litellm/litellm
Ishaan Jaff 3cc032bc0c
Add French language support for EU AI Act Article 5 guardrail (#21427)
* Add French language support for EU AI Act Article 5 template

- Create eu_ai_act_article5_fr.yaml with comprehensive French keywords
- Includes identifier words: concevoir, créer, développer, noter, classer, etc.
- Includes block words: crédit social, comportement social, émotion des employés, etc.
- Includes always-block keywords for explicit prohibited practices
- Includes exceptions for research, compliance, and legitimate use cases
- Catches circumvention attempts with phrase variations

* Add comprehensive tests for French EU AI Act guardrail

- Test 3 critical scenarios: blocked query, circumvention attempt, safe query
- Test edge cases: case-insensitive, mixed language, research exceptions
- All 7 tests passing
- Validates both blocking and allowing behavior

* Fix content filter to support conditional matching without inherit_from

- Enable conditional matching when identifier_words + additional_block_words are present
- Previously required inherit_from, but EU AI Act templates are self-contained
- Fixes Greptile feedback: conditional matching now works as documented

* Add pure conditional matching test for French guardrail

- Test identifier + block word combinations not in always_block_keywords
- Verifies conditional matching works independently
- Addresses Greptile feedback about test coverage gap

* Fix exception word bypass risk in French template

- Replace short words (film, jeu, juste) with context-specific phrases
- Prevents substring matching bypasses (e.g., enjeu matching jeu)
- Add tests for bypass prevention and legitimate game context
- Addresses Greptile security feedback

* Make conditional match assertion more robust

- Use getattr to safely access exception detail field
- Check if detail is dict before calling .get()
- Addresses Greptile feedback about brittle string assertion
2026-02-17 16:07:54 -08:00
..
a2a_protocol fix(mypy): resolve type checking errors in 5 files (#20627) 2026-02-06 18:34:55 -08:00
anthropic_interface [bug fix] do not fallback to token counter if disable_token_counter is enabled (#19041) 2026-01-13 16:53:38 -08:00
assistants
batch_completion fix(batch_completion): submit all model futures before waiting (#20705) 2026-02-11 15:44:37 +05:30
batches fix mypy 2026-02-16 19:05:06 +05:30
caching fix(cache): prevent DualCache async batch check-then-act race (#20986) 2026-02-13 18:32:41 +05:30
completion_extras fix openai tests 2026-02-12 17:53:47 -08:00
containers Make keepalive_timeout parameter work for Gunicorn (#19087) 2026-01-16 03:32:59 +05:30
endpoints/speech/speech_to_completion_bridge
evals Add eval run endpoints and methods 2026-02-17 19:32:15 +05:30
experimental_mcp_client fix(mcp): resolve OAuth2 'Capabilities: none' bug for upstream MCP servers (#20602) 2026-02-06 15:00:35 -08:00
files fix(mypy): fix type errors in files, opentelemetry, gemini transformation, and key management (#20161) 2026-01-31 10:09:07 -08:00
fine_tuning
google_genai Add custom vertex ai mapping to the output 2026-01-22 15:18:24 +05:30
images Add None as default image value 2026-01-19 09:08:38 +05:30
integrations fix: remove unused asyncio imports (F401) 2026-02-17 18:54:18 -03:00
interactions [Feat] Interactions API - allow using all litellm providers (interactions -> responses api bridge) (#18373) 2025-12-23 22:30:22 +05:30
litellm_core_utils fix: complete asyncio.iscoroutinefunction deprecation fix across codebase 2026-02-17 18:20:44 -03:00
llms Fix mypy issues 2026-02-17 20:51:51 +05:30
ocr [Feat] /ocr - Add VertexAI OCR provider support + cost tracking (#16216) 2025-11-03 15:56:49 -08:00
passthrough [Fix] VertexAI Pass through - Ensure only anthropic betas are forwarded down to LLM API (#19542) 2026-01-21 19:12:04 -08:00
proxy Add French language support for EU AI Act Article 5 guardrail (#21427) 2026-02-17 16:07:54 -08:00
proxy_auth Fix mypy issues 2026-02-05 09:44:05 +05:30
rag fix(ci): Fix ruff lint error - unused import in vertex_ai_ingestion (#21178) 2026-02-13 18:14:23 -08:00
realtime_api [Feat] Add xAI /realtime API Support - works with LiveKitSDK (#20381) 2026-02-03 19:58:28 -08:00
rerank_api Add tests for header forwarding 2025-12-12 17:54:17 +05:30
responses fix: remove unused Reasoning import from transformation.py 2026-02-15 12:37:41 -03:00
router_strategy fix(router): remove repeated provider parsing in budget limiter hot path (#21043) 2026-02-12 20:05:55 -08:00
router_utils feat: add support for anthropic_messages call type in prompt caching (#19233) 2026-02-07 22:51:06 -08:00
search
secret_managers Fix authorization issues, same alias; verified working 2026-02-05 14:14:20 +05:30
skills [Feat] Unified Skills API - works across Anthropic, Vertex, Azure, Bedrock (#18232) 2025-12-19 18:55:59 +05:30
types Add eval run endpoints methods 2026-02-17 19:32:15 +05:30
vector_store_files Vector store files Stable Release (#16643) 2025-11-15 13:00:33 -08:00
vector_stores Fix vector store configuration synchronization failure 2025-12-05 11:46:14 +05:30
videos fix: respect videos content db creds 2025-12-10 23:00:01 +05:30
__init__.py Add support for evals via sdk 2026-02-17 19:32:15 +05:30
_lazy_imports.py refactor: migrate utils.py lazy imports to registry pattern (#18657) 2026-01-05 09:55:48 -08:00
_lazy_imports_registry.py fix: remove duplicate PerplexityResponsesConfig from LLM_CONFIG_NAMES (#21105) 2026-02-13 08:38:50 -08:00
_logging.py fix:Parse embedded JSON in the message field of logs (#20366) 2026-02-10 16:13:33 +05:30
_redis.py
_service_logger.py Add comments 2026-02-14 19:46:56 -05:00
_uuid.py
_version.py
anthropic_beta_headers_config.json Merge branch 'main' into litellm_litellm_anthropic_remote_url3 2026-02-13 18:40:00 +05:30
anthropic_beta_headers_manager.py Add support sync with remote URL for beta headers 2026-02-13 15:49:42 +05:30
budget_manager.py
constants.py Add Claude Sonnet 4.6 pricing (#21395) 2026-02-17 11:46:24 -08:00
cost.json
cost_calculator.py fix mypy 2026-02-16 18:55:21 +05:30
exceptions.py Fix unsafe access to request attribute (#19573) 2026-01-22 10:58:29 -08:00
main.py feat(vertex_ai): preserve usageMetadata in _hidden_params (#20559) 2026-02-12 20:44:59 -08:00
model_prices_and_context_window_backup.json add default version for opus 4.6 (#21397) 2026-02-17 11:46:44 -08:00
mypy.ini
policy_templates_backup.json Fix EU AI Act template: add missing category_file path (#21424) 2026-02-17 15:07:23 -08:00
py.typed
router.py Merge pull request #21089 from BerriAI/litellm_vector_store_endpoints 2026-02-13 22:18:29 +05:30
scheduler.py fix(scheduler): remove orphan entries from queue - causing memory leak. (#20866) 2026-02-10 22:34:52 -08:00
timeout.py
utils.py fix(token-counter): normalize encode() return type and handle HF tokenizer fallback 2026-02-17 19:19:27 -03:00