scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-15 23:31:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 165
litellm/tests
History
Tin Chi Lo 37c8efde0f fix(bedrock): retry once, re-signed, when Converse rejects extra tool fields 
Bedrock validates some Claude models through an Anthropic-compatible validator
that accepts a narrower `toolSpec` than the Converse API documents, and rejects
the surplus members by presence with
`tools.<i>.<variant>.<field>: Extra inputs are not permitted`. Today the only
defence is `bedrock_converse_supports_strict_tools` in the pricing map, which
means every newly released Claude model is broken for tool calling until a human
notices and adds the flag.

Converse now drops the fields the provider just named and sends the request
again. `parse_rejected_tool_fields` reads the verdict off the error and is shared
with azure_ai, which already had this behaviour for the same error string; the
Bedrock side owns applying it to the Converse request shape. The retry is
single-shot, and a failure that is not an extra-tool-field rejection raises
exactly what it raised before.

The Bedrock-specific part is signing. SigV4 commits to a hash of the body, so a
retry that edits the body has to be signed again; reusing the original headers
would fail as SignatureDoesNotMatch rather than succeed. All four Converse call
paths (sync and async, streaming and not) go through one wrapper that takes the
transport as a parameter, so the retry logic exists once and each path keeps its
own error contract. The two error shapes Converse can raise are both handled:
the non-streaming paths convert to BedrockError themselves, the streaming paths
surface the transport's MaskedHTTPStatusError.

Verified against real Bedrock in us-east-1 on `us.anthropic.claude-sonnet-5`,
which currently lacks the pricing-map flag: all four paths returned a real tool
call for a request carrying `strict: true` that 400s without this change.
2026-08-03 18:43:22 -07:00
..
agent_tests feat(a2a): support a2a-sdk 1.x proxy routing for 0.3 and 1.0 agents (#30950) 2026-06-29 09:32:39 +05:30
audio_tests fix(tests): stabilize image-edit VCR cassettes to stop live gpt-image-1 spend (#28110) 2026-05-18 09:15:39 -07:00
base_sdk_tests fix(deps): move pydantic-settings into the base dependencies 2026-08-01 15:25:26 -07:00
basic_proxy_startup_tests
batches_tests refactor(batches): aggregate batch output cost, usage, and models in a single pass 2026-07-29 22:00:01 -07:00
benchmarks test(benchmarks): run shared logging executor inline to make CodSpeed measurements deterministic (#32435) 2026-07-09 11:14:22 +03:00
code_coverage_tests feat(guardrails): scan and mask MCP tool results via post_mcp_call (#35155) 2026-07-30 14:10:26 -07:00
documentation_tests feat(messages): route native Anthropic /messages through Rust behind LITELLM_RUST env var (#33848) 2026-07-18 18:27:20 -07:00
e2e refactor(ui): rename the create MCP server component to PascalCase (#35686) 2026-08-03 13:39:07 -07:00
enterprise feat(prometheus): add global exclude_metrics and exclude_labels options (#34201) 2026-08-01 11:52:27 -07:00
guardrails_tests feat: add deepkeep as custom guardrail (#33844) 2026-07-20 19:27:40 -07:00
image_gen_tests test(bedrock): switch image gen live test off EOL Titan to Nova Canvas (#31937) 2026-07-01 22:57:29 -07:00
integration feat(sandbox): e2b code execution primitive (#30898) 2026-06-20 16:30:01 -07:00
litellm fix(anthropic): strip all remaining output_format schema keywords rejected by Anthropic 2026-07-22 16:16:04 -07:00
litellm-proxy-extras
litellm_core_utils
litellm_utils_tests refactor(bedrock): remove the dead BedrockLLM invoke code path 2026-07-29 20:25:36 -07:00
llm_responses_api_testing Revert "chore(ci): sync litellm_internal_staging into daily OSS branch (#33337)" (#33339) 2026-07-14 19:32:25 -07:00
llm_translation test(bedrock): port the openai-route invoke tests onto the live config 2026-07-29 20:59:08 -07:00
load_tests
local_testing test(proxy): assert _delete_deployment's still-desired id set instead of a delete count 2026-08-01 14:36:30 -07:00
logging_callback_tests test(logging): pin routing_decision and internal_call_origin in the gcs pubsub spend log fixture 2026-08-01 14:34:04 -07:00
mcp_tests fix(mcp): keep REST tool listing in step with key/team grant enforcement 2026-07-30 22:13:10 -07:00
multi_instance_e2e_tests
ocr_tests test(ocr): use mistral-document-ai-2512 in azure_ai OCR tests 2026-07-15 18:13:22 -07:00
old_proxy_tests/tests
openai_endpoints_tests chore(ci): modernize model references in tests and configs (#27856) 2026-05-15 15:44:28 -07:00
otel_tests fix(cli): mint per-session agent credential on lite login (#31072) 2026-06-26 09:05:15 -07:00
pass_through_tests test(pass-through): de-flake vertex spend-log test by routing through the proxy (#31689) 2026-06-30 15:27:48 -07:00
pass_through_unit_tests fix(streaming): stamp completion_start_time on first chunk for /v1/messages and /v1/responses (#32284) 2026-07-06 19:30:31 -07:00
proxy_admin_ui_tests Revert "chore: remove _experimental/out (#31546)" 2026-07-01 13:25:47 -07:00
proxy_behavior feat(teams): apply default organization to new teams from default team settings (#35540) 2026-08-03 12:57:12 -07:00
proxy_e2e_anthropic_messages_tests Extend the record/replay proxy to chat, embeddings, moderations, rerank, and Anthropic (#29847) 2026-06-06 14:33:42 -07:00
proxy_migration_tests fix(docker): bake prisma offline in the componentized migrations image (#35485) 2026-08-01 14:12:31 -07:00
proxy_security_tests test(proxy): stop running real-DB tests in GitHub Actions unit jobs (#29700) 2026-06-04 14:56:02 -07:00
proxy_unit_tests Merge pull request #35406 from BerriAI/devin_ai_fix_batch_output_file_id_encoding_lit4964 2026-08-01 14:12:43 -07:00
router_unit_tests fix(auth): resolve managed batch/file deployment model_id to model name for team access checks 2026-07-28 15:35:19 +00:00
scim_tests
search_tests feat(tinyfish): make search provider permissive, attribute errors (#31997) 2026-07-03 10:17:11 -07:00
spend_tracking_tests chore(ci): modernize model references in tests and configs (#27856) 2026-05-15 15:44:28 -07:00
store_model_in_db_tests test: de-flake langfuse callbacks-in-db e2e test 2026-07-04 19:14:53 -07:00
test_litellm fix(bedrock): retry once, re-signed, when Converse rejects extra tool fields 2026-08-03 18:43:22 -07:00
unified_google_tests fix(google_genai): preserve complete SSE events in Vertex/Gemini image streaming (#30270) 2026-06-12 07:49:30 -07:00
vector_store_tests Revert "chore(tests): migrate Bedrock CI to AWS account 941277531214 (#28728)" (#29326) 2026-05-30 11:26:24 -07:00
windows_tests ci: reproduce default-Windows wheel install to guard MAX_PATH (#29597) 2026-06-03 11:28:08 -07:00
__init__.py
_fake_openai_endpoint_server.py test: point router/completion/triton tests at the local fake OpenAI endpoint (#30900) 2026-06-20 16:20:35 -07:00
_flush_vcr_cache.py tests(vcr): isolate cassette redis to CASSETTE_REDIS_URL 2026-05-01 12:32:59 -07:00
_live_test_helpers.py test(vcr): close out the remaining VCR live-call leaks (#29603) 2026-06-03 13:46:43 -07:00
_openai_record_replay_proxy.py Extend the record/replay proxy to chat, embeddings, moderations, rerank, and Anthropic (#29847) 2026-06-06 14:33:42 -07:00
_vcr_conftest_common.py test(vcr): close out the remaining VCR live-call leaks (#29603) 2026-06-03 13:46:43 -07:00
_vcr_redis_persister.py test(vcr): stop refreshing cassette TTL on read so cassettes lapse after 24h (#29784) 2026-06-05 10:22:41 -07:00
_ws_vcr.py test(realtime): record and replay websocket traffic in redis vcr cassettes (#32390) 2026-07-08 00:19:06 -07:00
eval_swe_bench.py
fake_openai_endpoint.py test: point router/completion/triton tests at the local fake OpenAI endpoint (#30900) 2026-06-20 16:20:35 -07:00
gettysburg.wav
large_text.py
openai_batch_completions.jsonl
pyrightconfig.json test: add e2e tests for spend, budgets and llms (#30869) 2026-06-24 15:01:57 -07:00
README.MD
test_anthropic_compaction_usage.py chore(oss): litellm oss staging 150626 (#30463) 2026-06-16 12:06:41 -07:00
test_budget_management.py
test_callbacks_on_proxy.py test(callbacks): harden flaky proxy callback-leak detector (#28195) 2026-05-18 16:39:02 -07:00
test_config.py
test_debug_warning.py
test_default_encoding_non_root.py
test_end_users.py chore(ci): modernize model references in tests and configs (#27856) 2026-05-15 15:44:28 -07:00
test_entrypoint.py
test_fallbacks.py
test_gpt5_azure_temperature_support.py
test_health.py fix(tests): swap dall-e to gpt-image-1 after openai deprecation 2026-05-12 16:55:18 -07:00
test_keys.py test: harden remaining pass-through CI flakes (image-gen spend poll, ruby assistants timeout) (#30685) 2026-06-17 14:35:47 -07:00
test_litellm_proxy_responses_config.py chore(ci): modernize model references in tests and configs (#27856) 2026-05-15 15:44:28 -07:00
test_logging.conf
test_models.py
test_new_vector_store_endpoints.py
test_openai_endpoints.py Extend the record/replay proxy to chat, embeddings, moderations, rerank, and Anthropic (#29847) 2026-06-06 14:33:42 -07:00
test_organizations.py
test_otel_thread_leak.py
test_passthrough_endpoints.py
test_presidio_latency.py
test_proxy_server_non_root.py
test_ratelimit.py chore(ci): modernize model references in tests and configs (#27856) 2026-05-15 15:44:28 -07:00
test_resource_cleanup.py
test_service_logger_otel.py fix(langfuse): send v4 ingestion header for otel callback (#33907) 2026-07-18 20:36:51 -07:00
test_spend_logs.py Litellm oss staging 04 21 2026 2 (#26569) 2026-05-20 21:25:19 -07:00
test_team.py
test_team_logging.py
test_team_members.py Litellm oss staging 04 21 2026 2 (#26569) 2026-05-20 21:25:19 -07:00
test_users.py Fix: tag budget reset must drop stale management-cache entry (#27568) 2026-05-10 00:18:55 +00:00

README.MD

In total litellm runs 1000+ tests

[02/20/2025] Update:

To make it easier to contribute and map what behavior is tested,

we've started mapping the litellm directory in tests/test_litellm

This folder can only run mock tests.