llama-stack 0.0.42__py3-none-any.whl → 0.3.4__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- llama_stack/__init__.py +5 -0
- llama_stack/apis/agents/__init__.py +1 -1
- llama_stack/apis/agents/agents.py +700 -281
- llama_stack/apis/agents/openai_responses.py +1311 -0
- llama_stack/{providers/adapters/memory/sample/config.py → apis/batches/__init__.py} +2 -5
- llama_stack/apis/batches/batches.py +100 -0
- llama_stack/apis/benchmarks/__init__.py +7 -0
- llama_stack/apis/benchmarks/benchmarks.py +108 -0
- llama_stack/apis/common/content_types.py +143 -0
- llama_stack/apis/common/errors.py +103 -0
- llama_stack/apis/common/job_types.py +38 -0
- llama_stack/apis/common/responses.py +36 -0
- llama_stack/apis/common/training_types.py +36 -5
- llama_stack/apis/common/type_system.py +158 -0
- llama_stack/apis/conversations/__init__.py +31 -0
- llama_stack/apis/conversations/conversations.py +286 -0
- llama_stack/apis/datasetio/__init__.py +7 -0
- llama_stack/apis/datasetio/datasetio.py +59 -0
- llama_stack/apis/datasets/__init__.py +7 -0
- llama_stack/apis/datasets/datasets.py +251 -0
- llama_stack/apis/datatypes.py +160 -0
- llama_stack/apis/eval/__init__.py +7 -0
- llama_stack/apis/eval/eval.py +169 -0
- llama_stack/apis/files/__init__.py +7 -0
- llama_stack/apis/files/files.py +199 -0
- llama_stack/apis/inference/__init__.py +1 -1
- llama_stack/apis/inference/inference.py +1169 -113
- llama_stack/apis/inspect/__init__.py +1 -1
- llama_stack/apis/inspect/inspect.py +69 -16
- llama_stack/apis/models/__init__.py +1 -1
- llama_stack/apis/models/models.py +148 -21
- llama_stack/apis/post_training/__init__.py +1 -1
- llama_stack/apis/post_training/post_training.py +265 -120
- llama_stack/{providers/adapters/agents/sample/config.py → apis/prompts/__init__.py} +2 -5
- llama_stack/apis/prompts/prompts.py +204 -0
- llama_stack/apis/providers/__init__.py +7 -0
- llama_stack/apis/providers/providers.py +69 -0
- llama_stack/apis/resource.py +37 -0
- llama_stack/apis/safety/__init__.py +1 -1
- llama_stack/apis/safety/safety.py +95 -12
- llama_stack/apis/scoring/__init__.py +7 -0
- llama_stack/apis/scoring/scoring.py +93 -0
- llama_stack/apis/scoring_functions/__init__.py +7 -0
- llama_stack/apis/scoring_functions/scoring_functions.py +208 -0
- llama_stack/apis/shields/__init__.py +1 -1
- llama_stack/apis/shields/shields.py +76 -33
- llama_stack/apis/synthetic_data_generation/__init__.py +1 -1
- llama_stack/apis/synthetic_data_generation/synthetic_data_generation.py +40 -17
- llama_stack/apis/telemetry/__init__.py +1 -1
- llama_stack/apis/telemetry/telemetry.py +322 -31
- llama_stack/apis/{dataset → tools}/__init__.py +2 -1
- llama_stack/apis/tools/rag_tool.py +218 -0
- llama_stack/apis/tools/tools.py +221 -0
- llama_stack/apis/vector_io/__init__.py +7 -0
- llama_stack/apis/vector_io/vector_io.py +960 -0
- llama_stack/apis/vector_stores/__init__.py +7 -0
- llama_stack/apis/vector_stores/vector_stores.py +51 -0
- llama_stack/apis/version.py +9 -0
- llama_stack/cli/llama.py +13 -5
- llama_stack/cli/stack/_list_deps.py +182 -0
- llama_stack/cli/stack/list_apis.py +1 -1
- llama_stack/cli/stack/list_deps.py +55 -0
- llama_stack/cli/stack/list_providers.py +24 -10
- llama_stack/cli/stack/list_stacks.py +56 -0
- llama_stack/cli/stack/remove.py +115 -0
- llama_stack/cli/stack/run.py +169 -56
- llama_stack/cli/stack/stack.py +18 -4
- llama_stack/cli/stack/utils.py +151 -0
- llama_stack/cli/table.py +23 -61
- llama_stack/cli/utils.py +29 -0
- llama_stack/core/access_control/access_control.py +131 -0
- llama_stack/core/access_control/conditions.py +129 -0
- llama_stack/core/access_control/datatypes.py +107 -0
- llama_stack/core/build.py +164 -0
- llama_stack/core/client.py +205 -0
- llama_stack/core/common.sh +37 -0
- llama_stack/{distribution → core}/configure.py +74 -55
- llama_stack/core/conversations/conversations.py +309 -0
- llama_stack/core/datatypes.py +625 -0
- llama_stack/core/distribution.py +276 -0
- llama_stack/core/external.py +54 -0
- llama_stack/core/id_generation.py +42 -0
- llama_stack/core/inspect.py +86 -0
- llama_stack/core/library_client.py +539 -0
- llama_stack/core/prompts/prompts.py +234 -0
- llama_stack/core/providers.py +137 -0
- llama_stack/core/request_headers.py +115 -0
- llama_stack/core/resolver.py +506 -0
- llama_stack/core/routers/__init__.py +101 -0
- llama_stack/core/routers/datasets.py +73 -0
- llama_stack/core/routers/eval_scoring.py +155 -0
- llama_stack/core/routers/inference.py +645 -0
- llama_stack/core/routers/safety.py +85 -0
- llama_stack/core/routers/tool_runtime.py +91 -0
- llama_stack/core/routers/vector_io.py +442 -0
- llama_stack/core/routing_tables/benchmarks.py +62 -0
- llama_stack/core/routing_tables/common.py +254 -0
- llama_stack/core/routing_tables/datasets.py +91 -0
- llama_stack/core/routing_tables/models.py +163 -0
- llama_stack/core/routing_tables/scoring_functions.py +66 -0
- llama_stack/core/routing_tables/shields.py +61 -0
- llama_stack/core/routing_tables/toolgroups.py +129 -0
- llama_stack/core/routing_tables/vector_stores.py +292 -0
- llama_stack/core/server/auth.py +187 -0
- llama_stack/core/server/auth_providers.py +494 -0
- llama_stack/core/server/quota.py +110 -0
- llama_stack/core/server/routes.py +141 -0
- llama_stack/core/server/server.py +542 -0
- llama_stack/core/server/tracing.py +80 -0
- llama_stack/core/stack.py +546 -0
- llama_stack/core/start_stack.sh +117 -0
- llama_stack/core/storage/datatypes.py +283 -0
- llama_stack/{cli/model → core/store}/__init__.py +1 -1
- llama_stack/core/store/registry.py +199 -0
- llama_stack/core/testing_context.py +49 -0
- llama_stack/core/ui/app.py +55 -0
- llama_stack/core/ui/modules/api.py +32 -0
- llama_stack/core/ui/modules/utils.py +42 -0
- llama_stack/core/ui/page/distribution/datasets.py +18 -0
- llama_stack/core/ui/page/distribution/eval_tasks.py +20 -0
- llama_stack/core/ui/page/distribution/models.py +18 -0
- llama_stack/core/ui/page/distribution/providers.py +27 -0
- llama_stack/core/ui/page/distribution/resources.py +48 -0
- llama_stack/core/ui/page/distribution/scoring_functions.py +18 -0
- llama_stack/core/ui/page/distribution/shields.py +19 -0
- llama_stack/core/ui/page/evaluations/app_eval.py +143 -0
- llama_stack/core/ui/page/evaluations/native_eval.py +253 -0
- llama_stack/core/ui/page/playground/chat.py +130 -0
- llama_stack/core/ui/page/playground/tools.py +352 -0
- llama_stack/core/utils/config.py +30 -0
- llama_stack/{distribution → core}/utils/config_dirs.py +3 -6
- llama_stack/core/utils/config_resolution.py +125 -0
- llama_stack/core/utils/context.py +84 -0
- llama_stack/core/utils/exec.py +96 -0
- llama_stack/{providers/impls/meta_reference/codeshield/config.py → core/utils/image_types.py} +4 -3
- llama_stack/{distribution → core}/utils/model_utils.py +2 -2
- llama_stack/{distribution → core}/utils/prompt_for_config.py +30 -63
- llama_stack/{apis/batch_inference → distributions/dell}/__init__.py +1 -1
- llama_stack/distributions/dell/build.yaml +33 -0
- llama_stack/distributions/dell/dell.py +158 -0
- llama_stack/distributions/dell/run-with-safety.yaml +141 -0
- llama_stack/distributions/dell/run.yaml +132 -0
- llama_stack/distributions/meta-reference-gpu/__init__.py +7 -0
- llama_stack/distributions/meta-reference-gpu/build.yaml +32 -0
- llama_stack/distributions/meta-reference-gpu/meta_reference.py +163 -0
- llama_stack/distributions/meta-reference-gpu/run-with-safety.yaml +154 -0
- llama_stack/distributions/meta-reference-gpu/run.yaml +139 -0
- llama_stack/{apis/evals → distributions/nvidia}/__init__.py +1 -1
- llama_stack/distributions/nvidia/build.yaml +29 -0
- llama_stack/distributions/nvidia/nvidia.py +154 -0
- llama_stack/distributions/nvidia/run-with-safety.yaml +137 -0
- llama_stack/distributions/nvidia/run.yaml +116 -0
- llama_stack/distributions/open-benchmark/__init__.py +7 -0
- llama_stack/distributions/open-benchmark/build.yaml +36 -0
- llama_stack/distributions/open-benchmark/open_benchmark.py +303 -0
- llama_stack/distributions/open-benchmark/run.yaml +252 -0
- llama_stack/distributions/postgres-demo/__init__.py +7 -0
- llama_stack/distributions/postgres-demo/build.yaml +23 -0
- llama_stack/distributions/postgres-demo/postgres_demo.py +125 -0
- llama_stack/distributions/postgres-demo/run.yaml +115 -0
- llama_stack/{apis/memory → distributions/starter}/__init__.py +1 -1
- llama_stack/distributions/starter/build.yaml +61 -0
- llama_stack/distributions/starter/run-with-postgres-store.yaml +285 -0
- llama_stack/distributions/starter/run.yaml +276 -0
- llama_stack/distributions/starter/starter.py +345 -0
- llama_stack/distributions/starter-gpu/__init__.py +7 -0
- llama_stack/distributions/starter-gpu/build.yaml +61 -0
- llama_stack/distributions/starter-gpu/run-with-postgres-store.yaml +288 -0
- llama_stack/distributions/starter-gpu/run.yaml +279 -0
- llama_stack/distributions/starter-gpu/starter_gpu.py +20 -0
- llama_stack/distributions/template.py +456 -0
- llama_stack/distributions/watsonx/__init__.py +7 -0
- llama_stack/distributions/watsonx/build.yaml +33 -0
- llama_stack/distributions/watsonx/run.yaml +133 -0
- llama_stack/distributions/watsonx/watsonx.py +95 -0
- llama_stack/env.py +24 -0
- llama_stack/log.py +314 -0
- llama_stack/models/llama/checkpoint.py +164 -0
- llama_stack/models/llama/datatypes.py +164 -0
- llama_stack/models/llama/hadamard_utils.py +86 -0
- llama_stack/models/llama/llama3/args.py +74 -0
- llama_stack/models/llama/llama3/chat_format.py +286 -0
- llama_stack/models/llama/llama3/generation.py +376 -0
- llama_stack/models/llama/llama3/interface.py +255 -0
- llama_stack/models/llama/llama3/model.py +304 -0
- llama_stack/models/llama/llama3/multimodal/__init__.py +12 -0
- llama_stack/models/llama/llama3/multimodal/encoder_utils.py +180 -0
- llama_stack/models/llama/llama3/multimodal/image_transform.py +409 -0
- llama_stack/models/llama/llama3/multimodal/model.py +1430 -0
- llama_stack/models/llama/llama3/multimodal/utils.py +26 -0
- llama_stack/models/llama/llama3/prompt_templates/__init__.py +22 -0
- llama_stack/models/llama/llama3/prompt_templates/base.py +39 -0
- llama_stack/models/llama/llama3/prompt_templates/system_prompts.py +319 -0
- llama_stack/models/llama/llama3/prompt_templates/tool_response.py +62 -0
- llama_stack/models/llama/llama3/quantization/loader.py +316 -0
- llama_stack/models/llama/llama3/template_data.py +116 -0
- llama_stack/models/llama/llama3/tokenizer.model +128000 -0
- llama_stack/models/llama/llama3/tokenizer.py +198 -0
- llama_stack/models/llama/llama3/tool_utils.py +266 -0
- llama_stack/models/llama/llama3_1/__init__.py +12 -0
- llama_stack/models/llama/llama3_1/prompt_format.md +358 -0
- llama_stack/models/llama/llama3_1/prompts.py +258 -0
- llama_stack/models/llama/llama3_2/prompts_text.py +229 -0
- llama_stack/models/llama/llama3_2/prompts_vision.py +126 -0
- llama_stack/models/llama/llama3_2/text_prompt_format.md +286 -0
- llama_stack/models/llama/llama3_2/vision_prompt_format.md +141 -0
- llama_stack/models/llama/llama3_3/prompts.py +259 -0
- llama_stack/models/llama/llama4/args.py +107 -0
- llama_stack/models/llama/llama4/chat_format.py +317 -0
- llama_stack/models/llama/llama4/datatypes.py +56 -0
- llama_stack/models/llama/llama4/ffn.py +58 -0
- llama_stack/models/llama/llama4/generation.py +313 -0
- llama_stack/models/llama/llama4/model.py +437 -0
- llama_stack/models/llama/llama4/moe.py +214 -0
- llama_stack/models/llama/llama4/preprocess.py +435 -0
- llama_stack/models/llama/llama4/prompt_format.md +304 -0
- llama_stack/models/llama/llama4/prompt_templates/system_prompts.py +136 -0
- llama_stack/models/llama/llama4/prompts.py +279 -0
- llama_stack/models/llama/llama4/quantization/__init__.py +5 -0
- llama_stack/models/llama/llama4/quantization/loader.py +226 -0
- llama_stack/models/llama/llama4/tokenizer.model +200000 -0
- llama_stack/models/llama/llama4/tokenizer.py +263 -0
- llama_stack/models/llama/llama4/vision/__init__.py +5 -0
- llama_stack/models/llama/llama4/vision/embedding.py +210 -0
- llama_stack/models/llama/llama4/vision/encoder.py +412 -0
- llama_stack/models/llama/prompt_format.py +191 -0
- llama_stack/models/llama/quantize_impls.py +316 -0
- llama_stack/models/llama/sku_list.py +1029 -0
- llama_stack/models/llama/sku_types.py +233 -0
- llama_stack/models/llama/tokenizer_utils.py +40 -0
- llama_stack/providers/datatypes.py +136 -107
- llama_stack/providers/inline/__init__.py +5 -0
- llama_stack/providers/inline/agents/__init__.py +5 -0
- llama_stack/providers/{impls/meta_reference/agents → inline/agents/meta_reference}/__init__.py +12 -5
- llama_stack/providers/inline/agents/meta_reference/agent_instance.py +1024 -0
- llama_stack/providers/inline/agents/meta_reference/agents.py +383 -0
- llama_stack/providers/inline/agents/meta_reference/config.py +37 -0
- llama_stack/providers/inline/agents/meta_reference/persistence.py +228 -0
- llama_stack/providers/inline/agents/meta_reference/responses/__init__.py +5 -0
- llama_stack/providers/inline/agents/meta_reference/responses/openai_responses.py +423 -0
- llama_stack/providers/inline/agents/meta_reference/responses/streaming.py +1226 -0
- llama_stack/providers/inline/agents/meta_reference/responses/tool_executor.py +449 -0
- llama_stack/providers/inline/agents/meta_reference/responses/types.py +194 -0
- llama_stack/providers/inline/agents/meta_reference/responses/utils.py +365 -0
- llama_stack/providers/inline/agents/meta_reference/safety.py +52 -0
- llama_stack/providers/inline/batches/__init__.py +5 -0
- llama_stack/providers/inline/batches/reference/__init__.py +36 -0
- llama_stack/providers/inline/batches/reference/batches.py +679 -0
- llama_stack/providers/inline/batches/reference/config.py +40 -0
- llama_stack/providers/inline/datasetio/__init__.py +5 -0
- llama_stack/providers/inline/datasetio/localfs/__init__.py +20 -0
- llama_stack/providers/inline/datasetio/localfs/config.py +23 -0
- llama_stack/providers/inline/datasetio/localfs/datasetio.py +113 -0
- llama_stack/providers/inline/eval/__init__.py +5 -0
- llama_stack/providers/inline/eval/meta_reference/__init__.py +28 -0
- llama_stack/providers/inline/eval/meta_reference/config.py +23 -0
- llama_stack/providers/inline/eval/meta_reference/eval.py +259 -0
- llama_stack/providers/inline/files/localfs/__init__.py +20 -0
- llama_stack/providers/inline/files/localfs/config.py +31 -0
- llama_stack/providers/inline/files/localfs/files.py +219 -0
- llama_stack/providers/inline/inference/__init__.py +5 -0
- llama_stack/providers/{impls/meta_reference/inference → inline/inference/meta_reference}/__init__.py +4 -4
- llama_stack/providers/inline/inference/meta_reference/common.py +24 -0
- llama_stack/providers/inline/inference/meta_reference/config.py +68 -0
- llama_stack/providers/inline/inference/meta_reference/generators.py +211 -0
- llama_stack/providers/inline/inference/meta_reference/inference.py +158 -0
- llama_stack/providers/inline/inference/meta_reference/model_parallel.py +96 -0
- llama_stack/providers/{impls/meta_reference/inference → inline/inference/meta_reference}/parallel_utils.py +56 -73
- llama_stack/providers/inline/inference/sentence_transformers/__init__.py +22 -0
- llama_stack/providers/{impls/meta_reference/agents → inline/inference/sentence_transformers}/config.py +6 -4
- llama_stack/providers/inline/inference/sentence_transformers/sentence_transformers.py +83 -0
- llama_stack/providers/inline/post_training/__init__.py +5 -0
- llama_stack/providers/inline/post_training/common/__init__.py +5 -0
- llama_stack/providers/inline/post_training/common/utils.py +35 -0
- llama_stack/providers/inline/post_training/common/validator.py +36 -0
- llama_stack/providers/inline/post_training/huggingface/__init__.py +27 -0
- llama_stack/providers/inline/post_training/huggingface/config.py +83 -0
- llama_stack/providers/inline/post_training/huggingface/post_training.py +208 -0
- llama_stack/providers/inline/post_training/huggingface/recipes/__init__.py +5 -0
- llama_stack/providers/inline/post_training/huggingface/recipes/finetune_single_device.py +519 -0
- llama_stack/providers/inline/post_training/huggingface/recipes/finetune_single_device_dpo.py +485 -0
- llama_stack/providers/inline/post_training/huggingface/utils.py +269 -0
- llama_stack/providers/inline/post_training/torchtune/__init__.py +27 -0
- llama_stack/providers/inline/post_training/torchtune/common/__init__.py +5 -0
- llama_stack/providers/inline/post_training/torchtune/common/checkpointer.py +240 -0
- llama_stack/providers/inline/post_training/torchtune/common/utils.py +99 -0
- llama_stack/providers/inline/post_training/torchtune/config.py +20 -0
- llama_stack/providers/inline/post_training/torchtune/datasets/__init__.py +5 -0
- llama_stack/providers/inline/post_training/torchtune/datasets/format_adapter.py +57 -0
- llama_stack/providers/inline/post_training/torchtune/datasets/sft.py +78 -0
- llama_stack/providers/inline/post_training/torchtune/post_training.py +178 -0
- llama_stack/providers/inline/post_training/torchtune/recipes/__init__.py +5 -0
- llama_stack/providers/inline/post_training/torchtune/recipes/lora_finetuning_single_device.py +588 -0
- llama_stack/providers/inline/safety/__init__.py +5 -0
- llama_stack/providers/{impls/meta_reference/codeshield → inline/safety/code_scanner}/__init__.py +4 -2
- llama_stack/providers/inline/safety/code_scanner/code_scanner.py +128 -0
- llama_stack/providers/{impls/meta_reference/memory → inline/safety/code_scanner}/config.py +5 -3
- llama_stack/providers/inline/safety/llama_guard/__init__.py +19 -0
- llama_stack/providers/inline/safety/llama_guard/config.py +19 -0
- llama_stack/providers/inline/safety/llama_guard/llama_guard.py +489 -0
- llama_stack/providers/{adapters/memory/sample → inline/safety/prompt_guard}/__init__.py +4 -4
- llama_stack/providers/inline/safety/prompt_guard/config.py +32 -0
- llama_stack/providers/inline/safety/prompt_guard/prompt_guard.py +131 -0
- llama_stack/providers/inline/scoring/__init__.py +5 -0
- llama_stack/providers/inline/scoring/basic/__init__.py +25 -0
- llama_stack/providers/{adapters/memory/weaviate → inline/scoring/basic}/config.py +5 -7
- llama_stack/providers/inline/scoring/basic/scoring.py +126 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/__init__.py +5 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/docvqa_scoring_fn.py +240 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/equality_scoring_fn.py +41 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/__init__.py +5 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/docvqa.py +21 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/equality.py +21 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/ifeval.py +23 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/regex_parser_math_response.py +27 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/regex_parser_multiple_choice_answer.py +71 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/fn_defs/subset_of.py +21 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/ifeval_scoring_fn.py +80 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/regex_parser_math_response_scoring_fn.py +66 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/regex_parser_scoring_fn.py +58 -0
- llama_stack/providers/inline/scoring/basic/scoring_fn/subset_of_scoring_fn.py +38 -0
- llama_stack/providers/inline/scoring/basic/utils/__init__.py +5 -0
- llama_stack/providers/inline/scoring/basic/utils/ifeval_utils.py +3319 -0
- llama_stack/providers/inline/scoring/basic/utils/math_utils.py +330 -0
- llama_stack/providers/inline/scoring/braintrust/__init__.py +27 -0
- llama_stack/providers/inline/scoring/braintrust/braintrust.py +230 -0
- llama_stack/providers/inline/scoring/braintrust/config.py +21 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/__init__.py +5 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/__init__.py +5 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/answer_correctness.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/answer_relevancy.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/answer_similarity.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_entity_recall.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_precision.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_recall.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_relevancy.py +23 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/factuality.py +24 -0
- llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/faithfulness.py +24 -0
- llama_stack/providers/inline/scoring/llm_as_judge/__init__.py +21 -0
- llama_stack/providers/inline/scoring/llm_as_judge/config.py +14 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring.py +113 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/__init__.py +5 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/__init__.py +5 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/llm_as_judge_405b_simpleqa.py +96 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/llm_as_judge_base.py +20 -0
- llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/llm_as_judge_scoring_fn.py +81 -0
- llama_stack/providers/inline/telemetry/__init__.py +5 -0
- llama_stack/providers/inline/telemetry/meta_reference/__init__.py +21 -0
- llama_stack/providers/inline/telemetry/meta_reference/config.py +47 -0
- llama_stack/providers/inline/telemetry/meta_reference/telemetry.py +252 -0
- llama_stack/providers/inline/tool_runtime/__init__.py +5 -0
- llama_stack/providers/inline/tool_runtime/rag/__init__.py +19 -0
- llama_stack/providers/{impls/meta_reference/telemetry → inline/tool_runtime/rag}/config.py +5 -3
- llama_stack/providers/inline/tool_runtime/rag/context_retriever.py +77 -0
- llama_stack/providers/inline/tool_runtime/rag/memory.py +332 -0
- llama_stack/providers/inline/vector_io/__init__.py +5 -0
- llama_stack/providers/inline/vector_io/chroma/__init__.py +19 -0
- llama_stack/providers/inline/vector_io/chroma/config.py +30 -0
- llama_stack/providers/inline/vector_io/faiss/__init__.py +21 -0
- llama_stack/providers/inline/vector_io/faiss/config.py +26 -0
- llama_stack/providers/inline/vector_io/faiss/faiss.py +293 -0
- llama_stack/providers/inline/vector_io/milvus/__init__.py +19 -0
- llama_stack/providers/inline/vector_io/milvus/config.py +29 -0
- llama_stack/providers/inline/vector_io/qdrant/__init__.py +20 -0
- llama_stack/providers/inline/vector_io/qdrant/config.py +29 -0
- llama_stack/providers/inline/vector_io/sqlite_vec/__init__.py +20 -0
- llama_stack/providers/inline/vector_io/sqlite_vec/config.py +26 -0
- llama_stack/providers/inline/vector_io/sqlite_vec/sqlite_vec.py +483 -0
- llama_stack/providers/registry/agents.py +16 -18
- llama_stack/providers/registry/batches.py +26 -0
- llama_stack/providers/registry/datasetio.py +49 -0
- llama_stack/providers/registry/eval.py +46 -0
- llama_stack/providers/registry/files.py +31 -0
- llama_stack/providers/registry/inference.py +273 -118
- llama_stack/providers/registry/post_training.py +69 -0
- llama_stack/providers/registry/safety.py +46 -41
- llama_stack/providers/registry/scoring.py +51 -0
- llama_stack/providers/registry/tool_runtime.py +87 -0
- llama_stack/providers/registry/vector_io.py +828 -0
- llama_stack/providers/remote/__init__.py +5 -0
- llama_stack/providers/remote/agents/__init__.py +5 -0
- llama_stack/providers/remote/datasetio/__init__.py +5 -0
- llama_stack/providers/{adapters/memory/chroma → remote/datasetio/huggingface}/__init__.py +7 -4
- llama_stack/providers/remote/datasetio/huggingface/config.py +23 -0
- llama_stack/providers/remote/datasetio/huggingface/huggingface.py +99 -0
- llama_stack/providers/remote/datasetio/nvidia/__init__.py +23 -0
- llama_stack/providers/remote/datasetio/nvidia/config.py +61 -0
- llama_stack/providers/remote/datasetio/nvidia/datasetio.py +116 -0
- llama_stack/providers/remote/eval/__init__.py +5 -0
- llama_stack/providers/remote/eval/nvidia/__init__.py +31 -0
- llama_stack/providers/remote/eval/nvidia/config.py +29 -0
- llama_stack/providers/remote/eval/nvidia/eval.py +162 -0
- llama_stack/providers/remote/files/s3/__init__.py +19 -0
- llama_stack/providers/remote/files/s3/config.py +42 -0
- llama_stack/providers/remote/files/s3/files.py +313 -0
- llama_stack/providers/remote/inference/__init__.py +5 -0
- llama_stack/providers/{adapters/safety/sample → remote/inference/anthropic}/__init__.py +4 -6
- llama_stack/providers/remote/inference/anthropic/anthropic.py +36 -0
- llama_stack/providers/remote/inference/anthropic/config.py +28 -0
- llama_stack/providers/{impls/meta_reference/telemetry → remote/inference/azure}/__init__.py +4 -4
- llama_stack/providers/remote/inference/azure/azure.py +25 -0
- llama_stack/providers/remote/inference/azure/config.py +61 -0
- llama_stack/providers/{adapters → remote}/inference/bedrock/__init__.py +18 -17
- llama_stack/providers/remote/inference/bedrock/bedrock.py +142 -0
- llama_stack/providers/{adapters/inference/sample → remote/inference/bedrock}/config.py +3 -4
- llama_stack/providers/remote/inference/bedrock/models.py +29 -0
- llama_stack/providers/remote/inference/cerebras/__init__.py +19 -0
- llama_stack/providers/remote/inference/cerebras/cerebras.py +28 -0
- llama_stack/providers/remote/inference/cerebras/config.py +30 -0
- llama_stack/providers/{adapters → remote}/inference/databricks/__init__.py +4 -5
- llama_stack/providers/remote/inference/databricks/config.py +37 -0
- llama_stack/providers/remote/inference/databricks/databricks.py +44 -0
- llama_stack/providers/{adapters → remote}/inference/fireworks/__init__.py +8 -4
- llama_stack/providers/remote/inference/fireworks/config.py +27 -0
- llama_stack/providers/remote/inference/fireworks/fireworks.py +27 -0
- llama_stack/providers/{adapters/memory/pgvector → remote/inference/gemini}/__init__.py +4 -4
- llama_stack/providers/remote/inference/gemini/config.py +28 -0
- llama_stack/providers/remote/inference/gemini/gemini.py +82 -0
- llama_stack/providers/remote/inference/groq/__init__.py +15 -0
- llama_stack/providers/remote/inference/groq/config.py +34 -0
- llama_stack/providers/remote/inference/groq/groq.py +18 -0
- llama_stack/providers/remote/inference/llama_openai_compat/__init__.py +15 -0
- llama_stack/providers/remote/inference/llama_openai_compat/config.py +34 -0
- llama_stack/providers/remote/inference/llama_openai_compat/llama.py +46 -0
- llama_stack/providers/remote/inference/nvidia/__init__.py +23 -0
- llama_stack/providers/remote/inference/nvidia/config.py +64 -0
- llama_stack/providers/remote/inference/nvidia/nvidia.py +61 -0
- llama_stack/providers/{adapters/safety/sample/config.py → remote/inference/nvidia/utils.py} +3 -4
- llama_stack/providers/{impls/vllm → remote/inference/ollama}/__init__.py +4 -6
- llama_stack/providers/remote/inference/ollama/config.py +25 -0
- llama_stack/providers/remote/inference/ollama/ollama.py +102 -0
- llama_stack/providers/{adapters/telemetry/opentelemetry → remote/inference/openai}/__init__.py +4 -4
- llama_stack/providers/remote/inference/openai/config.py +39 -0
- llama_stack/providers/remote/inference/openai/openai.py +38 -0
- llama_stack/providers/remote/inference/passthrough/__init__.py +23 -0
- llama_stack/providers/remote/inference/passthrough/config.py +34 -0
- llama_stack/providers/remote/inference/passthrough/passthrough.py +122 -0
- llama_stack/providers/remote/inference/runpod/__init__.py +16 -0
- llama_stack/providers/remote/inference/runpod/config.py +32 -0
- llama_stack/providers/remote/inference/runpod/runpod.py +42 -0
- llama_stack/providers/remote/inference/sambanova/__init__.py +16 -0
- llama_stack/providers/remote/inference/sambanova/config.py +34 -0
- llama_stack/providers/remote/inference/sambanova/sambanova.py +28 -0
- llama_stack/providers/{adapters → remote}/inference/tgi/__init__.py +3 -4
- llama_stack/providers/remote/inference/tgi/config.py +76 -0
- llama_stack/providers/remote/inference/tgi/tgi.py +85 -0
- llama_stack/providers/{adapters → remote}/inference/together/__init__.py +8 -4
- llama_stack/providers/remote/inference/together/config.py +27 -0
- llama_stack/providers/remote/inference/together/together.py +102 -0
- llama_stack/providers/remote/inference/vertexai/__init__.py +15 -0
- llama_stack/providers/remote/inference/vertexai/config.py +48 -0
- llama_stack/providers/remote/inference/vertexai/vertexai.py +54 -0
- llama_stack/providers/remote/inference/vllm/__init__.py +22 -0
- llama_stack/providers/remote/inference/vllm/config.py +59 -0
- llama_stack/providers/remote/inference/vllm/vllm.py +111 -0
- llama_stack/providers/remote/inference/watsonx/__init__.py +15 -0
- llama_stack/providers/remote/inference/watsonx/config.py +45 -0
- llama_stack/providers/remote/inference/watsonx/watsonx.py +336 -0
- llama_stack/providers/remote/post_training/__init__.py +5 -0
- llama_stack/providers/remote/post_training/nvidia/__init__.py +23 -0
- llama_stack/providers/remote/post_training/nvidia/config.py +113 -0
- llama_stack/providers/remote/post_training/nvidia/models.py +27 -0
- llama_stack/providers/remote/post_training/nvidia/post_training.py +430 -0
- llama_stack/providers/remote/post_training/nvidia/utils.py +63 -0
- llama_stack/providers/remote/safety/__init__.py +5 -0
- llama_stack/providers/remote/safety/bedrock/bedrock.py +111 -0
- llama_stack/providers/remote/safety/bedrock/config.py +14 -0
- llama_stack/providers/{adapters/inference/sample → remote/safety/nvidia}/__init__.py +5 -4
- llama_stack/providers/remote/safety/nvidia/config.py +40 -0
- llama_stack/providers/remote/safety/nvidia/nvidia.py +161 -0
- llama_stack/providers/{adapters/agents/sample → remote/safety/sambanova}/__init__.py +5 -4
- llama_stack/providers/remote/safety/sambanova/config.py +37 -0
- llama_stack/providers/remote/safety/sambanova/sambanova.py +98 -0
- llama_stack/providers/remote/tool_runtime/__init__.py +5 -0
- llama_stack/providers/remote/tool_runtime/bing_search/__init__.py +21 -0
- llama_stack/providers/remote/tool_runtime/bing_search/bing_search.py +112 -0
- llama_stack/providers/remote/tool_runtime/bing_search/config.py +22 -0
- llama_stack/providers/remote/tool_runtime/brave_search/__init__.py +20 -0
- llama_stack/providers/remote/tool_runtime/brave_search/brave_search.py +148 -0
- llama_stack/providers/remote/tool_runtime/brave_search/config.py +27 -0
- llama_stack/providers/remote/tool_runtime/model_context_protocol/__init__.py +15 -0
- llama_stack/providers/remote/tool_runtime/model_context_protocol/config.py +20 -0
- llama_stack/providers/remote/tool_runtime/model_context_protocol/model_context_protocol.py +73 -0
- llama_stack/providers/remote/tool_runtime/tavily_search/__init__.py +20 -0
- llama_stack/providers/remote/tool_runtime/tavily_search/config.py +27 -0
- llama_stack/providers/remote/tool_runtime/tavily_search/tavily_search.py +84 -0
- llama_stack/providers/remote/tool_runtime/wolfram_alpha/__init__.py +22 -0
- llama_stack/providers/remote/tool_runtime/wolfram_alpha/config.py +21 -0
- llama_stack/providers/remote/tool_runtime/wolfram_alpha/wolfram_alpha.py +140 -0
- llama_stack/providers/remote/vector_io/__init__.py +5 -0
- llama_stack/providers/remote/vector_io/chroma/__init__.py +17 -0
- llama_stack/providers/remote/vector_io/chroma/chroma.py +215 -0
- llama_stack/providers/remote/vector_io/chroma/config.py +28 -0
- llama_stack/providers/remote/vector_io/milvus/__init__.py +18 -0
- llama_stack/providers/remote/vector_io/milvus/config.py +35 -0
- llama_stack/providers/remote/vector_io/milvus/milvus.py +375 -0
- llama_stack/providers/remote/vector_io/pgvector/__init__.py +17 -0
- llama_stack/providers/remote/vector_io/pgvector/config.py +47 -0
- llama_stack/providers/remote/vector_io/pgvector/pgvector.py +460 -0
- llama_stack/providers/remote/vector_io/qdrant/__init__.py +17 -0
- llama_stack/providers/remote/vector_io/qdrant/config.py +37 -0
- llama_stack/providers/remote/vector_io/qdrant/qdrant.py +265 -0
- llama_stack/providers/remote/vector_io/weaviate/__init__.py +17 -0
- llama_stack/providers/remote/vector_io/weaviate/config.py +32 -0
- llama_stack/providers/remote/vector_io/weaviate/weaviate.py +393 -0
- llama_stack/providers/utils/bedrock/__init__.py +5 -0
- llama_stack/providers/utils/bedrock/client.py +74 -0
- llama_stack/providers/utils/bedrock/config.py +64 -0
- llama_stack/providers/utils/bedrock/refreshable_boto_session.py +112 -0
- llama_stack/providers/utils/common/__init__.py +5 -0
- llama_stack/providers/utils/common/data_schema_validator.py +103 -0
- llama_stack/providers/utils/datasetio/__init__.py +5 -0
- llama_stack/providers/utils/datasetio/url_utils.py +47 -0
- llama_stack/providers/utils/files/__init__.py +5 -0
- llama_stack/providers/utils/files/form_data.py +69 -0
- llama_stack/providers/utils/inference/__init__.py +8 -7
- llama_stack/providers/utils/inference/embedding_mixin.py +101 -0
- llama_stack/providers/utils/inference/inference_store.py +264 -0
- llama_stack/providers/utils/inference/litellm_openai_mixin.py +336 -0
- llama_stack/providers/utils/inference/model_registry.py +173 -23
- llama_stack/providers/utils/inference/openai_compat.py +1261 -49
- llama_stack/providers/utils/inference/openai_mixin.py +506 -0
- llama_stack/providers/utils/inference/prompt_adapter.py +365 -67
- llama_stack/providers/utils/kvstore/api.py +6 -6
- llama_stack/providers/utils/kvstore/config.py +28 -48
- llama_stack/providers/utils/kvstore/kvstore.py +61 -15
- llama_stack/providers/utils/kvstore/mongodb/__init__.py +9 -0
- llama_stack/providers/utils/kvstore/mongodb/mongodb.py +82 -0
- llama_stack/providers/utils/kvstore/postgres/__init__.py +7 -0
- llama_stack/providers/utils/kvstore/postgres/postgres.py +114 -0
- llama_stack/providers/utils/kvstore/redis/redis.py +33 -9
- llama_stack/providers/utils/kvstore/sqlite/config.py +2 -1
- llama_stack/providers/utils/kvstore/sqlite/sqlite.py +123 -22
- llama_stack/providers/utils/memory/file_utils.py +1 -1
- llama_stack/providers/utils/memory/openai_vector_store_mixin.py +1304 -0
- llama_stack/providers/utils/memory/vector_store.py +220 -82
- llama_stack/providers/utils/pagination.py +43 -0
- llama_stack/providers/utils/responses/__init__.py +5 -0
- llama_stack/providers/utils/responses/responses_store.py +292 -0
- llama_stack/providers/utils/scheduler.py +270 -0
- llama_stack/providers/utils/scoring/__init__.py +5 -0
- llama_stack/providers/utils/scoring/aggregation_utils.py +75 -0
- llama_stack/providers/utils/scoring/base_scoring_fn.py +114 -0
- llama_stack/providers/utils/scoring/basic_scoring_utils.py +26 -0
- llama_stack/providers/utils/sqlstore/__init__.py +5 -0
- llama_stack/providers/utils/sqlstore/api.py +128 -0
- llama_stack/providers/utils/sqlstore/authorized_sqlstore.py +319 -0
- llama_stack/providers/utils/sqlstore/sqlalchemy_sqlstore.py +343 -0
- llama_stack/providers/utils/sqlstore/sqlstore.py +70 -0
- llama_stack/providers/utils/telemetry/trace_protocol.py +142 -0
- llama_stack/providers/utils/telemetry/tracing.py +192 -53
- llama_stack/providers/utils/tools/__init__.py +5 -0
- llama_stack/providers/utils/tools/mcp.py +148 -0
- llama_stack/providers/utils/tools/ttl_dict.py +70 -0
- llama_stack/providers/utils/vector_io/__init__.py +5 -0
- llama_stack/providers/utils/vector_io/vector_utils.py +156 -0
- llama_stack/schema_utils.py +118 -0
- llama_stack/strong_typing/__init__.py +19 -0
- llama_stack/strong_typing/auxiliary.py +228 -0
- llama_stack/strong_typing/classdef.py +440 -0
- llama_stack/strong_typing/core.py +46 -0
- llama_stack/strong_typing/deserializer.py +877 -0
- llama_stack/strong_typing/docstring.py +409 -0
- llama_stack/strong_typing/exception.py +23 -0
- llama_stack/strong_typing/inspection.py +1085 -0
- llama_stack/strong_typing/mapping.py +40 -0
- llama_stack/strong_typing/name.py +182 -0
- llama_stack/strong_typing/py.typed +0 -0
- llama_stack/strong_typing/schema.py +792 -0
- llama_stack/strong_typing/serialization.py +97 -0
- llama_stack/strong_typing/serializer.py +500 -0
- llama_stack/strong_typing/slots.py +27 -0
- llama_stack/strong_typing/topological.py +89 -0
- llama_stack/testing/__init__.py +5 -0
- llama_stack/testing/api_recorder.py +956 -0
- llama_stack/ui/node_modules/flatted/python/flatted.py +149 -0
- llama_stack-0.3.4.dist-info/METADATA +261 -0
- llama_stack-0.3.4.dist-info/RECORD +625 -0
- {llama_stack-0.0.42.dist-info → llama_stack-0.3.4.dist-info}/WHEEL +1 -1
- llama_stack/apis/agents/client.py +0 -292
- llama_stack/apis/agents/event_logger.py +0 -184
- llama_stack/apis/batch_inference/batch_inference.py +0 -72
- llama_stack/apis/common/deployment_types.py +0 -31
- llama_stack/apis/dataset/dataset.py +0 -63
- llama_stack/apis/evals/evals.py +0 -122
- llama_stack/apis/inference/client.py +0 -197
- llama_stack/apis/inspect/client.py +0 -82
- llama_stack/apis/memory/client.py +0 -155
- llama_stack/apis/memory/memory.py +0 -65
- llama_stack/apis/memory_banks/__init__.py +0 -7
- llama_stack/apis/memory_banks/client.py +0 -101
- llama_stack/apis/memory_banks/memory_banks.py +0 -78
- llama_stack/apis/models/client.py +0 -83
- llama_stack/apis/reward_scoring/__init__.py +0 -7
- llama_stack/apis/reward_scoring/reward_scoring.py +0 -55
- llama_stack/apis/safety/client.py +0 -105
- llama_stack/apis/shields/client.py +0 -79
- llama_stack/cli/download.py +0 -340
- llama_stack/cli/model/describe.py +0 -82
- llama_stack/cli/model/download.py +0 -24
- llama_stack/cli/model/list.py +0 -62
- llama_stack/cli/model/model.py +0 -34
- llama_stack/cli/model/prompt_format.py +0 -112
- llama_stack/cli/model/safety_models.py +0 -52
- llama_stack/cli/stack/build.py +0 -299
- llama_stack/cli/stack/configure.py +0 -178
- llama_stack/distribution/build.py +0 -123
- llama_stack/distribution/build_conda_env.sh +0 -136
- llama_stack/distribution/build_container.sh +0 -142
- llama_stack/distribution/common.sh +0 -40
- llama_stack/distribution/configure_container.sh +0 -47
- llama_stack/distribution/datatypes.py +0 -139
- llama_stack/distribution/distribution.py +0 -58
- llama_stack/distribution/inspect.py +0 -67
- llama_stack/distribution/request_headers.py +0 -57
- llama_stack/distribution/resolver.py +0 -323
- llama_stack/distribution/routers/__init__.py +0 -48
- llama_stack/distribution/routers/routers.py +0 -158
- llama_stack/distribution/routers/routing_tables.py +0 -173
- llama_stack/distribution/server/endpoints.py +0 -48
- llama_stack/distribution/server/server.py +0 -343
- llama_stack/distribution/start_conda_env.sh +0 -42
- llama_stack/distribution/start_container.sh +0 -64
- llama_stack/distribution/templates/local-bedrock-conda-example-build.yaml +0 -10
- llama_stack/distribution/templates/local-build.yaml +0 -10
- llama_stack/distribution/templates/local-databricks-build.yaml +0 -10
- llama_stack/distribution/templates/local-fireworks-build.yaml +0 -10
- llama_stack/distribution/templates/local-hf-endpoint-build.yaml +0 -10
- llama_stack/distribution/templates/local-hf-serverless-build.yaml +0 -10
- llama_stack/distribution/templates/local-ollama-build.yaml +0 -10
- llama_stack/distribution/templates/local-tgi-build.yaml +0 -10
- llama_stack/distribution/templates/local-together-build.yaml +0 -10
- llama_stack/distribution/templates/local-vllm-build.yaml +0 -10
- llama_stack/distribution/utils/exec.py +0 -105
- llama_stack/providers/adapters/agents/sample/sample.py +0 -18
- llama_stack/providers/adapters/inference/bedrock/bedrock.py +0 -451
- llama_stack/providers/adapters/inference/bedrock/config.py +0 -55
- llama_stack/providers/adapters/inference/databricks/config.py +0 -21
- llama_stack/providers/adapters/inference/databricks/databricks.py +0 -125
- llama_stack/providers/adapters/inference/fireworks/config.py +0 -20
- llama_stack/providers/adapters/inference/fireworks/fireworks.py +0 -130
- llama_stack/providers/adapters/inference/ollama/__init__.py +0 -19
- llama_stack/providers/adapters/inference/ollama/ollama.py +0 -175
- llama_stack/providers/adapters/inference/sample/sample.py +0 -23
- llama_stack/providers/adapters/inference/tgi/config.py +0 -43
- llama_stack/providers/adapters/inference/tgi/tgi.py +0 -200
- llama_stack/providers/adapters/inference/together/config.py +0 -22
- llama_stack/providers/adapters/inference/together/together.py +0 -143
- llama_stack/providers/adapters/memory/chroma/chroma.py +0 -157
- llama_stack/providers/adapters/memory/pgvector/config.py +0 -17
- llama_stack/providers/adapters/memory/pgvector/pgvector.py +0 -211
- llama_stack/providers/adapters/memory/sample/sample.py +0 -23
- llama_stack/providers/adapters/memory/weaviate/__init__.py +0 -15
- llama_stack/providers/adapters/memory/weaviate/weaviate.py +0 -190
- llama_stack/providers/adapters/safety/bedrock/bedrock.py +0 -113
- llama_stack/providers/adapters/safety/bedrock/config.py +0 -16
- llama_stack/providers/adapters/safety/sample/sample.py +0 -23
- llama_stack/providers/adapters/safety/together/__init__.py +0 -18
- llama_stack/providers/adapters/safety/together/config.py +0 -26
- llama_stack/providers/adapters/safety/together/together.py +0 -101
- llama_stack/providers/adapters/telemetry/opentelemetry/config.py +0 -12
- llama_stack/providers/adapters/telemetry/opentelemetry/opentelemetry.py +0 -201
- llama_stack/providers/adapters/telemetry/sample/__init__.py +0 -17
- llama_stack/providers/adapters/telemetry/sample/config.py +0 -12
- llama_stack/providers/adapters/telemetry/sample/sample.py +0 -18
- llama_stack/providers/impls/meta_reference/agents/agent_instance.py +0 -844
- llama_stack/providers/impls/meta_reference/agents/agents.py +0 -161
- llama_stack/providers/impls/meta_reference/agents/persistence.py +0 -84
- llama_stack/providers/impls/meta_reference/agents/rag/context_retriever.py +0 -74
- llama_stack/providers/impls/meta_reference/agents/safety.py +0 -57
- llama_stack/providers/impls/meta_reference/agents/tests/code_execution.py +0 -93
- llama_stack/providers/impls/meta_reference/agents/tests/test_chat_agent.py +0 -305
- llama_stack/providers/impls/meta_reference/agents/tools/base.py +0 -20
- llama_stack/providers/impls/meta_reference/agents/tools/builtin.py +0 -375
- llama_stack/providers/impls/meta_reference/agents/tools/ipython_tool/code_env_prefix.py +0 -133
- llama_stack/providers/impls/meta_reference/agents/tools/ipython_tool/code_execution.py +0 -256
- llama_stack/providers/impls/meta_reference/agents/tools/ipython_tool/matplotlib_custom_backend.py +0 -87
- llama_stack/providers/impls/meta_reference/agents/tools/ipython_tool/utils.py +0 -21
- llama_stack/providers/impls/meta_reference/agents/tools/safety.py +0 -43
- llama_stack/providers/impls/meta_reference/codeshield/code_scanner.py +0 -58
- llama_stack/providers/impls/meta_reference/inference/config.py +0 -45
- llama_stack/providers/impls/meta_reference/inference/generation.py +0 -376
- llama_stack/providers/impls/meta_reference/inference/inference.py +0 -280
- llama_stack/providers/impls/meta_reference/inference/model_parallel.py +0 -99
- llama_stack/providers/impls/meta_reference/inference/quantization/fp8_impls.py +0 -184
- llama_stack/providers/impls/meta_reference/inference/quantization/fp8_txest_disabled.py +0 -76
- llama_stack/providers/impls/meta_reference/inference/quantization/loader.py +0 -97
- llama_stack/providers/impls/meta_reference/inference/quantization/scripts/quantize_checkpoint.py +0 -161
- llama_stack/providers/impls/meta_reference/memory/__init__.py +0 -19
- llama_stack/providers/impls/meta_reference/memory/faiss.py +0 -113
- llama_stack/providers/impls/meta_reference/safety/__init__.py +0 -17
- llama_stack/providers/impls/meta_reference/safety/base.py +0 -57
- llama_stack/providers/impls/meta_reference/safety/config.py +0 -48
- llama_stack/providers/impls/meta_reference/safety/llama_guard.py +0 -268
- llama_stack/providers/impls/meta_reference/safety/prompt_guard.py +0 -145
- llama_stack/providers/impls/meta_reference/safety/safety.py +0 -112
- llama_stack/providers/impls/meta_reference/telemetry/console.py +0 -89
- llama_stack/providers/impls/vllm/config.py +0 -35
- llama_stack/providers/impls/vllm/vllm.py +0 -241
- llama_stack/providers/registry/memory.py +0 -78
- llama_stack/providers/registry/telemetry.py +0 -44
- llama_stack/providers/tests/agents/test_agents.py +0 -210
- llama_stack/providers/tests/inference/test_inference.py +0 -257
- llama_stack/providers/tests/inference/test_prompt_adapter.py +0 -126
- llama_stack/providers/tests/memory/test_memory.py +0 -136
- llama_stack/providers/tests/resolver.py +0 -100
- llama_stack/providers/tests/safety/test_safety.py +0 -77
- llama_stack-0.0.42.dist-info/METADATA +0 -137
- llama_stack-0.0.42.dist-info/RECORD +0 -256
- /llama_stack/{distribution → core}/__init__.py +0 -0
- /llama_stack/{distribution/server → core/access_control}/__init__.py +0 -0
- /llama_stack/{distribution/utils → core/conversations}/__init__.py +0 -0
- /llama_stack/{providers/adapters → core/prompts}/__init__.py +0 -0
- /llama_stack/{providers/adapters/agents → core/routing_tables}/__init__.py +0 -0
- /llama_stack/{providers/adapters/inference → core/server}/__init__.py +0 -0
- /llama_stack/{providers/adapters/memory → core/storage}/__init__.py +0 -0
- /llama_stack/{providers/adapters/safety → core/ui}/__init__.py +0 -0
- /llama_stack/{providers/adapters/telemetry → core/ui/modules}/__init__.py +0 -0
- /llama_stack/{providers/impls → core/ui/page}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference → core/ui/page/distribution}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference/agents/rag → core/ui/page/evaluations}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference/agents/tests → core/ui/page/playground}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference/agents/tools → core/utils}/__init__.py +0 -0
- /llama_stack/{distribution → core}/utils/dynamic.py +0 -0
- /llama_stack/{distribution → core}/utils/serialize.py +0 -0
- /llama_stack/{providers/impls/meta_reference/agents/tools/ipython_tool → distributions}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference/inference/quantization → models}/__init__.py +0 -0
- /llama_stack/{providers/impls/meta_reference/inference/quantization/scripts → models/llama}/__init__.py +0 -0
- /llama_stack/{providers/tests → models/llama/llama3}/__init__.py +0 -0
- /llama_stack/{providers/tests/agents → models/llama/llama3/quantization}/__init__.py +0 -0
- /llama_stack/{providers/tests/inference → models/llama/llama3_2}/__init__.py +0 -0
- /llama_stack/{providers/tests/memory → models/llama/llama3_3}/__init__.py +0 -0
- /llama_stack/{providers/tests/safety → models/llama/llama4}/__init__.py +0 -0
- /llama_stack/{scripts → models/llama/llama4/prompt_templates}/__init__.py +0 -0
- /llama_stack/providers/{adapters → remote}/safety/bedrock/__init__.py +0 -0
- {llama_stack-0.0.42.dist-info → llama_stack-0.3.4.dist-info}/entry_points.txt +0 -0
- {llama_stack-0.0.42.dist-info → llama_stack-0.3.4.dist-info/licenses}/LICENSE +0 -0
- {llama_stack-0.0.42.dist-info → llama_stack-0.3.4.dist-info}/top_level.txt +0 -0
|
@@ -1,136 +0,0 @@
|
|
|
1
|
-
#!/bin/bash
|
|
2
|
-
|
|
3
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
4
|
-
# All rights reserved.
|
|
5
|
-
#
|
|
6
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
7
|
-
# the root directory of this source tree.
|
|
8
|
-
|
|
9
|
-
LLAMA_MODELS_DIR=${LLAMA_MODELS_DIR:-}
|
|
10
|
-
LLAMA_STACK_DIR=${LLAMA_STACK_DIR:-}
|
|
11
|
-
TEST_PYPI_VERSION=${TEST_PYPI_VERSION:-}
|
|
12
|
-
|
|
13
|
-
if [ -n "$LLAMA_STACK_DIR" ]; then
|
|
14
|
-
echo "Using llama-stack-dir=$LLAMA_STACK_DIR"
|
|
15
|
-
fi
|
|
16
|
-
if [ -n "$LLAMA_MODELS_DIR" ]; then
|
|
17
|
-
echo "Using llama-models-dir=$LLAMA_MODELS_DIR"
|
|
18
|
-
fi
|
|
19
|
-
|
|
20
|
-
if [ "$#" -lt 3 ]; then
|
|
21
|
-
echo "Usage: $0 <distribution_type> <build_name> <build_file_path> <pip_dependencies> [<special_pip_deps>]" >&2
|
|
22
|
-
echo "Example: $0 <distribution_type> mybuild ./my-stack-build.yaml 'numpy pandas scipy'" >&2
|
|
23
|
-
exit 1
|
|
24
|
-
fi
|
|
25
|
-
|
|
26
|
-
special_pip_deps="$4"
|
|
27
|
-
|
|
28
|
-
set -euo pipefail
|
|
29
|
-
|
|
30
|
-
build_name="$1"
|
|
31
|
-
env_name="llamastack-$build_name"
|
|
32
|
-
build_file_path="$2"
|
|
33
|
-
pip_dependencies="$3"
|
|
34
|
-
|
|
35
|
-
# Define color codes
|
|
36
|
-
RED='\033[0;31m'
|
|
37
|
-
GREEN='\033[0;32m'
|
|
38
|
-
NC='\033[0m' # No Color
|
|
39
|
-
|
|
40
|
-
# this is set if we actually create a new conda in which case we need to clean up
|
|
41
|
-
ENVNAME=""
|
|
42
|
-
|
|
43
|
-
SCRIPT_DIR=$(dirname "$(readlink -f "$0")")
|
|
44
|
-
source "$SCRIPT_DIR/common.sh"
|
|
45
|
-
|
|
46
|
-
ensure_conda_env_python310() {
|
|
47
|
-
local env_name="$1"
|
|
48
|
-
local pip_dependencies="$2"
|
|
49
|
-
local special_pip_deps="$3"
|
|
50
|
-
local python_version="3.10"
|
|
51
|
-
|
|
52
|
-
# Check if conda command is available
|
|
53
|
-
if ! command -v conda &>/dev/null; then
|
|
54
|
-
printf "${RED}Error: conda command not found. Is Conda installed and in your PATH?${NC}" >&2
|
|
55
|
-
exit 1
|
|
56
|
-
fi
|
|
57
|
-
|
|
58
|
-
# Check if the environment exists
|
|
59
|
-
if conda env list | grep -q "^${env_name} "; then
|
|
60
|
-
printf "Conda environment '${env_name}' exists. Checking Python version...\n"
|
|
61
|
-
|
|
62
|
-
# Check Python version in the environment
|
|
63
|
-
current_version=$(conda run -n "${env_name}" python --version 2>&1 | cut -d' ' -f2 | cut -d'.' -f1,2)
|
|
64
|
-
|
|
65
|
-
if [ "$current_version" = "$python_version" ]; then
|
|
66
|
-
printf "Environment '${env_name}' already has Python ${python_version}. No action needed.\n"
|
|
67
|
-
else
|
|
68
|
-
printf "Updating environment '${env_name}' to Python ${python_version}...\n"
|
|
69
|
-
conda install -n "${env_name}" python="${python_version}" -y
|
|
70
|
-
fi
|
|
71
|
-
else
|
|
72
|
-
printf "Conda environment '${env_name}' does not exist. Creating with Python ${python_version}...\n"
|
|
73
|
-
conda create -n "${env_name}" python="${python_version}" -y
|
|
74
|
-
|
|
75
|
-
ENVNAME="${env_name}"
|
|
76
|
-
# setup_cleanup_handlers
|
|
77
|
-
fi
|
|
78
|
-
|
|
79
|
-
eval "$(conda shell.bash hook)"
|
|
80
|
-
conda deactivate && conda activate "${env_name}"
|
|
81
|
-
|
|
82
|
-
if [ -n "$TEST_PYPI_VERSION" ]; then
|
|
83
|
-
# these packages are damaged in test-pypi, so install them first
|
|
84
|
-
$CONDA_PREFIX/bin/pip install fastapi libcst
|
|
85
|
-
$CONDA_PREFIX/bin/pip install --extra-index-url https://test.pypi.org/simple/ \
|
|
86
|
-
llama-models==$TEST_PYPI_VERSION llama-stack==$TEST_PYPI_VERSION \
|
|
87
|
-
$pip_dependencies
|
|
88
|
-
if [ -n "$special_pip_deps" ]; then
|
|
89
|
-
IFS='#' read -ra parts <<<"$special_pip_deps"
|
|
90
|
-
for part in "${parts[@]}"; do
|
|
91
|
-
echo "$part"
|
|
92
|
-
$CONDA_PREFIX/bin/pip install $part
|
|
93
|
-
done
|
|
94
|
-
fi
|
|
95
|
-
else
|
|
96
|
-
# Re-installing llama-stack in the new conda environment
|
|
97
|
-
if [ -n "$LLAMA_STACK_DIR" ]; then
|
|
98
|
-
if [ ! -d "$LLAMA_STACK_DIR" ]; then
|
|
99
|
-
printf "${RED}Warning: LLAMA_STACK_DIR is set but directory does not exist: $LLAMA_STACK_DIR${NC}\n" >&2
|
|
100
|
-
exit 1
|
|
101
|
-
fi
|
|
102
|
-
|
|
103
|
-
printf "Installing from LLAMA_STACK_DIR: $LLAMA_STACK_DIR\n"
|
|
104
|
-
$CONDA_PREFIX/bin/pip install --no-cache-dir -e "$LLAMA_STACK_DIR"
|
|
105
|
-
else
|
|
106
|
-
$CONDA_PREFIX/bin/pip install --no-cache-dir llama-stack
|
|
107
|
-
fi
|
|
108
|
-
|
|
109
|
-
if [ -n "$LLAMA_MODELS_DIR" ]; then
|
|
110
|
-
if [ ! -d "$LLAMA_MODELS_DIR" ]; then
|
|
111
|
-
printf "${RED}Warning: LLAMA_MODELS_DIR is set but directory does not exist: $LLAMA_MODELS_DIR${NC}\n" >&2
|
|
112
|
-
exit 1
|
|
113
|
-
fi
|
|
114
|
-
|
|
115
|
-
printf "Installing from LLAMA_MODELS_DIR: $LLAMA_MODELS_DIR\n"
|
|
116
|
-
$CONDA_PREFIX/bin/pip uninstall -y llama-models
|
|
117
|
-
$CONDA_PREFIX/bin/pip install --no-cache-dir -e "$LLAMA_MODELS_DIR"
|
|
118
|
-
fi
|
|
119
|
-
|
|
120
|
-
# Install pip dependencies
|
|
121
|
-
printf "Installing pip dependencies\n"
|
|
122
|
-
$CONDA_PREFIX/bin/pip install $pip_dependencies
|
|
123
|
-
if [ -n "$special_pip_deps" ]; then
|
|
124
|
-
IFS='#' read -ra parts <<<"$special_pip_deps"
|
|
125
|
-
for part in "${parts[@]}"; do
|
|
126
|
-
echo "$part"
|
|
127
|
-
$CONDA_PREFIX/bin/pip install $part
|
|
128
|
-
done
|
|
129
|
-
fi
|
|
130
|
-
fi
|
|
131
|
-
|
|
132
|
-
mv $build_file_path $CONDA_PREFIX/
|
|
133
|
-
echo "Build spec configuration saved at $CONDA_PREFIX/$build_name-build.yaml"
|
|
134
|
-
}
|
|
135
|
-
|
|
136
|
-
ensure_conda_env_python310 "$env_name" "$pip_dependencies" "$special_pip_deps"
|
|
@@ -1,142 +0,0 @@
|
|
|
1
|
-
#!/bin/bash
|
|
2
|
-
|
|
3
|
-
LLAMA_MODELS_DIR=${LLAMA_MODELS_DIR:-}
|
|
4
|
-
LLAMA_STACK_DIR=${LLAMA_STACK_DIR:-}
|
|
5
|
-
TEST_PYPI_VERSION=${TEST_PYPI_VERSION:-}
|
|
6
|
-
|
|
7
|
-
if [ "$#" -lt 4 ]; then
|
|
8
|
-
echo "Usage: $0 <build_name> <docker_base> <pip_dependencies> [<special_pip_deps>]" >&2
|
|
9
|
-
echo "Example: $0 my-fastapi-app python:3.9-slim 'fastapi uvicorn' " >&2
|
|
10
|
-
exit 1
|
|
11
|
-
fi
|
|
12
|
-
|
|
13
|
-
special_pip_deps="$6"
|
|
14
|
-
|
|
15
|
-
set -euo pipefail
|
|
16
|
-
|
|
17
|
-
build_name="$1"
|
|
18
|
-
image_name="llamastack-$build_name"
|
|
19
|
-
docker_base=$2
|
|
20
|
-
build_file_path=$3
|
|
21
|
-
host_build_dir=$4
|
|
22
|
-
pip_dependencies=$5
|
|
23
|
-
|
|
24
|
-
# Define color codes
|
|
25
|
-
RED='\033[0;31m'
|
|
26
|
-
GREEN='\033[0;32m'
|
|
27
|
-
NC='\033[0m' # No Color
|
|
28
|
-
|
|
29
|
-
SCRIPT_DIR=$(dirname "$(readlink -f "$0")")
|
|
30
|
-
REPO_DIR=$(dirname $(dirname "$SCRIPT_DIR"))
|
|
31
|
-
DOCKER_BINARY=${DOCKER_BINARY:-docker}
|
|
32
|
-
DOCKER_OPTS=${DOCKER_OPTS:-}
|
|
33
|
-
REPO_CONFIGS_DIR="$REPO_DIR/tmp/configs"
|
|
34
|
-
|
|
35
|
-
TEMP_DIR=$(mktemp -d)
|
|
36
|
-
|
|
37
|
-
llama stack configure $build_file_path
|
|
38
|
-
cp $host_build_dir/$build_name-run.yaml $REPO_CONFIGS_DIR
|
|
39
|
-
|
|
40
|
-
add_to_docker() {
|
|
41
|
-
local input
|
|
42
|
-
output_file="$TEMP_DIR/Dockerfile"
|
|
43
|
-
if [ -t 0 ]; then
|
|
44
|
-
printf '%s\n' "$1" >>"$output_file"
|
|
45
|
-
else
|
|
46
|
-
# If stdin is not a terminal, read from it (heredoc)
|
|
47
|
-
cat >>"$output_file"
|
|
48
|
-
fi
|
|
49
|
-
}
|
|
50
|
-
|
|
51
|
-
add_to_docker <<EOF
|
|
52
|
-
FROM $docker_base
|
|
53
|
-
WORKDIR /app
|
|
54
|
-
|
|
55
|
-
RUN apt-get update && apt-get install -y \
|
|
56
|
-
iputils-ping net-tools iproute2 dnsutils telnet \
|
|
57
|
-
curl wget telnet \
|
|
58
|
-
procps psmisc lsof \
|
|
59
|
-
traceroute \
|
|
60
|
-
bubblewrap \
|
|
61
|
-
&& rm -rf /var/lib/apt/lists/*
|
|
62
|
-
|
|
63
|
-
EOF
|
|
64
|
-
|
|
65
|
-
stack_mount="/app/llama-stack-source"
|
|
66
|
-
models_mount="/app/llama-models-source"
|
|
67
|
-
|
|
68
|
-
if [ -n "$LLAMA_STACK_DIR" ]; then
|
|
69
|
-
if [ ! -d "$LLAMA_STACK_DIR" ]; then
|
|
70
|
-
echo "${RED}Warning: LLAMA_STACK_DIR is set but directory does not exist: $LLAMA_STACK_DIR${NC}" >&2
|
|
71
|
-
exit 1
|
|
72
|
-
fi
|
|
73
|
-
|
|
74
|
-
# Install in editable format. We will mount the source code into the container
|
|
75
|
-
# so that changes will be reflected in the container without having to do a
|
|
76
|
-
# rebuild. This is just for development convenience.
|
|
77
|
-
add_to_docker "RUN pip install -e $stack_mount"
|
|
78
|
-
else
|
|
79
|
-
add_to_docker "RUN pip install llama-stack"
|
|
80
|
-
fi
|
|
81
|
-
|
|
82
|
-
if [ -n "$LLAMA_MODELS_DIR" ]; then
|
|
83
|
-
if [ ! -d "$LLAMA_MODELS_DIR" ]; then
|
|
84
|
-
echo "${RED}Warning: LLAMA_MODELS_DIR is set but directory does not exist: $LLAMA_MODELS_DIR${NC}" >&2
|
|
85
|
-
exit 1
|
|
86
|
-
fi
|
|
87
|
-
|
|
88
|
-
add_to_docker <<EOF
|
|
89
|
-
RUN pip uninstall -y llama-models
|
|
90
|
-
RUN pip install $models_mount
|
|
91
|
-
|
|
92
|
-
EOF
|
|
93
|
-
fi
|
|
94
|
-
|
|
95
|
-
if [ -n "$pip_dependencies" ]; then
|
|
96
|
-
add_to_docker "RUN pip install $pip_dependencies"
|
|
97
|
-
fi
|
|
98
|
-
|
|
99
|
-
if [ -n "$special_pip_deps" ]; then
|
|
100
|
-
IFS='#' read -ra parts <<< "$special_pip_deps"
|
|
101
|
-
for part in "${parts[@]}"; do
|
|
102
|
-
add_to_docker "RUN pip install $part"
|
|
103
|
-
done
|
|
104
|
-
fi
|
|
105
|
-
|
|
106
|
-
add_to_docker <<EOF
|
|
107
|
-
|
|
108
|
-
# This would be good in production but for debugging flexibility lets not add it right now
|
|
109
|
-
# We need a more solid production ready entrypoint.sh anyway
|
|
110
|
-
#
|
|
111
|
-
ENTRYPOINT ["python", "-m", "llama_stack.distribution.server.server"]
|
|
112
|
-
|
|
113
|
-
EOF
|
|
114
|
-
|
|
115
|
-
add_to_docker "ADD tmp/configs/$(basename "$build_file_path") ./llamastack-build.yaml"
|
|
116
|
-
add_to_docker "ADD tmp/configs/$build_name-run.yaml ./llamastack-run.yaml"
|
|
117
|
-
|
|
118
|
-
printf "Dockerfile created successfully in $TEMP_DIR/Dockerfile"
|
|
119
|
-
cat $TEMP_DIR/Dockerfile
|
|
120
|
-
printf "\n"
|
|
121
|
-
|
|
122
|
-
mounts=""
|
|
123
|
-
if [ -n "$LLAMA_STACK_DIR" ]; then
|
|
124
|
-
mounts="$mounts -v $(readlink -f $LLAMA_STACK_DIR):$stack_mount"
|
|
125
|
-
fi
|
|
126
|
-
if [ -n "$LLAMA_MODELS_DIR" ]; then
|
|
127
|
-
mounts="$mounts -v $(readlink -f $LLAMA_MODELS_DIR):$models_mount"
|
|
128
|
-
fi
|
|
129
|
-
|
|
130
|
-
if command -v selinuxenabled &> /dev/null && selinuxenabled; then
|
|
131
|
-
# Disable SELinux labels -- we don't want to relabel the llama-stack source dir
|
|
132
|
-
DOCKER_OPTS="$DOCKER_OPTS --security-opt label=disable"
|
|
133
|
-
fi
|
|
134
|
-
|
|
135
|
-
set -x
|
|
136
|
-
$DOCKER_BINARY build $DOCKER_OPTS -t $image_name -f "$TEMP_DIR/Dockerfile" "$REPO_DIR" $mounts
|
|
137
|
-
|
|
138
|
-
# clean up tmp/configs
|
|
139
|
-
rm -rf $REPO_CONFIGS_DIR
|
|
140
|
-
set +x
|
|
141
|
-
|
|
142
|
-
echo "Success! You can run it with: $DOCKER_BINARY $DOCKER_OPTS run -p 5000:5000 $image_name"
|
|
@@ -1,40 +0,0 @@
|
|
|
1
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
2
|
-
# All rights reserved.
|
|
3
|
-
#
|
|
4
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
5
|
-
# the root directory of this source tree.
|
|
6
|
-
|
|
7
|
-
cleanup() {
|
|
8
|
-
envname="$1"
|
|
9
|
-
|
|
10
|
-
set +x
|
|
11
|
-
echo "Cleaning up..."
|
|
12
|
-
conda deactivate
|
|
13
|
-
conda env remove --name $envname -y
|
|
14
|
-
}
|
|
15
|
-
|
|
16
|
-
handle_int() {
|
|
17
|
-
if [ -n $ENVNAME ]; then
|
|
18
|
-
cleanup $ENVNAME
|
|
19
|
-
fi
|
|
20
|
-
exit 1
|
|
21
|
-
}
|
|
22
|
-
|
|
23
|
-
handle_exit() {
|
|
24
|
-
if [ $? -ne 0 ]; then
|
|
25
|
-
echo -e "\033[1;31mABORTING.\033[0m"
|
|
26
|
-
if [ -n $ENVNAME ]; then
|
|
27
|
-
cleanup $ENVNAME
|
|
28
|
-
fi
|
|
29
|
-
fi
|
|
30
|
-
}
|
|
31
|
-
|
|
32
|
-
setup_cleanup_handlers() {
|
|
33
|
-
trap handle_int INT
|
|
34
|
-
trap handle_exit EXIT
|
|
35
|
-
|
|
36
|
-
__conda_setup="$('conda' 'shell.bash' 'hook' 2>/dev/null)"
|
|
37
|
-
eval "$__conda_setup"
|
|
38
|
-
|
|
39
|
-
conda deactivate
|
|
40
|
-
}
|
|
@@ -1,47 +0,0 @@
|
|
|
1
|
-
#!/bin/bash
|
|
2
|
-
|
|
3
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
4
|
-
# All rights reserved.
|
|
5
|
-
#
|
|
6
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
7
|
-
# the root directory of this source tree.
|
|
8
|
-
|
|
9
|
-
DOCKER_BINARY=${DOCKER_BINARY:-docker}
|
|
10
|
-
DOCKER_OPTS=${DOCKER_OPTS:-}
|
|
11
|
-
LLAMA_STACK_DIR=${LLAMA_STACK_DIR:-}
|
|
12
|
-
|
|
13
|
-
set -euo pipefail
|
|
14
|
-
|
|
15
|
-
error_handler() {
|
|
16
|
-
echo "Error occurred in script at line: ${1}" >&2
|
|
17
|
-
exit 1
|
|
18
|
-
}
|
|
19
|
-
|
|
20
|
-
trap 'error_handler ${LINENO}' ERR
|
|
21
|
-
|
|
22
|
-
if [ $# -lt 2 ]; then
|
|
23
|
-
echo "Usage: $0 <container name> <build file path>"
|
|
24
|
-
exit 1
|
|
25
|
-
fi
|
|
26
|
-
|
|
27
|
-
docker_image="$1"
|
|
28
|
-
host_build_dir="$2"
|
|
29
|
-
container_build_dir="/app/builds"
|
|
30
|
-
|
|
31
|
-
if command -v selinuxenabled &> /dev/null && selinuxenabled; then
|
|
32
|
-
# Disable SELinux labels
|
|
33
|
-
DOCKER_OPTS="$DOCKER_OPTS --security-opt label=disable"
|
|
34
|
-
fi
|
|
35
|
-
|
|
36
|
-
mounts=""
|
|
37
|
-
if [ -n "$LLAMA_STACK_DIR" ]; then
|
|
38
|
-
mounts="$mounts -v $(readlink -f $LLAMA_STACK_DIR):/app/llama-stack-source"
|
|
39
|
-
fi
|
|
40
|
-
|
|
41
|
-
set -x
|
|
42
|
-
$DOCKER_BINARY run $DOCKER_OPTS -it \
|
|
43
|
-
--entrypoint "/usr/local/bin/llama" \
|
|
44
|
-
-v $host_build_dir:$container_build_dir \
|
|
45
|
-
$mounts \
|
|
46
|
-
$docker_image \
|
|
47
|
-
stack configure ./llamastack-build.yaml --output-dir $container_build_dir
|
|
@@ -1,139 +0,0 @@
|
|
|
1
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
2
|
-
# All rights reserved.
|
|
3
|
-
#
|
|
4
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
5
|
-
# the root directory of this source tree.
|
|
6
|
-
|
|
7
|
-
from datetime import datetime
|
|
8
|
-
|
|
9
|
-
from typing import Dict, List, Optional, Union
|
|
10
|
-
|
|
11
|
-
from pydantic import BaseModel, Field
|
|
12
|
-
|
|
13
|
-
from llama_stack.providers.datatypes import * # noqa: F403
|
|
14
|
-
from llama_stack.apis.models import * # noqa: F403
|
|
15
|
-
from llama_stack.apis.shields import * # noqa: F403
|
|
16
|
-
from llama_stack.apis.memory_banks import * # noqa: F403
|
|
17
|
-
from llama_stack.apis.inference import Inference
|
|
18
|
-
from llama_stack.apis.memory import Memory
|
|
19
|
-
from llama_stack.apis.safety import Safety
|
|
20
|
-
|
|
21
|
-
|
|
22
|
-
LLAMA_STACK_BUILD_CONFIG_VERSION = "2"
|
|
23
|
-
LLAMA_STACK_RUN_CONFIG_VERSION = "2"
|
|
24
|
-
|
|
25
|
-
|
|
26
|
-
RoutingKey = Union[str, List[str]]
|
|
27
|
-
|
|
28
|
-
|
|
29
|
-
RoutableObject = Union[
|
|
30
|
-
ModelDef,
|
|
31
|
-
ShieldDef,
|
|
32
|
-
MemoryBankDef,
|
|
33
|
-
]
|
|
34
|
-
|
|
35
|
-
RoutableObjectWithProvider = Union[
|
|
36
|
-
ModelDefWithProvider,
|
|
37
|
-
ShieldDefWithProvider,
|
|
38
|
-
MemoryBankDefWithProvider,
|
|
39
|
-
]
|
|
40
|
-
|
|
41
|
-
RoutedProtocol = Union[
|
|
42
|
-
Inference,
|
|
43
|
-
Safety,
|
|
44
|
-
Memory,
|
|
45
|
-
]
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
# Example: /inference, /safety
|
|
49
|
-
class AutoRoutedProviderSpec(ProviderSpec):
|
|
50
|
-
provider_type: str = "router"
|
|
51
|
-
config_class: str = ""
|
|
52
|
-
|
|
53
|
-
docker_image: Optional[str] = None
|
|
54
|
-
routing_table_api: Api
|
|
55
|
-
module: str
|
|
56
|
-
provider_data_validator: Optional[str] = Field(
|
|
57
|
-
default=None,
|
|
58
|
-
)
|
|
59
|
-
|
|
60
|
-
@property
|
|
61
|
-
def pip_packages(self) -> List[str]:
|
|
62
|
-
raise AssertionError("Should not be called on AutoRoutedProviderSpec")
|
|
63
|
-
|
|
64
|
-
|
|
65
|
-
# Example: /models, /shields
|
|
66
|
-
class RoutingTableProviderSpec(ProviderSpec):
|
|
67
|
-
provider_type: str = "routing_table"
|
|
68
|
-
config_class: str = ""
|
|
69
|
-
docker_image: Optional[str] = None
|
|
70
|
-
|
|
71
|
-
router_api: Api
|
|
72
|
-
module: str
|
|
73
|
-
pip_packages: List[str] = Field(default_factory=list)
|
|
74
|
-
|
|
75
|
-
|
|
76
|
-
class DistributionSpec(BaseModel):
|
|
77
|
-
description: Optional[str] = Field(
|
|
78
|
-
default="",
|
|
79
|
-
description="Description of the distribution",
|
|
80
|
-
)
|
|
81
|
-
docker_image: Optional[str] = None
|
|
82
|
-
providers: Dict[str, Union[str, List[str]]] = Field(
|
|
83
|
-
default_factory=dict,
|
|
84
|
-
description="""
|
|
85
|
-
Provider Types for each of the APIs provided by this distribution. If you
|
|
86
|
-
select multiple providers, you should provide an appropriate 'routing_map'
|
|
87
|
-
in the runtime configuration to help route to the correct provider.""",
|
|
88
|
-
)
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
class Provider(BaseModel):
|
|
92
|
-
provider_id: str
|
|
93
|
-
provider_type: str
|
|
94
|
-
config: Dict[str, Any]
|
|
95
|
-
|
|
96
|
-
|
|
97
|
-
class StackRunConfig(BaseModel):
|
|
98
|
-
version: str = LLAMA_STACK_RUN_CONFIG_VERSION
|
|
99
|
-
built_at: datetime
|
|
100
|
-
|
|
101
|
-
image_name: str = Field(
|
|
102
|
-
...,
|
|
103
|
-
description="""
|
|
104
|
-
Reference to the distribution this package refers to. For unregistered (adhoc) packages,
|
|
105
|
-
this could be just a hash
|
|
106
|
-
""",
|
|
107
|
-
)
|
|
108
|
-
docker_image: Optional[str] = Field(
|
|
109
|
-
default=None,
|
|
110
|
-
description="Reference to the docker image if this package refers to a container",
|
|
111
|
-
)
|
|
112
|
-
conda_env: Optional[str] = Field(
|
|
113
|
-
default=None,
|
|
114
|
-
description="Reference to the conda environment if this package refers to a conda environment",
|
|
115
|
-
)
|
|
116
|
-
apis: List[str] = Field(
|
|
117
|
-
default_factory=list,
|
|
118
|
-
description="""
|
|
119
|
-
The list of APIs to serve. If not specified, all APIs specified in the provider_map will be served""",
|
|
120
|
-
)
|
|
121
|
-
|
|
122
|
-
providers: Dict[str, List[Provider]] = Field(
|
|
123
|
-
description="""
|
|
124
|
-
One or more providers to use for each API. The same provider_type (e.g., meta-reference)
|
|
125
|
-
can be instantiated multiple times (with different configs) if necessary.
|
|
126
|
-
""",
|
|
127
|
-
)
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
class BuildConfig(BaseModel):
|
|
131
|
-
version: str = LLAMA_STACK_BUILD_CONFIG_VERSION
|
|
132
|
-
name: str
|
|
133
|
-
distribution_spec: DistributionSpec = Field(
|
|
134
|
-
description="The distribution spec to build including API providers. "
|
|
135
|
-
)
|
|
136
|
-
image_type: str = Field(
|
|
137
|
-
default="conda",
|
|
138
|
-
description="Type of package to build (conda | container)",
|
|
139
|
-
)
|
|
@@ -1,58 +0,0 @@
|
|
|
1
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
2
|
-
# All rights reserved.
|
|
3
|
-
#
|
|
4
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
5
|
-
# the root directory of this source tree.
|
|
6
|
-
|
|
7
|
-
import importlib
|
|
8
|
-
from typing import Dict, List
|
|
9
|
-
|
|
10
|
-
from pydantic import BaseModel
|
|
11
|
-
|
|
12
|
-
from llama_stack.providers.datatypes import Api, ProviderSpec, remote_provider_spec
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
def stack_apis() -> List[Api]:
|
|
16
|
-
return [v for v in Api]
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
class AutoRoutedApiInfo(BaseModel):
|
|
20
|
-
routing_table_api: Api
|
|
21
|
-
router_api: Api
|
|
22
|
-
|
|
23
|
-
|
|
24
|
-
def builtin_automatically_routed_apis() -> List[AutoRoutedApiInfo]:
|
|
25
|
-
return [
|
|
26
|
-
AutoRoutedApiInfo(
|
|
27
|
-
routing_table_api=Api.models,
|
|
28
|
-
router_api=Api.inference,
|
|
29
|
-
),
|
|
30
|
-
AutoRoutedApiInfo(
|
|
31
|
-
routing_table_api=Api.shields,
|
|
32
|
-
router_api=Api.safety,
|
|
33
|
-
),
|
|
34
|
-
AutoRoutedApiInfo(
|
|
35
|
-
routing_table_api=Api.memory_banks,
|
|
36
|
-
router_api=Api.memory,
|
|
37
|
-
),
|
|
38
|
-
]
|
|
39
|
-
|
|
40
|
-
|
|
41
|
-
def providable_apis() -> List[Api]:
|
|
42
|
-
routing_table_apis = set(
|
|
43
|
-
x.routing_table_api for x in builtin_automatically_routed_apis()
|
|
44
|
-
)
|
|
45
|
-
return [api for api in Api if api not in routing_table_apis and api != Api.inspect]
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
def get_provider_registry() -> Dict[Api, Dict[str, ProviderSpec]]:
|
|
49
|
-
ret = {}
|
|
50
|
-
for api in providable_apis():
|
|
51
|
-
name = api.name.lower()
|
|
52
|
-
module = importlib.import_module(f"llama_stack.providers.registry.{name}")
|
|
53
|
-
ret[api] = {
|
|
54
|
-
"remote": remote_provider_spec(api),
|
|
55
|
-
**{a.provider_type: a for a in module.available_providers()},
|
|
56
|
-
}
|
|
57
|
-
|
|
58
|
-
return ret
|
|
@@ -1,67 +0,0 @@
|
|
|
1
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
2
|
-
# All rights reserved.
|
|
3
|
-
#
|
|
4
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
5
|
-
# the root directory of this source tree.
|
|
6
|
-
|
|
7
|
-
from typing import Dict, List
|
|
8
|
-
from llama_stack.apis.inspect import * # noqa: F403
|
|
9
|
-
from pydantic import BaseModel
|
|
10
|
-
|
|
11
|
-
from llama_stack.distribution.server.endpoints import get_all_api_endpoints
|
|
12
|
-
from llama_stack.providers.datatypes import * # noqa: F403
|
|
13
|
-
from llama_stack.distribution.datatypes import * # noqa: F403
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
class DistributionInspectConfig(BaseModel):
|
|
17
|
-
run_config: StackRunConfig
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
async def get_provider_impl(config, deps):
|
|
21
|
-
impl = DistributionInspectImpl(config, deps)
|
|
22
|
-
await impl.initialize()
|
|
23
|
-
return impl
|
|
24
|
-
|
|
25
|
-
|
|
26
|
-
class DistributionInspectImpl(Inspect):
|
|
27
|
-
def __init__(self, config, deps):
|
|
28
|
-
self.config = config
|
|
29
|
-
self.deps = deps
|
|
30
|
-
|
|
31
|
-
async def initialize(self) -> None:
|
|
32
|
-
pass
|
|
33
|
-
|
|
34
|
-
async def list_providers(self) -> Dict[str, List[ProviderInfo]]:
|
|
35
|
-
run_config = self.config.run_config
|
|
36
|
-
|
|
37
|
-
ret = {}
|
|
38
|
-
for api, providers in run_config.providers.items():
|
|
39
|
-
ret[api] = [
|
|
40
|
-
ProviderInfo(
|
|
41
|
-
provider_id=p.provider_id,
|
|
42
|
-
provider_type=p.provider_type,
|
|
43
|
-
)
|
|
44
|
-
for p in providers
|
|
45
|
-
]
|
|
46
|
-
|
|
47
|
-
return ret
|
|
48
|
-
|
|
49
|
-
async def list_routes(self) -> Dict[str, List[RouteInfo]]:
|
|
50
|
-
run_config = self.config.run_config
|
|
51
|
-
|
|
52
|
-
ret = {}
|
|
53
|
-
all_endpoints = get_all_api_endpoints()
|
|
54
|
-
for api, endpoints in all_endpoints.items():
|
|
55
|
-
providers = run_config.providers.get(api.value, [])
|
|
56
|
-
ret[api.value] = [
|
|
57
|
-
RouteInfo(
|
|
58
|
-
route=e.route,
|
|
59
|
-
method=e.method,
|
|
60
|
-
provider_types=[p.provider_type for p in providers],
|
|
61
|
-
)
|
|
62
|
-
for e in endpoints
|
|
63
|
-
]
|
|
64
|
-
return ret
|
|
65
|
-
|
|
66
|
-
async def health(self) -> HealthInfo:
|
|
67
|
-
return HealthInfo(status="OK")
|
|
@@ -1,57 +0,0 @@
|
|
|
1
|
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
|
2
|
-
# All rights reserved.
|
|
3
|
-
#
|
|
4
|
-
# This source code is licensed under the terms described in the LICENSE file in
|
|
5
|
-
# the root directory of this source tree.
|
|
6
|
-
|
|
7
|
-
import json
|
|
8
|
-
import threading
|
|
9
|
-
from typing import Any, Dict
|
|
10
|
-
|
|
11
|
-
from .utils.dynamic import instantiate_class_type
|
|
12
|
-
|
|
13
|
-
_THREAD_LOCAL = threading.local()
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
class NeedsRequestProviderData:
|
|
17
|
-
def get_request_provider_data(self) -> Any:
|
|
18
|
-
spec = self.__provider_spec__
|
|
19
|
-
assert spec, f"Provider spec not set on {self.__class__}"
|
|
20
|
-
|
|
21
|
-
provider_type = spec.provider_type
|
|
22
|
-
validator_class = spec.provider_data_validator
|
|
23
|
-
if not validator_class:
|
|
24
|
-
raise ValueError(f"Provider {provider_type} does not have a validator")
|
|
25
|
-
|
|
26
|
-
val = getattr(_THREAD_LOCAL, "provider_data_header_value", None)
|
|
27
|
-
if not val:
|
|
28
|
-
return None
|
|
29
|
-
|
|
30
|
-
validator = instantiate_class_type(validator_class)
|
|
31
|
-
try:
|
|
32
|
-
provider_data = validator(**val)
|
|
33
|
-
return provider_data
|
|
34
|
-
except Exception as e:
|
|
35
|
-
print("Error parsing provider data", e)
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
def set_request_provider_data(headers: Dict[str, str]):
|
|
39
|
-
keys = [
|
|
40
|
-
"X-LlamaStack-ProviderData",
|
|
41
|
-
"x-llamastack-providerdata",
|
|
42
|
-
]
|
|
43
|
-
for key in keys:
|
|
44
|
-
val = headers.get(key, None)
|
|
45
|
-
if val:
|
|
46
|
-
break
|
|
47
|
-
|
|
48
|
-
if not val:
|
|
49
|
-
return
|
|
50
|
-
|
|
51
|
-
try:
|
|
52
|
-
val = json.loads(val)
|
|
53
|
-
except json.JSONDecodeError:
|
|
54
|
-
print("Provider data not encoded as a JSON object!", val)
|
|
55
|
-
return
|
|
56
|
-
|
|
57
|
-
_THREAD_LOCAL.provider_data_header_value = val
|