agentjido / req_llm / d856b37f3667c0a313b73f60c3fed8dd5282553b-PR-169

Builds Branch Commit Type Ran Committer Via Coverage
d856b37f3667c0a313b73f60c3fed8dd5282553b-PR-169 chore/refresh-fixtures-before-1.0 Update test configuration for catalog_allow system - Configure test.exs to allow all FakeRegistry providers (:all) - Update ModelMatrix test expectations to match new default models - Removed references to old default models (gpt-4-turbo, claude-... Pull #169 mikehostetler github pending completion   set done
74abc43f5e428c095fef3946b35949669bfe8455-PR-168 fix/bedrock-claude-exclusions fix(bedrock): Exclude old Claude models from test suite Adds exclusion file to filter out Claude 3.x and legacy 4.0 models that lack fixtures. Only Claude 4.5 (Haiku, Sonnet) and Claude 4.1 (Opus) remain in test coverage. Fixes test failures whe... Pull #168 neilberkman github
52.47
94e1b59bbfbbeb8fdbbe66c4928ed99747da4a45 main fix(vertex): Rename decode_sse_event to decode_stream_event (#167) Update function name to match Provider behavior callback. Fixes compilation warnings. push web-flow github
52.54
6e1b4675d06f0e099a832cedbd4478b17eba449a-PR-167 fix/vertex-decode-stream-event fix(vertex): Rename decode_sse_event to decode_stream_event Update function name to match Provider behavior callback. Fixes compilation warnings. Pull #167 neilberkman github
52.57
64e00ce70969c7dd3ac8dc7655bd21b877725d37 main refactor: Rename decode_sse_event to decode_stream_event for protocol agnosticity (#166) Renames the Provider callback from decode_sse_event to decode_stream_event to reflect that providers use different streaming protocols: - OpenAI, Anthropic,... push web-flow github
54.02
2fcd2769b57f95c8b1fd6d0f16e8c27911cddc6c-PR-152 fix/bedrock-inference-profiles test: Re-record Bedrock fixtures with thinking_constraints implementation Re-recorded 72 Bedrock fixtures with updated provider implementation. Results: 59/65 tests passing. Expected failures: - Opus token_limit (AWS signature expiry on long-runn... Pull #152 neilberkman github pending completion   set done
5a1398105bc7994d902205c4d18d454d99369265-PR-152 fix/bedrock-inference-profiles test: Re-record Vertex AI fixtures with final implementation Re-recorded all 36 fixtures (12 per model) for Haiku 4.5, Sonnet 4.5, and Opus 4.1 with the complete provider implementation including thinking_constraints callback. All 30 tests passing. Pull #152 neilberkman github pending completion   set done
6ea6929fa1cbaae7ce544fa109bcc7c6080f3a79-PR-152 fix/bedrock-inference-profiles refactor: Replace hardcoded Bedrock checks with thinking_constraints callback Add thinking_constraints/0 callback to Provider behavior to handle platform-specific requirements for extended thinking (e.g., required temperature and min max_tokens).... Pull #152 neilberkman github pending completion   set done
6025614f88473265c113f3f72d7487ae3b57d06b-PR-152 fix/bedrock-inference-profiles refactor: Replace hardcoded Bedrock checks with thinking_constraints callback Add thinking_constraints/0 callback to Provider behavior to handle platform-specific requirements for extended thinking (e.g., required temperature and min max_tokens).... Pull #152 neilberkman github pending completion   set done
4baf52ab9e461c626dfc46719975a9d2c6f6e64d-PR-152 fix/bedrock-inference-profiles Add exclusions for older Vertex Claude models without fixtures Pull #152 neilberkman github pending completion   set done
99866d945db559c708b6d5f0c68c6218aead0ed2-PR-152 fix/bedrock-inference-profiles docs(model_sync): Clarify patch file provider ID requirements Explain why patch files must use original hyphenated IDs from models.dev (e.g., google-vertex-anthropic) rather than normalized underscored IDs. The models_data map keys come from mode... Pull #152 neilberkman github pending completion   set done
49055e9c007b489353f0b2fe05548a4e8e867ab6-PR-152 fix/bedrock-inference-profiles Add exclusions for older Vertex Claude models without fixtures Pull #152 neilberkman github pending completion   set done
d53adff3465b91e39b38925bac799a82d7210432-PR-152 fix/bedrock-inference-profiles feat(vertex): Add Google Vertex AI provider with complete Claude support Implements Google Vertex AI provider for Anthropic Claude models with streaming, extended thinking, tool calling, structured output, and prompt caching. Core implementation... Pull #152 neilberkman github pending completion   set done
3953ef56b19866879ec23aabf6126597ae28960e-PR-152 fix/bedrock-inference-profiles feat(vertex): Add Google Vertex AI provider with complete Claude support Implements Google Vertex AI provider for Anthropic Claude models with streaming, extended thinking, tool calling, structured output, and prompt caching. Core implementation... Pull #152 neilberkman github pending completion   set done
2ead740a2079278ae05fdd8244caf9fbec7f1830-PR-152 fix/bedrock-inference-profiles feat(vertex): Add Google Vertex AI provider with complete Claude support Implements Google Vertex AI provider for Anthropic Claude models with streaming, extended thinking, tool calling, structured output, and prompt caching. Core implementation... Pull #152 neilberkman github pending completion   set done