responses): look up moved turn items in a set so a long compaction-split turn trims in linear time * test(responses): measure the compaction-split turn move as a CodSpeed benchmark instead of a clock read in unit tests * fix(responses): move only calls still open at the compaction, and bill a disconnected stream from the newest compaction item * fix(responses): restore only the newest stored call when a reused call_id is answered after a compaction item * fix(spend logs): bound a Responses compaction item's summary like a chat block instead of cutting its encrypted_content * test(proxy): price a failed dispatched call through a real ProxyLogging instead of mocked hooks * fix(responses): finish a stream whose Claude compaction failed and group answered call ids by turn once * fix(responses): count input_tokens after a compaction item the way inference sends it |
||
|
Mateo Wang
|
66c8f5b1a9
|
fix(llms): join api_base and route with one shared helper (#47535)
* fix(llms): sanitize trailing slashes in provider get_complete_url * test(llms): move trailing-slash url tests into each provider's mapped unit file * fix(llms): strip trailing slashes in Perplexity search and Ollama embedding URLs and run the Topaz tests in CI * fix(llms): strip trailing slashes in search, DashScope, VolcEngine, GigaChat and Gradient AI URL builders * fix(fastcrw): strip a trailing slash from api_base and expect one slash in the GigaChat tests * fix(gigachat): strip a trailing slash from the pass-through api_base * fix(llms): join api_base and route with one shared helper A deployment api_base that ends in / made several get_complete_url builders emit a double slash, which Envoy and other gateways answer with 404. Add join_api_base in url_utils and use it in those builders, with a regression test that drives every get_complete_url under litellm/llms with a trailing-slash api_base * refactor(llms): move per-provider api_base strips onto join_api_base Use the shared join_api_base helper in the search, chat, embedding, image variation, GigaChat and VolcEngine URL builders instead of a per-provider rstrip and suffix check, so one helper owns the single-slash join * ci: run the trailing-slash URL builder test in the llms shard * test(integration): cover trailing-slash api_base across provider URL builders Wire cells for chat, streaming, messages, responses, embeddings, search and pass-through with and without a trailing slash, an owned-proxy module for the flag-gated providers, env bases and a worker kill, two chaos bursts, and a Topaz SDK cell * fix(llms): send every provider route through join_api_base Every place under litellm/llms and litellm/main.py that joined a caller-settable api_base with a route by concatenation or f-string now goes through join_api_base, so an api_base with a trailing slash sends the same upstream path as one without it Routes made only of a caller-supplied id, model, or passthrough path stay on a plain join over an already normalized base, because the helper's suffix dedup would turn a lookup by id into a list or create call The sweep test now also reaches the factory-made JSON provider builders, and a new loopback test drives the public entrypoints with an api_base ending in /gw/ and checks that every request path keeps /gw/ with no double slash * fix(llms): keep the route skip only where a builder already had it join_api_base is a pure join again. The new ensure_api_route skips the route only when the URL path already ends in it, and it is used only where the merge base or the trailing-slash test PR head already skipped. Evals cancel, Predibase, and Azure AI threads append their route again, EdenAI model listing and Gemini RAG ingestion join through join_api_base, and Azure AI Anthropic checks the path so a host named anthropic-* builds /anthropic/v1/messages * fix(azure_ai): read /anthropic from the URL path when building the count_tokens URL * test(llms): run the Gemini RAG trailing-slash case through aingest The sync ingest wrapper needs a current event loop, which an earlier async test in the same worker clears * test(llms): cover the Azure realtime, Bedrock rerank, and oobabooga joins with a trailing-slash api_base * test(integration): cover api_base joins across providers, query strings and named hosts * feat(url_utils): add join_api_base and ensure_api_route helpers * fix(url_utils): keep the scheme slashes when a stripped suffix was the whole host join_api_base and ensure_api_route now keep the // of a base left with only a scheme, so http://v1 with /v1 stripped joins to http://v1/files as before. The Azure v1 client base_url goes through ensure_api_route so a base ending in /openai/v1 no longer doubles it * fix(azure_ai): build the count_tokens URL from the messages URL * fix(azure_ai): match anthropic as a whole path segment in the messages URL * fix(llms): keep main's Azure AI anthropic match and non-string Bedrock endpoints Azure AI chat and messages go back to cutting the base at the first /anthropic, as on main, and count_tokens keeps its path check, so a gateway path like /anthropic-proxy builds the same URLs as main. A base that already ends in /v1/messages still gets count_tokens next to it. A non-string aws_bedrock_runtime_endpoint (an httpx.URL VPC endpoint) is used as its text again in AgentCore, Bedrock responses and the RAGFlow base, instead of falling back to the default host. get_volcengine_base_url and GigaChat's get_api_base strip a trailing slash again for outside callers. The cache-hit wire test is skipped as a known proxy bug, since one cache hit writes two cache_hit spend rows. * fix(llms): drop a slash before the api_base query and tighten the route test checks * test(llms): annotate new test locals as Final and flatten the builder generator * fix(url_utils): keep a scheme-only api_base hostless so it fails before sending --------- Co-authored-by: Raushankumar0720 <raushan.singh.cg@gmail.com> |
||
|
alberto
|
f768fb32ef
|
fix(streaming)!: send the include_usage chunk with an empty choices array (#47109)
* fix: Send the include_usage chunk with an empty choices array Signed-off-by: Alberto Perdomo <aperdomo@redhat.com> * test: Skip chunks with no choices when reading streams Signed-off-by: Alberto Perdomo <aperdomo@redhat.com> * fix: Strip usage from plain-iterator async stream chunks Signed-off-by: Alberto Perdomo <aperdomo@redhat.com> * test: Expect empty choices on the usage chunk in the gemma tracing test Signed-off-by: Alberto Perdomo <aperdomo@redhat.com> * fix(streaming): send the empty-choices usage chunk only to callers that asked for include_usage The usage chunk now has choices=[] only when the caller set stream_options.include_usage itself. When the OpenAI handler, the proxy's always_include_stream_usage setting, or the harness added include_usage on the caller's behalf, the chunk keeps its legacy single empty-delta choice, so clients that read choices[0] on every chunk keep working. The proxy and harness mark their injection with the private _litellm_stream_usage_injected flag, and the proxy scrubs a client-sent copy. Plain-iterator streams keep usage on their chunk when usage was not requested, and the LLM Shield trailing flush ignores chunks with no choices. * fix(proxy): mark the default stream usage injection so hooks keep the legacy usage chunk shape The proxy's default strip-mode injection adds include_usage for clients that did not ask and drops the usage chunk before it reaches them, but streaming iterator hooks still saw it with choices=[]. Marking that injection with _litellm_stream_usage_injected keeps the old single empty choice for them. * fix(router): treat deployment and router-default stream_options as injected, and drop the client flag on the queue route * refactor(router): move the stream usage injection check into router_utils * fix(router): read stream_options from pydantic deployment params when marking injected usage Update the mantle, openinference, and text-completion integration tests to expect the include_usage chunk with empty choices. * fix(proxy): keep the stream usage injection marker out of stored request bodies * test(streaming): cover who asked for stream usage on chat, queue, deployment and spend log paths * test: type the queue flag test, pin the logging start time, and build the deployment usage config without mutation --------- Signed-off-by: Alberto Perdomo <aperdomo@redhat.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
91724bb0a7
|
fix(responses)!: count input tokens the way inference sends the request (#47363)
* fix(count-tokens): translate chat tool use before Claude token counts Responses and chat tool use (function tools, function_call items, tool messages) is now converted the way Claude chat inference converts it before the Anthropic, Azure AI and Vertex partner counters call count_tokens. Translation errors answer 400 before any provider call. Anthropic-shaped requests and /v1/messages/count_tokens keep their body as sent. * test(count-tokens): compare Claude count bodies with inference for Responses tool use * fix(responses): forward the caller's Responses body to OpenAI token counts OpenAI deployments now get the caller's /v1/responses/input_tokens body as sent, minus LiteLLM-only fields, so OpenAI's own count and errors come back. previous_response_id is checked for ownership like inference and routes to the deployment that served it * fix(responses)!: count Claude Responses bodies the way inference sends them Non-native Responses counts now run the body through the inference bridge, so text.format, tool_choice, reasoning, built-in tools, and file_search count the way inference sends them on Anthropic, Azure AI, Vertex, and Bedrock. The OpenAI native count body uses the drop_params-resolved tool fields. * fix(responses): honor drop_params and replay previous turns in Claude counts Claude input_tokens counts now read the same drop_params setting inference merges from the deployment, router defaults, and request, so a forced tool_choice downgrades to auto under drop_params. previous_response_id replays the stored turns through inference's session loader, and chat messages carrying fields beyond role and content go through the provider's chat translation * fix(responses): keep proxy-only request keys out of OpenAI counts and map chat tools in Bedrock counts * fix(count-tokens): import on Python 3.11 and keep stored prompts in OpenAI counts under drop_params * fix(responses): leave tool_choice out of counts when drop_params drops every tool * fix(count-tokens): count Mantle options, drop cache marks, and reject input inference cannot send The Bedrock Mantle fallback now counts thinking, tool_choice and output_config like the runtime count. Claude Responses counts leave out cache_control marks again, so more than four marks count instead of failing. Counts answer 400 before any provider call for user blocks chat inference refuses and for Responses tools with no Chat Completions equivalent, instead of dropping the block or counting a placeholder tool. Adds the wire tests for these paths * fix(count-tokens): count Responses tools the way the inference bridge sends them A Claude count now hoists additional_tools input items into tools, strips a malformed tool flag under drop_params and rejects it without, and leaves tools out when the deployment lists tools in additional_drop_params, so the count matches the body Claude inference sends for the same request * fix(count-tokens): keep Bedrock runtime counts option-free and send OpenAI the provider file id * test(count-tokens): capture the Anthropic count request at the HTTP layer instead of patching the handler * fix(count-tokens): ignore a caller-sent file id mapping and keep non-string file ids as sent |
||
|
Mateo Wang
|
b51b9b7d43
|
fix(anthropic)!: answer 400 for a malformed tool_result text on /v1/messages and count_tokens (#47347)
* fix(anthropic): answer 400 for a malformed tool_result text on the Responses bridge and count_tokens A tool_result text block whose text is null, a number, or missing made /v1/messages and /v1/messages/count_tokens answer 500 on OpenAI and Azure models. Both routes now answer 400 naming the block, and drop the block when drop_params is on. * fix(anthropic): keep a missing tool_result text as empty text, bound the error, honor body drop_params on count_tokens * fix(anthropic): apply the tool_result text check on the chat/completions bridge and every count route The chat/completions bridge used for Azure, Gemini and other providers now answers the same 400 for a null or numeric tool_result text and drops the block under drop_params, so a model group that mixes OpenAI with another provider answers the same way on every deployment. Gemini and Vertex count_tokens apply the same check, count_tokens turns on drop_params for Claude Code and Codex the way /v1/messages does, and a tool_result without a tool_use_id is checked too * fix(proxy): answer 400 for a malformed tool_result text on count_tokens models without a count API Providers with no token counting API fall back to the local tokenizer, which counted a null or numeric tool_result text as fine while /v1/messages on the same model answers 400. The fallback now runs the same check, unless drop_params applies to the deployment or the request * fix(proxy): honor request drop_params on the local count fallback when no deployment resolves * test(integration): cover non-string tool_result text on both /v1/messages bridges and count routes * fix(anthropic): keep a dropped tool_result on Mistral and reject it before a guardrail rewrite Under drop_params a tool_result whose only text block was null or a number became an empty tool message, and Mistral's empty-message filter removed it, so Mistral answered 400 for a missing tool reply. The filter now only removes empty assistant messages, which is what Mistral rejects. A guardrail that rewrites structured_messages wrote the dropped text back as "", so a null or numeric text was silently blanked even without drop_params. The rewrite path now answers the same 400 as the bridges unless the request or global drop_params is set. * test(integration): cover the Mistral empty tool reply and the guardrail rewrite tool_result text check * fix(anthropic): keep a malformed tool_result text through a guardrail rewrite and keep Mistral's null-message drop A guardrail that rewrites structured_messages now writes the caller's null or numeric tool_result text back in place, so the route's bridge and the deployment's drop_params decide after routing, the same as with no guardrail. The rewrite path no longer answers its own 400 before routing. Mistral drops a null-content system or user message again, as main does, and only a tool reply is exempt from that filter so a dropped tool_result keeps its tool call. * test(mistral): expect empty system and user turns dropped and the empty tool reply kept * perf(anthropic): restore kept tool_result text parts in one linear pass * test(integration): cover the tool_result text restore across tool_results, the chat bridge, and a batch upload behind a rewriting guardrail * fix(anthropic): let guardrails scan malformed tool_result text and honor drop_params on native Anthropic A rewriting guardrail now sees a non-string tool_result text as its JSON text and the write-back restores the caller's value by the sanitized tool_use_id when the guardrail left it unchanged. Native Anthropic drops the bad block under drop_params on /v1/messages and count_tokens, and Mistral keeps an empty tool reply only when it answers a tool call, sent as empty text. * fix(anthropic): compare a kept tool_result text to its scanned form once and align the wire tests with drop_params * test(integration): cover native Anthropic drop_params resolution, guardrail scanning of numeric tool_result text, and the Mistral orphan tool message * fix(count_tokens): honor the router's default drop_params when counting a malformed tool_result text * test(count_tokens): send the tool_result count tests through a real request body instead of a mocked parser * test(count_tokens): send the native Claude count test through a real request body * fix(anthropic): keep each non-string tool_result text when tool_use_ids sanitize to the same id * fix(anthropic): keep non-ASCII characters in the guardrail's view of a non-string tool_result text * test(gemini): mock the count_tokens upstream with respx instead of patching the client factory * fix(mistral): keep an empty tool reply only for an open call, and forward non-object message items on the native drop An empty tool reply re-answering a call the conversation already moved past made Mistral reject the request, and a tool call id that is not a string crashed the filter. Native Anthropic under drop_params assumed every message item was an object. * fix(mistral): find answering tool replies in one pass instead of copying the open call set per reply * fix(mistral): ignore dropped empty placeholders when tracking open tool calls * fix(mistral): drop an empty tool reply when another reply to the same call carries the result * refactor(anthropic): share tool_result text helpers to shrink the diff * fix(anthropic): answer 400 for a non-string tool_result text on the OpenAI Responses bridge |
||
|
Paul El Khoury
|
d7620c77a1
|
fix(openai): drop unsupported reasoning effort for gpt-6.1-sol (#43935)
* fix(openai): drop unsupported reasoning effort for gpt-6.1-sol gpt-6.1-sol accepts low, medium, high, xhigh, and max. none and minimal were still forwarded, because chat completions ignored an explicit supports_none_reasoning_effort=false and the Responses mapper never applied that contract. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(openai): mark the stripped reasoning object as outgoing JSON Co-authored-by: Cursor <cursoragent@cursor.com> * Fix Azure responses reading the OpenAI effort map gpt-6-astra on Azure accepts reasoning.effort none. The responses gate was checking the OpenAI entry, which rejects it. Signed-off-by: Paul El Khoury <82139421+pOpO1-9@users.noreply.github.com> Co-authored-by: Cursor <cursoragent@cursor.com> * fix(openai): gate every reasoning_effort level from the cost map on chat and Responses One gate reads the model's cost-map row for whatever effort level a request asks for, on Chat Completions and the Responses API, for OpenAI and Azure. xhigh stays opt-in on chat (Azure also keeps none opt-in); every other level is refused only when its supports_<level>_reasoning_effort flag is false. A refused level is dropped under drop_params or answered with a local 400, and Responses keeps the rest of the reasoning object when it drops effort. The guards that follow judge temperature by the level the provider will see. Tests move into the mapped test files and derive the expected outcome for each level from the map row, so a new GPT-6 model needs only its row. * test(openai): pin gpt-6.1-sol's refused effort levels and prove the row flags alone drive the gate * fix(cost-map): azure gpt-6-astra refuses none and honors max on Responses, as live calls show * test(reasoning_effort): pin the astra max and none flips through the adapter and Foundry paths * fix(reasoning_effort): keep max out of the route gates and correct the gpt-5.3-codex rows The same model honors max on Responses and refuses it on Chat Completions, so one row flag cannot gate it on either route. The gates now skip max, the Azure gpt-6-astra rows keep max off so the /v1/messages adapter degrades to xhigh and /model_group/info stops short of it as on main, and chat still forwards max for the Responses bridge. The gpt-5.3-codex rows said none and xhigh off and minimal on; the live model answers the other way round (2026-10-01), and an opt-out gate would have dropped the levels Codex CLI sends. * fix(reasoning_effort): bridge function tools to Responses when the chat gate drops none * docs(reasoning_effort): correct the max advertisement note * fix(azure_ai): read the Foundry row on the Responses effort gate and turn none off for Azure gpt-6.1-sol * fix(azure): gate a prefixed Foundry deployment's Responses effort off its row litellm.responses(model="azure_ai/<deployment>", api_base=<Foundry host>) is re-routed to the azure provider with the azure_ai/ prefix intact, so the Azure Responses config saw a prefixed name, skipped it as another provider's pass-through, and forwarded none and minimal to Azure even under drop_params. The Azure gpt-5 resolver also read no row at all for a prefixed name. The Azure Responses gate now detects the family the way the Azure chat gate does, and an azure_ai/ name on the Azure OpenAI configs reads its azure_ai/ row when the map has one, else the azure/ twin, the rule the Foundry chat config already applies to the same deployment. * fix(cost-map): turn none off on azure/eu/gpt-6-astra like its sibling rows * docs(reasoning_effort): say how each gate reads none and shorten the module docstring * fix(azure_ai): call the public effort-level check from the Foundry Responses gate * test(reasoning_effort): fold the Foundry Responses tests into the mapped file and type the row helpers --------- Signed-off-by: Paul El Khoury <82139421+pOpO1-9@users.noreply.github.com> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Co-authored-by: ryan-crabbe-berri <ryan@berri.ai> |
||
|
Mateo Wang
|
75faaeab64
|
feat(count_tokens)!: configurable token_count_framing and token_count_model, count source in the response (#46777)
* feat(count_tokens): configurable token_count_framing and token_count_fallback_model, count source in the response
Count endpoints now say where each number came from in litellm_count_source
(provider, fallback_model, local_estimate) and name the fallback model in
litellm_count_fallback_model. When the provider answers 400 or 404 for a model
it cannot count, the proxy counts the same request with token_count_fallback_model.
token_count_framing adds a per-model constant to local estimates. Both settings
ship in the cost map for Bedrock Claude Sonnet/Haiku 5.5, Bedrock Kimi K3, and
Databricks Kimi K3 and DeepSeek V4 Flash, and can be overridden in model_info.
* test(count_tokens): classify the new cost map keys in the schema generator and expect count_source in exact-shape tests
* test(count_tokens): e2e matrix against real inference and integration wire tests for count source, fallback and framing
* fix(count_tokens): keep the provider error when local counting is disabled, name the fallback error, keep settings in cooldown
Moves the cost map lookup off the event loop
* ci(cost map): bound token_count_framing to non-negative and token_count_fallback_model to non-empty
* fix(count_tokens): keep a failed fallback's own status, ship fallbacks only where they count, add system.ai framing
A fallback model that answers 404 returns the requested model's error unchanged, and any other fallback
failure answers with its own status and message. The eu, apac, au and jp Sonnet and Haiku 5.5 entries
no longer ship the Opus 5.5 fallback, which AWS serves only in us-east-1. A configured empty
token_count_fallback_model turns the shipped fallback off. Config framing accepts whole-number floats and
rejects negatives. The Databricks system.ai Kimi K3 and DeepSeek V4 Flash names carry the same framing as
their serving endpoint names, and the e2e matrix runs on Bedrock Haiku 5.5 too.
* test(databricks): list the system.ai Kimi K3 and DeepSeek V4 Flash names with their published DBU rates
* fix(count_tokens): skip a token count fallback model for an hour after it answers 404 on that deployment
A us. or global. Sonnet 5.5 deployment outside us-east-1 paid two extra Bedrock
calls and about 200 ms on every count, since Opus 5.5 cannot count there either.
The Kimi K3 e2e also runs the large system prompt shape now.
* test(count_tokens): use the shared integration port allocator in the fallback wire test
* fix(cost map): keep the Databricks system.ai entries to framing only, so their spend stays as today
* test(count_tokens): cover the fallback 404 skip, inference profile base_model, shipped framing, and deployment-scoped token count settings
Adds integration cells for a fallback model that answers 404, an application inference profile ARN deployment whose base_model carries the shipped fallback, and the shipped Kimi K3 framing with token_count_framing 0 turning it off. Adds unit tests that keep token count settings off the shared backend cost map key and that store a token count setting on PATCH only when the operator changed it
* fix(count_tokens): keep the counter's own 400 and the provider's error when a framed estimate cannot read the body
A 400 the OpenAI or Gemini counter returns before calling the provider is the caller's answer, so it no longer tries the fallback model or a framed estimate. When the framed local estimate cannot read a content block, the client gets the provider's error instead of a 500. Adds /audit round 2's integration tests for the fallback leg's converted body, the Anthropic envelope on fallback failures, and both framing cases.
* fix(count_tokens): answer the provider's error whenever a framed estimate raises
A tool schema the local estimate cannot format, such as a boolean subschema or an array with no items, raised AttributeError or KeyError and answered a 500 on a framed deployment whose provider had already answered 400. Any estimate failure now answers the provider's own error.
* fix(count_tokens): load the custom tokenizer inside the guarded framed estimate
A framed deployment whose custom_tokenizer could not load answered a 500 or 503 where the provider had already answered 400, because the tokenizer was selected outside the estimate's error handling. Selecting it inside answers the provider's error, logs a warning with the estimate's error, and still raises the Rust extension's fork-guard errors.
* fix(count_tokens): look up framing under the deployment's custom_llm_provider
* fix(count_tokens): share one tokenizer download per worker and bound the wait after a provider error
A framed deployment whose custom_tokenizer is not cached locally made every
count wait out the Hugging Face hub's whole retry cycle when the hub could not
be reached, and a burst of those loads held the default thread limiter that
every count's cost map lookup also needs. Counts on a worker now share one
download per minute, a failed download is reused for the rest of that minute,
and after a provider error a count waits at most 5 s before answering the
provider's error. The swallowed-estimate warning logs only the error type.
* fix(count_tokens): keep one tokenizer download per key running and load the local route per request
* test(count_tokens): type the scripted counter and wait on events instead of wall time where possible
* fix(count_tokens): run framed tokenizer loads on their own capped threads and answer a tokenizer panic with the provider error
Framed loads used anyio's default 40 threads, so a hub that drops packets let about 40 hanging loads stall every count on the worker. Loads now run on a per-worker pool of 4, and a count that finds 4 loads running answers the provider's error at once. Any finished load is reused for a minute after it finishes, so a load that fell back to tiktoken is not retried on every count. A pyo3 PanicException from the tokenizer, while loading or counting, now answers the provider's error instead of a plain 500
* feat(proxy): count tokens with token_count_model in its pinned region
token_count_model {model, region} replaces token_count_fallback_model. Bedrock Sonnet and Haiku 5.5 count with Opus 5.5 in us-east-1, or Melbourne for au. profiles, and a failure names the region to enable
* fix(count_tokens): use the renamed count model constant in the e2e test and list token_count_model with the object keys
* fix(proxy): keep the upstream error envelope when adding the region hint
* test(proxy): annotate the token count setting echo test
* ci(cost map): require non-blank token_count_model strings and test the region rules
* fix(count_tokens): answer the token_count_model's error even when framing is set
* fix(count_tokens): keep the provider error for Google contents instead of a framed estimate
* fix(count_tokens): estimate Google contents the route translated to messages
* fix(count_tokens): keep the provider error for Google contents with non-text parts
* test(proxy): drive tokenizer wait and cost map tests by events and a skippable loop clock
* fix(count_tokens): check Google text-only parts without a KeysView comparison basedpyright flags
* feat(proxy): count Bedrock Kimi K3 tokens exactly with its pinned tokenizer
* test(bedrock): keep the Kimi K3 image check in-process
* fix(count_tokens): join the region hint to an upstream message that already ends with a period
Also pins that a blank token_count_model in config falls through to the cost map default
* fix(count_tokens): count Gemini tool-call turns and Responses tools, and stop tokenizer retries from stalling counts
The Google countTokens route estimates contents with functionCall, functionResponse and thoughtSignature parts again, and still answers the provider error for media, thought and unknown parts
The Kimi K3 exact count renders Responses API and Anthropic tools through the chat tool translation the Vertex count already used, now shared from the prompt factory, and falls back to the estimate when the Converse body cannot carry a tool
A failed Kimi K3 tokenizer load is retried in the background without making counts wait, and a load still running after the wait logs a warning
* fix(count_tokens): count Kimi K3 tools the way Bedrock inference sends them
A flat Responses API function or custom tool becomes the chat tool the
Responses route sends to inference, and every other tool goes to the
Converse body as is, so hosted tools Bedrock drops are left out of the
count instead of being counted or turning it into an estimate.
* test(count_tokens): annotate new token counter test locals with Final
* fix(count_tokens): keep shipped count regions inside the deployment's geography
A plain Bedrock Sonnet or Haiku 5.5 deployment in a non-US region no longer sends its count prompt to us-east-1. Global count models and configured token_count_model keep their pinned region. Also expect litellm_count_source in the Vertex Llama count test main added
* fix(count_tokens): resolve count settings off the event loop
|
||
|
Kerry Lu
|
e8550c1454
|
fix(openai): send chat completions through client.post so the SDK's per-message param transform stops holding the event loop (#46967)
* fix(openai): send chat completions through client.post so the SDK's per-message param transform stops holding the event loop create() walks every message, content part and tool of the request through openai._utils._transform before posting, 91 ms at 300k tokens and ~300 ms at 1M tokens on the event loop, and returns the same dict for the plain dicts litellm builds. client.post is what create() calls after that walk, same request, headers, retries, raw response and stream objects, so the chat helpers now call it directly through the helper embeddings already used * fix(openai): type the chat client.post result as LegacyAPIResponse so the type gate sees .headers and .parse * fix(openai): assert the chat client.post result is a LegacyAPIResponse instead of casting it * test(openai): stub client.post with a LegacyAPIResponse spec in the empty-response and no-leak tests * test(core-utils): stub client.post with a LegacyAPIResponse spec in the agentic loop leak test * test: every client.post stub on the OpenAI chat path returns a LegacyAPIResponse-spec'd mock --------- Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
b39eadcb82
|
feat(openai)!: honor per-deployment workload identity on non-chat routes (#39966)
* feat(anthropic): workload identity federation and pluggable identity sources Backend half of #38818 (internal copy of the fork PR #38013), rebuilt as one commit on top of litellm_internal_staging without the dashboard changes. Deployments on anthropic/ without a static api_key can exchange an OIDC workload assertion for a short-lived sk-ant-oat01 token through a shared RFC 7523 JWT-bearer engine. The assertion comes from a mounted token file, an env token, a LiteLLM-signed issuer, or Keycloak, chosen per deployment, per named credential, or through ANTHROPIC_IDENTITY_SOURCE. The federation fields are server-owned: refused inline in request bodies and on POST /model/new, proxy-admin only on credentials, and the token exchange is pinned to api.anthropic.com unless LITELLM_ANTHROPIC_WIF_ALLOWED_HOSTS adds a host. GET /credentials/{name}/jwks exports the public key set of a LiteLLM-signed credential for the Claude Console. The OpenAI federation trio from #39613 rides along on the backend side with the same server-owned handling. Fixes #28607 Resolves LIT-6107 Co-authored-by: derhornspieler <15236687+derhornspieler@users.noreply.github.com> * fix(anthropic): let batch-result downloads mint from deployment params and accept host:port allowlist entries The files handler enabled workload identity on batch-result downloads but never received the deployment's litellm_params, so a deployment authenticating through a named credential could only mint from process-wide env vars. It now threads litellm_params through to the auth header the way the batch retrieve path already does. LITELLM_ANTHROPIC_WIF_ALLOWED_HOSTS entries written as host:port were read by urlsplit as a scheme, so the allowlist kept the raw entry while the exchange compared bare hostnames and refused the gateway. Entries are now parsed as network locations whether or not they carry a scheme. * fix(types): move the WIF kwargs key sets to a leaf module so the kwargs funnel imports without a cycle * test(anthropic): pin case-insensitive matching of WIF exchange-host allowlist entries * fix(anthropic): end workload identity federation errors without a period so the router suffix reads cleanly * fix(proxy): decrypt stored litellm_params before the WIF write gate * fix(proxy): hide WIF secret references from /health output * fix(proxy): keep the proxy error shape on credential endpoint refusals * fix(proxy): hide identity token file paths from /health output * feat(openai): honor per-deployment workload identity on non-chat routes Image generation, audio speech and transcription, moderation, text completions, files, batches, and fine-tuning built their OpenAI client from the env vars alone, so a keyless deployment carrying openai_identity_provider_id, openai_service_account_id, and openai_identity_token_file failed with the api_key error on every route other than chat, Responses, and embeddings. Those handlers now build their client through the shared workload identity resolver, and the managed batch file-access credential extraction keeps the trio * fix(openai): cache workload identity clients and let deployment identity outrank process-wide keys Non-chat handlers built a fresh SDK client per call, so files, batches, fine-tuning, text completion and sync moderation exchanged the OIDC token on every request. The builders now reuse the cached workload identity client per identity, base, timeout, retries and organization; static-key clients are still built fresh. A process-wide OPENAI_API_KEY (or litellm.api_key / litellm.openai_key) no longer silently overrides a deployment's openai_identity_* params. An explicit distinct api_key on the deployment or call still wins. * fix(openai): keep shared workload identity clients on the process http session The workload identity client is cached and shared across requests, so it must never borrow the transport a single caller passed in. Static-key clients keep using the caller's http client; federated clients always ride the process-wide session. Also makes the http client helpers on BaseOpenAILLM public since the builder now needs them from outside the class. * test(openai): pin partial deployment identity beating the process key * test(openai): type the managed-object invoke callback * fix(anthropic): rename the federation workspace param so Bedrock's anthropic_workspace_id keeps working The Bedrock Claude Platform route already reads anthropic_workspace_id from optional_params, so banning that spelling as a server-owned federation parameter broke a pre-existing client capability. The federation field is now anthropic_federation_workspace_id (env ANTHROPIC_FEDERATION_WORKSPACE_ID), which restores the base branch's behavior for Bedrock callers, drops the Bedrock-specific hint from the refusal message, and deletes the unconditional ban constant that no longer had a reader * fix(auth): share one exchanged token across workers reading the same assertion Anthropic accepts each identity assertion exactly once, so two uvicorn workers reading the same token file both minting from it means the second exchange is denied with jti_reused. Minted tokens now land in a per-user 0700 cache directory guarded by a file lock, so workers on the same host reuse one exchange until the token expires or the assertion rotates. A 401 is only retried when the re-read assertion actually differs, and the denial hint explains jti_reused. LITELLM_TOKEN_EXCHANGE_CACHE_DIR moves the cache and an empty value disables it * feat(openai): honor deployment workload identity on image edits, vector stores, containers, videos, and evals The httpx-based OpenAI handlers built their Authorization header from the deployment api_key or OPENAI_API_KEY alone, so a keyless deployment carrying openai_identity_provider_id, openai_service_account_id, and openai_identity_token_file sent "Bearer None". They now resolve the bearer through resolve_openai_bearer_token, which exchanges the deployment's identity token the same way the SDK-based routes do. A deployment that carries only part of the identity trio now raises a 500 naming the missing field instead of silently falling back to the process key, and the SDK client builders take a factory for the static-key http client so the federated branch never touches litellm.client_session. * fix: keep anthropic federation from being shadowed or leaked An empty or whitespace-only ANTHROPIC_API_KEY counted as set, so a federated deployment sent an empty x-api-key on every call instead of minting a token. Blank values now read as unset, and a real static key on a federated deployment logs once that it outranks federation and nothing is being federated. The exchange-host allowlist matched hostnames only, so a second process on another port of an allowed host was trusted with the workload's identity token. An entry that names a port now trusts that port alone, while a bare host still trusts every port. The shared token store exists so the workers reading one projected token file do not each spend its single-use jti. A source that mints its own assertion per exchange shares nothing with another worker, so it no longer writes a live token to disk for a lookup that can never hit. * feat(openai): resolve per-deployment workload identity for the assistants routes OpenAIAssistantsAPI built its OpenAI and AsyncOpenAI clients from a locals() scrape, so a deployment carrying openai_identity_provider_id, openai_service_account_id and openai_identity_token_file never exchanged a token. Thread litellm_params through the class and build both clients with the shared workload identity builders. * fix: unlink a staged token file a failed write leaves behind The 401 denial hint now also says federation ignores ANTHROPIC_WORKSPACE_ID, which the Bedrock Claude platform provider already reads. * refactor: move anthropic jwks derivation behind a provider-owned tagged union * feat(openai): resolve per-deployment workload identity for image variations Build the image variations client through the shared workload identity builders instead of scraping locals(), and give the async path the async http session it was missing. Let a federation trio serve the model-less batch fallback, and hold the standard LRU size of exchanged tokens. * fix: unlink the staged token file when its write fails at close A buffered write only reaches the disk when the handle closes, so a full disk surfaces at close and left the staging file behind holding a usable token. * fix(anthropic): close the staging descriptor before writing the shared token file * fix(wif): judge federation writes by what they set, not what is stored The admin gate read the stored deployment, so a team admin lost edit, delete and Test Connection on any deployment carrying federation params. It now returns early unless the submitted fields touch the federation surface, and a Test Connection probe that points the deployment at its own api_base is still refused, with the 403 no longer wrapped into a 500 The rest of the same review pass: POST /model/new refuses only a blocking value of `blocked`, so a client that always sends `blocked: false` is not turned away; a request body can no longer pick which federated identity to mint as by naming a stored credential; an advisory refresh the executor refuses disarms the entry instead of wedging the identity until the follower timeout; the static-key shadow warning resolves its env fallback inside the cache instead of once per request; credential writes drop nulls before storing them; the token exchange validates the endpoint URL before reading an assertion and keeps refusing redirects across a client heal; /health hides every server-owned federation field from non-admins; and the async create_file and create_batch paths say which setting is missing when the provider resolves no URL * style(tests): keep two blank lines before the batch_cost_is_final block * test(batches): type the workload identity fallback test fixtures * fix(moderation): stop validating unrelated kwargs on the sync workload identity path * fix(auth): refuse server-owned workload identity keys in array form and inside a client router config * fix(openai): keep a deployment's own key ahead of workload identity on every route Thread the caller's own api_key into the identity resolver for embedding, transcription, moderation, image generation, and assistants so a deployment key equal to the process key still wins. A partial identity trio now falls back to a configured static key, unset max_retries keeps the SDK default, and /v1/decisions plus OpenAI count_tokens resolve the exchanged bearer for identity-only deployments * fix(containers): pass deployment params to container file auth headers Container file routes built their auth headers without the deployment's litellm_params, so a keyless deployment with an OpenAI workload identity sent Bearer None on upload, list, retrieve, delete and content calls * fix(openai): keep process-wide key precedence and move token exchange off the event loop Restore the merge base order where an explicit or process-wide OpenAI key wins over federation, so exporting OPENAI_API_KEY keeps working the way it did before. Run the header-building validate_environment calls (image edit, videos, containers, vector stores, vector store files) and the decisions and System One request preparation in a worker thread, so a slow token exchange no longer stalls the loop. Federate realtime websocket sessions, live sessions, realtime client secrets and transcription sessions, and the OpenAI moderation guardrail when no key is set. * fix(openai): answer 401 for a partial identity trio and federate realtime health checks A deployment with part of the workload identity trio and no static key now fails as an auth error naming the missing params, so the router does not retry it as a server error. The realtime health check sends the exchanged bearer for a keyless federated OpenAI deployment, and API token counting reads the same static keys as every other route before minting. * fix(openai): keep local token counting without an identity trio and federate the moderation guardrail from its own config * fix(openai): skip a partial-trio moderation guardrail and let vault keys win in token counting * test(integration): refuse caller-chosen identity params in user_config and bracketed form keys * fix(openai): send the honored secret manager key from SDK client builders A deployment whose OPENAI_API_KEY lives only in the secret manager now sends that key on image, assistants, files, batches, and other SDK-built routes instead of failing with no api_key, matching what the workload identity resolver already treats as a static key * test(openai): check the token exchange runs off the event loop instead of timing the loop --------- Co-authored-by: derhornspieler <15236687+derhornspieler@users.noreply.github.com> |
||
|
ryan-crabbe-berri
|
9cf09c18fe
|
fix(providers): route Azure Responses past dated api versions, count Grok reasoning tokens on azure_ai, refresh provider sets after register_model, fix gpt-6.1-sol Azure rows (#46377)
* fix(azure): add responses_api_version so the Responses route stops using the deployment chat api_version * fix(azure_ai): fold Grok reasoning tokens into completion_tokens like the xai provider * fix(utils): refresh provider model sets after register_model * fix(cost-map): align azure gpt-6.1-sol reasoning effort flags with the openai row * test(e2e): cover Azure Responses on a deployment pinned to a dated chat api_version * fix(utils): keep router deployment ids out of provider model sets register_model now refreshes the provider sets, so the per-request and router registrations keyed by a deployment id put those opaque ids into wildcard model listings. Those callers now pass add_to_provider_model_sets=False. * fix(azure): carry responses_api_version through the chat to Responses bridge completion() dropped responses_api_version from the litellm_params it hands the Responses bridge, so azure/responses/<model> on a dated chat api_version still built the 404ing /openai/responses URL. Also cover the auto-router marker never forwarding the new param. * fix(azure): stop listing a wrong set of reasoning efforts when none is rejected The message said only low, medium and high are supported, which is wrong for models that also take xhigh such as gpt-6.1-sol. * test(azure): use the latest grok and gpt-5.4 rows in the new azure tests * fix(proxy): list each wildcard-expanded model once A provider set can hold a model both bare and provider-prefixed, which now happens after any priced request registers its openai/<model> key, so the openai/* listing showed that model twice. * fix(azure): route Responses past dated GA api_versions by default A deployment pinned to a dated GA api_version such as 2024-10-21 for chat now sends Responses calls (create, get, delete, cancel, input_items and the azure/responses/ bridge) to the default Responses version, because those dated versions have no Responses route on Azure and 404. responses_api_version still overrides it, -preview, v1, latest and preview versions are kept, an api-version in the api_base query still wins, and chat keeps api_version as before * chore(azure_ai): mark the grok chunk_parser dict as an override contract * fix(azure): rewrite only the route's /openai segment when moving a call to v1 The v1 rewrite in the Azure URL builder replaced every "/openai" substring in the path, so a gateway api_base like https://apim.example.com/openai-gw/team turned into /openai/v1-gw/team/openai/v1/responses. Dated GA Responses calls now land on the v1 route by default, which made this reachable for them too, so the rewrite now only touches the last /openai path segment. The Responses version resolver no longer duplicates the builder's default for an unset api_version. Tests cover the gateway api_base case, the env AZURE_API_BASE query-string precedence, and the e2e suite now has one test for the dated GA default and one where only responses_api_version can make the call work * test(integration): cover Azure Responses api_version, Grok reasoning spend, wildcard listing and reasoning_effort none Wire tests pin the Azure route and query for chat, Responses and the azure/responses bridge on a dated GA api_version, and check responses_api_version never reaches the request body. Cost cases check azure_ai Grok folds reasoning tokens into completion_tokens and spend, streaming and not. A config-driven proxy checks /v1/models next to an openai/* wildcard lists each id once and never a deployment id. Azure gpt-6.1-sol rejects reasoning_effort none before reaching upstream and drops it with drop_params * fix(utils): list ai21 chat and vertex code models under their provider wildcard register_model now files ai21 chat models into ai21_chat_models, which models_by_provider never read, so a registered ai21 chat model dropped out of ai21/* listings. models_by_provider now reads ai21_chat_models and the vertex code chat and code text sets too. * fix(main): keep caller-chosen model names out of provider wildcard listings --------- Co-authored-by: yuneng-jiang <yuneng@berri.ai> |
||
|
Mateo Wang
|
7ddc9f1c96
|
fix(bedrock)!: keep reasoning, thinking blocks and the json_mode finish_reason on fake streams (#46863)
* fix(bedrock): keep reasoning, thinking blocks and the json_mode finish_reason on fake streams A Bedrock Converse model without native structured output fakes the stream when response_format carries a json_schema. The fake stream transformed the response without json_mode and rebuilt a chunk from text, tool_use and token counts only, so it dropped reasoning_content, the signed thinking_blocks and the reasoning token count, and ended on finish_reason tool_calls. The three fake-stream sites now pass json_mode to _transform_response and stream the resulting ModelResponse through the shared convert_model_response_to_streaming converter, which now numbers each streamed thinking block with its index the way real streams do * fix(bedrock): parse text tool calls on Converse fake streams like the non-stream answer * fix(responses): stream the answer text when one chat chunk carries reasoning and the answer * fix(streaming): keep every field on the OpenAI and openai_like fake streams and split multi-block thinking Azure o-series and openai_like fake streams now use the shared MockResponseIterator, so extra tool calls, reasoning and cached and reasoning tokens survive; the Bedrock-local iterator is deleted. The /v1/responses bridge also splits a chunk whose thinking block is signed or redacted with no reasoning text, and the sync iterator closes the reasoning item before the answer opens, like the async one. The /v1/messages bridge gives each thinking block in a chunk its own content block, starts a new block when the thinking block index changes, and streams redacted thinking as redacted_thinking. Snowflake replays drop the stream index from thinking blocks. * fix(anthropic): stream each thinking block's text once on /v1/messages An unsigned thinking chunk now opens its content block with empty text, since its thinking_delta carries the text. A multi-block chunk keeps each piece's thinking block, so an unsigned block followed by a signed one keeps its own text and index. * docs(anthropic): describe which thinking block start carries its text * fix(streaming): close merged reasoning before the answer when one chunk carries both * test(streaming): cover fake-stream reasoning across the proxy bridges and type the Snowflake replay test |
||
|
Mateo Wang
|
b21e659c24
|
fix(proxy)!: count endpoints return errors in their source API shape instead of silent local counts (#46815)
* fix(count_tokens): convert Anthropic-format messages for the OpenAI counter the way inference does * test(count_tokens): check /v1/messages/count_tokens sends OpenAI the input /v1/messages inference sends * fix(count_tokens): build the Anthropic count request without a cast and sync dashboard API types * fix(count_tokens): count Anthropic-format requests for Gemini the way /v1/messages inference converts them * fix(count_tokens): return a 400 naming Anthropic blocks the OpenAI and Gemini count can't convert * test(count_tokens): compare tools by value in the public api test * fix(count_tokens): reject unconvertible blocks nested inside a tool_result * fix(count_tokens): count Anthropic bodies the way /v1/messages inference sends them to OpenAI Unwrap affinity-encoded reasoning, count the structured-output schema, flatten replayed web search results, join a list system prompt, and answer a message without a role with a 400 before calling OpenAI. messages_format is now internal to the Anthropic count route instead of a public /utils/token_counter field. * test(count_tokens): expect the image detail the adapter now sends after merging main * fix(count_tokens): fetch Gemini remote media without blocking the event loop * fix(count_tokens): type normalize_native_messages without the unimported Any On Python 3.10 to 3.13 the annotation is evaluated at import, so the proxy failed to start with NameError * chore(count_tokens): give the Gemini count helpers' list returns a reason * fix(count_tokens): normalize Anthropic messages for Gemini counts the way /v1/messages does * fix(count_tokens): count Gemini bodies after inference's history rewrites and answer converter errors with 400 * test(count_tokens): cover the Gemini count error paths and compaction with a text-block system * fix(proxy): count endpoints return missing-key and request errors in their source API shape * fix(proxy): drop imports the count error change left unused * fix(proxy): let Entra-authenticated Azure AI Claude count tokens and keep Google count errors in the Google envelope * fix(proxy): send Azure environment keys as x-api-key on Azure AI Claude counts and log unexpected Google count failures * fix(proxy): count with the global OpenAI and Anthropic keys chat already uses * fix(proxy): resolve the OpenAI count base with the key so a global key stays on its gateway * fix(proxy): keep unexpected Google count failures generic so internal details stay out of the response * fix(proxy): resolve the Azure AI Claude count base from the same settings chat reads * fix(proxy): count OpenAI tokens with the workload identity bearer chat already uses * fix(proxy): mint the OpenAI workload identity bearer inside the count error handling * fix(proxy): report a failed OpenAI workload identity exchange on the count endpoints as a 401 --------- Co-authored-by: Mrinal Chanshetty <mrinal@berri.ai> |
||
|
Mateo Wang
|
d25b58f744
|
fix(count_tokens): count Anthropic-format requests for OpenAI the way /v1/messages inference converts them (#46759)
* fix(count_tokens): convert Anthropic-format messages for the OpenAI counter the way inference does * test(count_tokens): check /v1/messages/count_tokens sends OpenAI the input /v1/messages inference sends * fix(count_tokens): build the Anthropic count request without a cast and sync dashboard API types * test(count_tokens): compare tools by value in the public api test * fix(count_tokens): count Anthropic bodies the way /v1/messages inference sends them to OpenAI Unwrap affinity-encoded reasoning, count the structured-output schema, flatten replayed web search results, join a list system prompt, and answer a message without a role with a 400 before calling OpenAI. messages_format is now internal to the Anthropic count route instead of a public /utils/token_counter field. * test(count_tokens): expect the image detail the adapter now sends after merging main * fix(count_tokens): type normalize_native_messages without the unimported Any On Python 3.10 to 3.13 the annotation is evaluated at import, so the proxy failed to start with NameError * test(count_tokens): expect the converted Responses input in the budgeted-key count integration test --------- Co-authored-by: Mrinal Chanshetty <mrinal@berri.ai> |
||
|
ryan-crabbe-berri
|
1f041a3265
|
fix(openai): strip prompt_cache_breakpoint under drop_params for OpenAI models that cannot read it (#46453)
A chat completions request to gpt-4.1 (or any OpenAI model older than 5.6) with drop_params on still sent the marker inside messages and tools. It is now removed there, while drop_params off keeps sending it through and models that read it keep it either way Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
2c4ec9c6fa
|
fix(responses): open faked stream items empty so tool arguments and text arrive once (#46533)
* fix(responses): open faked stream items empty so tool arguments and text arrive once * fix(responses): open faked stream items in progress with a fresh empty list, and open mid-stream guardrail blocks empty * fix(responses): open faked refusals, reasoning summaries, custom tool calls and MCP calls empty so each value arrives once * test(integration): cover the empty-open Responses stream items on every fake, replay and block path Integration cells for the faked /v1/responses stream, the chat and messages bridges, cache replay, guardrail block events and the MCP gateway call. Each opened item starts empty and each value arrives once, checked with raw httpx, the OpenAI SDK and the Anthropic SDK, plus upstream errors, 5 KB values, a client drop, a dropped upstream connection and a worker SIGKILL |
||
|
devin-ai-integration[bot]
|
17f621e116
|
fix(vertex_ai): route Model Garden models through the generateContent adapter (#41836)
* fix(vertex_ai): route Model Garden models through the generateContent adapter Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(vertex_ai): annotate generateContent route test parameters Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(vertex_ai): type litellm_params in get_vertex_ai_model_route * fix(vertex_ai): bill and log Model Garden generateContent calls like chat completions Streamed Model Garden calls through the generateContent adapter now request stream usage where the deployment supports it, keep cached-token details, and look up the provider-relative cost key, so they bill like the non-stream call. The adapter forwards router metadata, passes upstream errors through with their status, countTokens falls back to local counting for non-Gemini routes, and the native Gemini config is limited to the Gemini routes. * fix(google_genai): keep adapter stream usage, admin opt-out, single failure count, and base_model routing Streamed generateContent through the chat adapter now ends with the upstream usage, honors always_include_stream_usage false and the caller's stream_options, counts one failure per failed call, and routes Vertex deployments by their base_model like chat does * fix(google_genai): ask for stream usage on every adapter stream so Claude streams end with usage * test(google_genai): expect the trailing usage chunk on bridged OpenAI streams The OpenAI bridge test now sends the usage chunk OpenAI returns for include_usage and expects it translated into usageMetadata. The response_format support helper goes back to its single-purpose form since its only other caller is gone. * fix(google_genai): end adapter streams with one STOP event carrying usage, and keep upstream usage after the finish chunk * fix(google_genai): send the held STOP event when no usage arrives, keep a caller's stream_options, and route countTokens like generateContent * test(integration): cover the moved Vertex routes, adapter stream usage, the shared iterator and usage settings over the wire * test(integration): add cache-hit twins for the stream usage setting and non-stream cells for every iterator member --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
mrinal-berri
|
08a24d391a
|
fix(count_tokens): count system and tools on Vertex Gemini, fall back to Mantle on Bedrock 404 (#46430)
* fix: align token counting with inference context Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: preserve Vertex Gemini count message shapes Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: forward Google countTokens tools in shapes every provider counter accepts Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test: add package marker for Vertex count-token unit tests Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(vertex_ai): ignore deployment system_instruction when counting tokens Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(count_tokens): forward system and tools on Google AI Studio, keep strict on Anthropic tools Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * style(count_tokens): ruff format gemini count handler Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(count_tokens): return Mapping from count request body builders Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(count_tokens): truncate long Anthropic tool names on Vertex count, update stale strict test expectations Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(count_tokens): simplify Gemini count context and tests Put the Vertex Gemini count branch back in the same shape as main and the partner model branch, keeping only the system and tools context, the provider error mapping and the gemini tokenizer default Share one error response helper between the Vertex and AI Studio callers, drop the unread kind fields and the extra validation of map_function output, and say Gemini instead of Vertex AI in the error that AI Studio users also see Tighten the tests by removing a stray parametrize on a helper, sharing the SDK count args, asserting the exact forwarded system and tools, and folding the Anthropic tool shape tests into one table * fix(count_tokens): reuse inference tool translation for Vertex Gemini count Vertex and AI Studio count now convert tools with the same functions inference uses, translate_anthropic_tools_to_openai and the Responses tool converter, before VertexGeminiConfig.map_function. Unnamed and Anthropic hosted tools now count the same declarations inference sends instead of being dropped Tools that cannot be converted, such as a non list tools value, a non object tool or a schema that is not a JSON object, now return a 400 before any request instead of being counted without them --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
606ecf56de
|
fix(decisions): keep box answers from OpenAI-format providers instead of turning them into refusals (#46633)
* fix(decisions): keep box answers from OpenAI-format providers instead of turning them into refusals Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(integration): prove an OpenAI-format box answer reaches the caller as a box Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(integration): give the box wire test a deployment that advertises supports_box_questions Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Wang Yunzhuo
|
a2f792634e
|
fix: preserve explicit provider for async image, audio, and realtime routes (#36346)
* fix(images): preserve explicit provider for async generation * test(images): prove a bare model name reaches its explicit provider on the async path * fix(images): validate async provider hints instead of reading untyped kwargs Read custom_llm_provider and api_base in aimage_generation through a TypeAdapter over a small TypedDict, so the provider lookup gets typed values and the basedpyright gate stays within its merge-base ceiling * fix: answer bad provider hints with a 400 and keep the explicit provider for async speech and transcription * fix: keep the explicit provider for realtime and cover the audio bad-hint 400 * fix: keep the explicit provider on the realtime HTTP routes and read base_url in async audio * fix: stop async image generation from rejecting a base_url it never reads * fix(audio): skip base_url validation when api_base is set on async speech and transcription * test(integration): cover a declared provider with a bare model on image, audio, and realtime routes --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Mateo Wang
|
d6e6c6f26c
|
fix(responses): map Anthropic tool search to OpenAI server tool search (#46441)
* fix(responses): translate OpenAI tool_search to Anthropic tool search for Claude models * fix(responses): define the tool_search_call output item locally for older openai SDKs * fix(responses): gate tool search on the Anthropic request builder and validate deferred tool fields Bedrock Converse builds its own request and has no tool search, so the bridge now drops the tool_search tool there with a warning instead of sending a tool the transport loses, while Bedrock invoke, Anthropic, and Vertex keep it. A replayed tool_search_call with no stored result is dropped from the Anthropic request so a previous_response_id follow-up no longer answers 400. Only srvtoolu_ ids whose name is one of Anthropic's search tools count as search calls, so a caller function named like one stays a function call. A non-boolean defer_loading or a non-list allowed_callers now answers 400 instead of 500 * fix(responses): keep an answered search call on replay and send tools undeferred when no search tool goes out A replayed Claude tool search call that the caller answered with a tool message kept its tool_result in the user turn while its tool_use was dropped from the assistant turn, so Anthropic answered 400 on the OpenAI-SDK tool loop. The search call is now dropped only when no tool message answers it. The Responses bridge now strips defer_loading from every tool whenever no tool search tool is sent, whether the caller sent none or the route cannot run it, since a deferred tool with no search tool can never be loaded. Namespace members had started carrying the flag to every bridge provider, which providers that reject unknown tool keys answered 422. The audit round-1 integration cells for the bridge wire and chaos paths ride in this commit. * fix(responses): gate only the OpenAI tool_search shape and keep native Anthropic search tools intact A native tool_search_tool_regex_20251119 or tool_search_tool_bm25_20251119 tool the caller sends through /v1/responses passes through the bridge on every route as before, and the other tools keep defer_loading whenever any Anthropic search tool goes out. The route gate now drops the OpenAI tool_search item before translation instead of filtering the translated regex tool, so bm25 counts as a search tool. answered_tool_call_ids walks the history with islice instead of a slice copy. The tool_search_call output item drops its created_by field, which OpenAI rejects on a replayed input item and the streaming item never carried * fix(responses): drop malformed tool flags under drop_params and keep stored search calls off non-Anthropic follow-ups With drop_params on, a non-boolean defer_loading or a non-list allowed_callers on a Responses API tool (or a namespace member) is dropped before the bridge instead of answering 400, so only a request without drop_params gets the 400. A custom tool now carries defer_loading to the chat tool like a function tool does A previous_response_id follow-up routed to a provider that does not build an Anthropic request no longer replays the stored Claude search call (srvtoolu_ id) as a client tool call. Mistral answered that history 400 "Unexpected tool call id". An assistant turn that held only the search call is dropped there. The Anthropic route keeps the call, since its factory pairs it with the stored result The tool_search_call output model is renamed OutputToolSearchCall so the proxy's OpenAPI schema keeps the OpenAI SDK's ResponseToolSearchCall name. answered_tool_call_ids reads the history by index, so the lookup is linear in the history length. The streaming done event reads the response it is given, which is the same object it used to read from the iterator * fix(responses): validate function allowed_callers, send one search tool, and keep replays clean off Anthropic A function tool whose allowed_callers is not a list of strings now answers 400 instead of 500. The guardrail scan reads tools without malformed flags, so the route decides between a 400 and a drop under drop_params. When the caller sends a native Anthropic search tool beside tool_search, only the native one goes out. Databricks Claude no longer counts as an Anthropic request route. A previous_response_id follow-up off Anthropic also drops a search call the client replayed as a function_call with a srvtoolu_ id, and its output. The answered tool id scan in the Anthropic message factory runs once per assistant run instead of once per message. A null created_by on tool_search_call and tool_search_output output items is left out of the serialized response, so a cache hit returns the same keys as the first call. * fix(guardrails): keep OpenAI tool_search out of the Responses guardrail tool scan The scan translated tool_search into Anthropic's regex search tool on every route, so a guardrail returning only function tools removed tool_search from native OpenAI requests and OpenAI answered 400 for the deferred tools * fix(anthropic): drop answered tool search replays beside client calls, place system before dropped search turns, flag tool search on azure_ai * fix(anthropic): keep answered tool search replays paired with their answers * fix(anthropic): place mid-conversation system messages in one pass and scan tool_search_call in output guardrails * fix(guardrails): send empty incomplete tool_search_call arguments when a rewrite is not a JSON object * fix(responses): keep the caller's defer_loading on Claude routes when no server tool_search is sent * fix(guardrails): deliver streamed rewrites when OpenAI's tool_search_call has no call_id * fix(responses): answer 400 to a tool_search execution that is not server or client OpenAI's own Responses API answers 400 invalid_value to it. With drop_params the malformed execution is dropped like the other malformed tool flags, which leaves a server search * test(responses): cover the 400 and drop_params paths for a malformed tool_search execution * fix(responses): map Anthropic tool search to OpenAI server tool search Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses): sync generated tool search API types Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses): type tool search mapping predicates Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses): validate tool-search mappings Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses): use the openai SDK ToolSearchToolParam for the mapped tool search * fix(responses): keep deferred chat tools beside a tool search on the responses bridge Restore the hand-rolled OpenAIToolSearchToolParam, since openai 2.20.0 (the supported floor) has no ToolSearchToolParam, and carry defer_loading from chat function tools onto the Responses request when a tool search tool is present, which OpenAI requires for tool_search * fix(responses): leave the bridge's defer_loading carry to its own PR #46439 owns keeping defer_loading on chat function tools sent through the Responses bridge, so this PR keeps only the tool search mapping and the openai 2.20.0 compatible tool type * fix(openai): keep tool search mapping to the OpenAI API and drop it where the model or history cannot take it * fix(openai): narrow the tool search drop to Anthropic's search tool, drop it when nothing is deferred, and run the wire tests against the OpenAI host * fix(openai): pass malformed tools to OpenAI as sent and keep a caller's own tool_search under drop_params * perf(openai): scan Responses history for Claude search calls only when Anthropic's search tool is sent * test(responses): cover the Anthropic tool search mapping to OpenAI Responses across routes, settings, and chaos * test(integration): define the fallback cell's deployments in the gateway config so both workers load them * test(integration): keep the worker-kill burst held past the replacement wait * fix(responses bridge): count an Anthropic search tool as hosted tool search on the chat bridge main's #46439 refuses defer_loading when no tool_search tool is in the request, and the chat bridge only counted {"type": "tool_search"}, so the Anthropic bm25 or regex tool this PR maps to OpenAI's tool search was refused before the mapping ran Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(integration): expect the chat bridge to map the search tool now that defer_loading survives --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Mateo Wang
|
52fc37888b
|
fix(responses bridge): round-trip tool_search_call and tool_search_output items between turns (#46444)
* fix(responses bridge): keep defer_loading on function tools sent to the Responses API Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): give the defer_loading casts a cast-ok reason Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(responses bridge): build defer_loading outside the tool loop so Final stays valid Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): round-trip tool_search_call and tool_search_output items between turns Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * refactor(responses bridge): read replayed tool-search and namespace fields through typed helpers Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): keep tool_search items on a turn with no message or tool call Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): keep every reasoning item on a search-only turn Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): leave content unset on a search-only turn so replay adds no empty message Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): replay tool search items only with tool_search, count replayed tools in the allowlist, keep namespace off guardrails * fix(responses bridge): honor drop_params for defer_loading and route tool_search to Responses Under drop_params, defer_loading is dropped when no tool_search tool is sent, so the 400 it would cause upstream goes away as before. A non-bool value answers 400 without drop_params and is dropped with it. A null tool-level value falls back to the nested function value. A request carrying a tool_search tool now bridges to Responses on OpenAI and Azure endpoints even with reasoning_effort none, since chat completions rejects the tool_search type. * fix(responses bridge): replay tool search items in output order and count replayed tools in tool policy Store the ordered output prefix through the last tool search item, reasoning included, so a second reasoning item no longer makes OpenAI reject the replay. The shared chat-to-Responses converter now replays tool search items and namespace only when the bridge opts in, so Noma, the complexity router and guardrail payloads keep their shape. The tool policy guardrail and the allowlist count replayed tools only when the request can send them. * fix(responses bridge): forward defer_loading only beside a tool_search tool defer_loading only means something with tool_search, and backends without it reject the key (xAI answers 403). A true value with no tool_search tool now answers 400 naming drop_params and is dropped under drop_params; a false value there is a no-op and is dropped. * fix(tool_loop): keep tool search items and namespace between rounds * fix(responses bridge): check replayed tool search loads when a deployment adds tool_search * fix(chat): drop litellm response fields before strict OpenAI-compatible providers A chat reply from the Responses bridge carries provider_specific_fields on the message and on each tool call, plus reasoning_items. Groq, Fireworks, Mistral and Together reject unknown fields, so sending that reply back to them as history failed with a 400 or 422. Strip them in the shared helper and use it in those providers. * fix(caching): keep reasoning_items and tool search items on response cache hits A cached bridge response lost reasoning_items on a non-streaming hit and lost both reasoning_items and provider_specific_fields (tool_search_items) on a streamed hit, so the next turn could not replay the search or the reasoning around it * test(integration): cover tool search item replay on the chat to Responses bridge Proxy, SDK, authorization and chaos cells for returning and replaying tool_search_call and tool_search_output items and the tool call namespace * test(integration): add response cache twins for the tool search guardrail cells Each guardrail cell now has a twin where the answer is already in the Redis response cache: a key opted out of global guardrails caches a replay that loads a policy-blocked tool, the same replay without guardrails is cached before block mode sees it, rewrite mode is not served the unrewritten conversation's cached answer, and an allowed replay is served from the cache. * fix(responses bridge): forward defer_loading beside a client-executed tool_search The bridge now replays a client tool_search_call and the client's tool_search_output through provider_specific_fields.tool_search_items, so a deferred function beside a client tool_search keeps defer_loading like it does beside a hosted one * test(responses bridge): pin the client tool_search turn and its replay, streamed and not * fix(responses bridge): strip nested namespace for strict providers and keep streamed reasoning_items on text replies Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix(responses bridge): keep dropping defer_loading beside a client-executed tool_search Forwarding defer_loading beside a client tool_search made GPT stop with a search call the chat client has to answer by hand, so a plain question that used to get tool calls got none. Keep main's behavior of dropping it there. --------- Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
devin-ai-integration[bot]
|
314f17986a
|
fix(responses): translate OpenAI tool_search to Anthropic tool search for Claude models (#45882)
* fix(responses): translate OpenAI tool_search to Anthropic tool search for Claude models * fix(responses): define the tool_search_call output item locally for older openai SDKs * fix(responses): gate tool search on the Anthropic request builder and validate deferred tool fields Bedrock Converse builds its own request and has no tool search, so the bridge now drops the tool_search tool there with a warning instead of sending a tool the transport loses, while Bedrock invoke, Anthropic, and Vertex keep it. A replayed tool_search_call with no stored result is dropped from the Anthropic request so a previous_response_id follow-up no longer answers 400. Only srvtoolu_ ids whose name is one of Anthropic's search tools count as search calls, so a caller function named like one stays a function call. A non-boolean defer_loading or a non-list allowed_callers now answers 400 instead of 500 * fix(responses): keep an answered search call on replay and send tools undeferred when no search tool goes out A replayed Claude tool search call that the caller answered with a tool message kept its tool_result in the user turn while its tool_use was dropped from the assistant turn, so Anthropic answered 400 on the OpenAI-SDK tool loop. The search call is now dropped only when no tool message answers it. The Responses bridge now strips defer_loading from every tool whenever no tool search tool is sent, whether the caller sent none or the route cannot run it, since a deferred tool with no search tool can never be loaded. Namespace members had started carrying the flag to every bridge provider, which providers that reject unknown tool keys answered 422. The audit round-1 integration cells for the bridge wire and chaos paths ride in this commit. * fix(responses): gate only the OpenAI tool_search shape and keep native Anthropic search tools intact A native tool_search_tool_regex_20251119 or tool_search_tool_bm25_20251119 tool the caller sends through /v1/responses passes through the bridge on every route as before, and the other tools keep defer_loading whenever any Anthropic search tool goes out. The route gate now drops the OpenAI tool_search item before translation instead of filtering the translated regex tool, so bm25 counts as a search tool. answered_tool_call_ids walks the history with islice instead of a slice copy. The tool_search_call output item drops its created_by field, which OpenAI rejects on a replayed input item and the streaming item never carried * fix(responses): drop malformed tool flags under drop_params and keep stored search calls off non-Anthropic follow-ups With drop_params on, a non-boolean defer_loading or a non-list allowed_callers on a Responses API tool (or a namespace member) is dropped before the bridge instead of answering 400, so only a request without drop_params gets the 400. A custom tool now carries defer_loading to the chat tool like a function tool does A previous_response_id follow-up routed to a provider that does not build an Anthropic request no longer replays the stored Claude search call (srvtoolu_ id) as a client tool call. Mistral answered that history 400 "Unexpected tool call id". An assistant turn that held only the search call is dropped there. The Anthropic route keeps the call, since its factory pairs it with the stored result The tool_search_call output model is renamed OutputToolSearchCall so the proxy's OpenAPI schema keeps the OpenAI SDK's ResponseToolSearchCall name. answered_tool_call_ids reads the history by index, so the lookup is linear in the history length. The streaming done event reads the response it is given, which is the same object it used to read from the iterator * fix(responses): validate function allowed_callers, send one search tool, and keep replays clean off Anthropic A function tool whose allowed_callers is not a list of strings now answers 400 instead of 500. The guardrail scan reads tools without malformed flags, so the route decides between a 400 and a drop under drop_params. When the caller sends a native Anthropic search tool beside tool_search, only the native one goes out. Databricks Claude no longer counts as an Anthropic request route. A previous_response_id follow-up off Anthropic also drops a search call the client replayed as a function_call with a srvtoolu_ id, and its output. The answered tool id scan in the Anthropic message factory runs once per assistant run instead of once per message. A null created_by on tool_search_call and tool_search_output output items is left out of the serialized response, so a cache hit returns the same keys as the first call. * fix(guardrails): keep OpenAI tool_search out of the Responses guardrail tool scan The scan translated tool_search into Anthropic's regex search tool on every route, so a guardrail returning only function tools removed tool_search from native OpenAI requests and OpenAI answered 400 for the deferred tools * fix(anthropic): drop answered tool search replays beside client calls, place system before dropped search turns, flag tool search on azure_ai * fix(anthropic): keep answered tool search replays paired with their answers * fix(anthropic): place mid-conversation system messages in one pass and scan tool_search_call in output guardrails * fix(guardrails): send empty incomplete tool_search_call arguments when a rewrite is not a JSON object * fix(responses): keep the caller's defer_loading on Claude routes when no server tool_search is sent * fix(guardrails): deliver streamed rewrites when OpenAI's tool_search_call has no call_id * fix(responses): answer 400 to a tool_search execution that is not server or client OpenAI's own Responses API answers 400 invalid_value to it. With drop_params the malformed execution is dropped like the other malformed tool flags, which leaves a server search * test(responses): cover the 400 and drop_params paths for a malformed tool_search execution * test(responses): cover a Claude search call replayed to Bedrock Converse with cache marks --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Co-authored-by: mateo <mateo@berri.ai> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
Tobias Mayr
|
d03923e6a9
|
fix(responses): propagate cache_control through responses->chat completion bridge (#32511)
* fix(responses): propagate cache_control through responses->chat completion bridge User/assistant messages, tool call outputs, and function calls built from Responses API input items were dropping cache_control, breaking prompt caching for Bedrock/Anthropic when routed through the litellm_completion transform. Also propagate cache_control onto the last content element of user/assistant messages, since Anthropic/Bedrock read it from the content block rather than (or in addition to) the message level. Adds the missing cache_control field to ChatCompletionToolMessage and ChatCompletionResponseMessage TypedDicts to keep the assignments type-safe. * chore(ui): regenerate schema.d.ts for ChatCompletionToolMessage.cache_control Adding cache_control to the ChatCompletionToolMessage TypedDict changed the proxy's OpenAPI spec; regenerated via `npm run gen:api`. * fix(responses): don't overwrite existing content-block cache_control; add list-content test Per Greptile review on the PR: - The message-level cache_control was unconditionally overwriting the last content block's cache_control when that block already carried its own (preserved by _transform_responses_api_content_to_chat_completion_content). Now only propagates onto the last block when it doesn't already have one. - Added the missing test coverage for the content-block propagation path with list content, which no existing test exercised (all used plain string content, so the isinstance(content, list) branch was never hit). * style(responses): extract cache_control into a named variable, trim comment Per style review: the message-building branch read cache_control 3 different ways in 5 lines (.get() twice, then [] indexing once), unlike the sibling tool-call-output/function-call functions this PR touches, which each extract it into a named variable up front (output_cache_control, function_call_cache_control). Matched that convention here too. Also trimmed the 3-line inline comment to one line, matching the file's usual comment length for similarly subtle behavior. * chore: refresh lazy OpenAPI snapshot * fix: type cache control propagation * fix: place cache control type suppressions * style: format cache control type conversion * refactor(responses): build bridged cache_control messages in one shot * fix(responses): keep cache_control marks on bridged tool calls and within the provider cap A function_call item's mark now rides on the tool call the bridge builds, an assistant text item merged behind a function_call keeps its own mark, a developer item keeps its mark when it is mapped to a system message, and a previous_response_id replay yields its stale marks to the newest ones so the provider cap holds. The Responses layer counts the marks on function_call and function_call_output items before it places an ordinal injection point, and the chat pass replaces that count with its own instead of adding to it. * fix(responses): count each cache_control mark once and keep the bridge within the provider's limits List content carries an item-level cache_control on its last block only, so the provider sees one cache point and the breakpoint census counts one. Marked function tools and SDK-typed input items are handled by the Responses-layer reserve, assistant text merged into a function call turn keeps its mark on its own block, and the previous_response_id replay budget drops marks per turn, newest kept, instead of stripping every older turn once one overshoots. The request's own marks now fall under the same budget, newest kept, so five marked items no longer answer the provider's 400. A cache_control value that is not an ephemeral mark with an optional ttl is left off the bridged message instead of reaching the provider as sent, and a developer message maps to the system role with its content, name, and cache_control only, so unknown keys never reach a strict provider. * fix(responses): send item cache_control only to cache-point targets and follow drop_params Item marks reach the provider only when the target turns them into cache points, and every other target keeps main's behavior of stripping them. Without drop_params a mark the provider would refuse reaches it as sent, so the client gets the same 400 as on /v1/chat/completions. With drop_params a malformed mark is dropped and only the newest 4 are kept. The shared input-to-messages helper and the developer mapping in base_utils go back to main, and a marked developer item reaches a cache-point target as a system message in the bridge. * fix(responses): budget top-level marks, follow drop_params on Nova tool marks, and match chat-keeping providers * fix(responses): validate replayed marks and count input tokens without item marks * fix(responses): leave unknown Bedrock ARNs unmarked under drop_params, match OpenRouter's list, and reserve only forwarded marks * test(responses): cover item cache_control parity per chat route, drop_params trimming, and replay budgets * fix(responses): keep item marks off inspection paths and match each chat route's keep rule Only the request the bridge sends gets item marks now. Guardrails, token counting, rate limiting, and input_tokens read the input with no item marks, as on main, so an off-spec mark cannot fail a guardrail call. Tool message and tool call marks are typed inside the bridge, leaving the shared chat types and guardrail schemas as on main. Databricks keeps marks only for Claude, Snowflake uses its chat route's Claude check, and every Anthropic deployment keeps them whatever its model name. The injection census skips SDK-typed tools, and the owned-proxy test expects the marked messages on /v1/responses. * test(responses): cover guardrail reads without item marks, Snowflake name parity, and SDK-typed tools in the census * fix(responses): leave a non-list input untouched when stripping item marks * test(responses): cover string, list, object and tuple input with and without item marks kept Each input shape now answers the same whether the unmapped Bedrock ARN keeps item marks or strips them under drop_params, through the proxy and the SDK. Object input also runs through the OpenAI-compatible chat bridge, /v1/responses/input_tokens and the project input-token limiter * test(cache): derive owned-proxy cell timeouts from the harness boot budget The capped-burst cells and the uncapped response-cache cells boot their own two-worker proxy, and a fixed 240 or 360 second timeout cut them during boot on a slow runner before the harness readiness budget could decide. They now carry the same derived budget as the sibling chaos and guardrail files, with four boot and stop budgets for the restart cell that boots twice --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
devin-ai-integration[bot]
|
2f3ccbe7c7
|
feat(decisions)!: send images to Cloudflare Clef through its images field (#45734)
* feat(decisions): send images to Cloudflare Clef OpenAI-format input_image parts and the System One images field now reach Cloudflare clef and clef-flash in its images field, with remote URLs downloaded and inlined first. Providers that cannot take images drop them under drop_params and answer 400 otherwise. * fix(decisions): follow drop_params for images and use the call timeout for image downloads * fix(decisions): drop a malformed images entry on its own under drop_params * fix(decisions): refuse images next to input whatever drop_params says * fix(decisions): bound clef image fetches, honor drop_params for images, and log caller URLs Refuse more than 4 images for Cloudflare clef before any download, keep the caller's timeout as one deadline across image fetch retries, accept images next to input under drop_params, validate image strings and objects, and pass the caller's image URLs (not inlined bytes) to logging callbacks. * fix(decisions): check input_image URLs and clef image sizes * fix(decisions): accept data URL media type parameters and check clef image types * fix(decisions): count clef images after dropping unsupported types and flag clef vision --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
ryan-crabbe-berri
|
365c4885de
|
fix(openai): carry cache_control to prompt_cache_breakpoint on chat completions to gpt-5.6 (#46101)
* fix(openai): carry cache_control to prompt_cache_breakpoint on chat completions to gpt-5.6 A cache_control mark on a system, developer, user or tool message sent to /v1/chat/completions now reaches an OpenAI deployment that reads prompt_cache_breakpoint (gpt-5.6+ on api.openai.com, or one opted in through prompt_cache_options) as a prompt_cache_breakpoint on the same content part, instead of being stripped. Marks on assistant turns and tool definitions are still dropped, and deployments that do not read breakpoints keep today's behavior. * test(openai): rewrite chat prompt_cache_breakpoint wire test as sent-vs-received literals Each test now shows the request the caller sends and the exact body OpenAI receives as plain JSON, with one helper that returns the upstream body. Drops the per-test marker plumbing and the spend-log check, which did not test the cache mark translation. * test(openai): cover a marked chat message with null content on the breakpoint path A system, developer, user or tool message that carries cache_control but whose content is neither a string nor a list of parts is handed to the existing strip untouched instead of raising inside the transform. The new case pins that the request still goes out with the mark stripped and no breakpoint added. * refactor(openai): reuse the cache control hook's breakpoint placement on chat completions The chat transform had its own copy of how a message-level cache_control lands on a part for OpenAI's prompt_cache_breakpoint. It now calls one hook helper that the injection path also uses, so which roles and part types can carry a breakpoint is decided in one place. A message-level mark on list content now lands on the last part OpenAI accepts a breakpoint on, matching injection. * test(e2e): cover cache_control reaching gpt-5.6 explicit prompt caching on chat completions Four live cases through the proxy: a marked system part, a marked string system message, a marked user part and a marked tool result. Each must write the cache on the first call and read it on the second. * fix(openai): parse content parts instead of a banned TypeGuard when renaming cache marks The strict ruff gate bans typing.TypeGuard as unverified narrowing. The part rename now validates an accepted part with pydantic before handing it to the shared rename helper. |