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
|
@@ -1,52 +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 Any, Dict, Optional
|
|
8
|
-
|
|
9
|
-
from pydantic import BaseModel, ConfigDict, Field
|
|
10
|
-
|
|
11
|
-
from llama_models.datatypes import * # noqa: F403
|
|
12
|
-
from llama_models.sku_list import LlamaDownloadInfo
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
class PromptGuardModel(BaseModel):
|
|
16
|
-
"""Make a 'fake' Model-like object for Prompt Guard. Eventually this will be removed."""
|
|
17
|
-
|
|
18
|
-
model_id: str = "Prompt-Guard-86M"
|
|
19
|
-
description: str = (
|
|
20
|
-
"Prompt Guard. NOTE: this model will not be provided via `llama` CLI soon."
|
|
21
|
-
)
|
|
22
|
-
is_featured: bool = False
|
|
23
|
-
huggingface_repo: str = "meta-llama/Prompt-Guard-86M"
|
|
24
|
-
max_seq_length: int = 2048
|
|
25
|
-
is_instruct_model: bool = False
|
|
26
|
-
quantization_format: CheckpointQuantizationFormat = (
|
|
27
|
-
CheckpointQuantizationFormat.bf16
|
|
28
|
-
)
|
|
29
|
-
arch_args: Dict[str, Any] = Field(default_factory=dict)
|
|
30
|
-
recommended_sampling_params: Optional[SamplingParams] = None
|
|
31
|
-
|
|
32
|
-
def descriptor(self) -> str:
|
|
33
|
-
return self.model_id
|
|
34
|
-
|
|
35
|
-
model_config = ConfigDict(protected_namespaces=())
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
def prompt_guard_model_sku():
|
|
39
|
-
return PromptGuardModel()
|
|
40
|
-
|
|
41
|
-
|
|
42
|
-
def prompt_guard_download_info():
|
|
43
|
-
return LlamaDownloadInfo(
|
|
44
|
-
folder="Prompt-Guard",
|
|
45
|
-
files=[
|
|
46
|
-
"model.safetensors",
|
|
47
|
-
"special_tokens_map.json",
|
|
48
|
-
"tokenizer.json",
|
|
49
|
-
"tokenizer_config.json",
|
|
50
|
-
],
|
|
51
|
-
pth_size=1,
|
|
52
|
-
)
|
llama_stack/cli/stack/build.py
DELETED
|
@@ -1,299 +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 argparse
|
|
8
|
-
|
|
9
|
-
from llama_stack.cli.subcommand import Subcommand
|
|
10
|
-
from llama_stack.distribution.datatypes import * # noqa: F403
|
|
11
|
-
import os
|
|
12
|
-
from functools import lru_cache
|
|
13
|
-
from pathlib import Path
|
|
14
|
-
|
|
15
|
-
TEMPLATES_PATH = (
|
|
16
|
-
Path(os.path.relpath(__file__)).parent.parent.parent / "distribution" / "templates"
|
|
17
|
-
)
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
@lru_cache()
|
|
21
|
-
def available_templates_specs() -> List[BuildConfig]:
|
|
22
|
-
import yaml
|
|
23
|
-
|
|
24
|
-
template_specs = []
|
|
25
|
-
for p in TEMPLATES_PATH.rglob("*build.yaml"):
|
|
26
|
-
with open(p, "r") as f:
|
|
27
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
28
|
-
template_specs.append(build_config)
|
|
29
|
-
|
|
30
|
-
return template_specs
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
class StackBuild(Subcommand):
|
|
34
|
-
def __init__(self, subparsers: argparse._SubParsersAction):
|
|
35
|
-
super().__init__()
|
|
36
|
-
self.parser = subparsers.add_parser(
|
|
37
|
-
"build",
|
|
38
|
-
prog="llama stack build",
|
|
39
|
-
description="Build a Llama stack container",
|
|
40
|
-
formatter_class=argparse.RawTextHelpFormatter,
|
|
41
|
-
)
|
|
42
|
-
self._add_arguments()
|
|
43
|
-
self.parser.set_defaults(func=self._run_stack_build_command)
|
|
44
|
-
|
|
45
|
-
def _add_arguments(self):
|
|
46
|
-
self.parser.add_argument(
|
|
47
|
-
"--config",
|
|
48
|
-
type=str,
|
|
49
|
-
default=None,
|
|
50
|
-
help="Path to a config file to use for the build. You can find example configs in llama_stack/distribution/example_configs. If this argument is not provided, you will be prompted to enter information interactively",
|
|
51
|
-
)
|
|
52
|
-
|
|
53
|
-
self.parser.add_argument(
|
|
54
|
-
"--template",
|
|
55
|
-
type=str,
|
|
56
|
-
default=None,
|
|
57
|
-
help="Name of the example template config to use for build. You may use `llama stack build --list-templates` to check out the available templates",
|
|
58
|
-
)
|
|
59
|
-
|
|
60
|
-
self.parser.add_argument(
|
|
61
|
-
"--list-templates",
|
|
62
|
-
type=bool,
|
|
63
|
-
default=False,
|
|
64
|
-
action=argparse.BooleanOptionalAction,
|
|
65
|
-
help="Show the available templates for building a Llama Stack distribution",
|
|
66
|
-
)
|
|
67
|
-
|
|
68
|
-
self.parser.add_argument(
|
|
69
|
-
"--name",
|
|
70
|
-
type=str,
|
|
71
|
-
help="Name of the Llama Stack build to override from template config. This name will be used as paths to store configuration files, build conda environments/docker images. If not specified, will use the name from the template config. ",
|
|
72
|
-
)
|
|
73
|
-
|
|
74
|
-
self.parser.add_argument(
|
|
75
|
-
"--image-type",
|
|
76
|
-
type=str,
|
|
77
|
-
help="Image Type to use for the build. This can be either conda or docker. If not specified, will use the image type from the template config.",
|
|
78
|
-
choices=["conda", "docker"],
|
|
79
|
-
)
|
|
80
|
-
|
|
81
|
-
def _get_build_config_from_name(self, args: argparse.Namespace) -> Optional[Path]:
|
|
82
|
-
if os.getenv("CONDA_PREFIX", ""):
|
|
83
|
-
conda_dir = (
|
|
84
|
-
Path(os.getenv("CONDA_PREFIX")).parent / f"llamastack-{args.name}"
|
|
85
|
-
)
|
|
86
|
-
else:
|
|
87
|
-
cprint(
|
|
88
|
-
"Cannot find CONDA_PREFIX. Trying default conda path ~/.conda/envs...",
|
|
89
|
-
color="green",
|
|
90
|
-
)
|
|
91
|
-
conda_dir = (
|
|
92
|
-
Path(os.path.expanduser("~/.conda/envs")) / f"llamastack-{args.name}"
|
|
93
|
-
)
|
|
94
|
-
build_config_file = Path(conda_dir) / f"{args.name}-build.yaml"
|
|
95
|
-
if build_config_file.exists():
|
|
96
|
-
return build_config_file
|
|
97
|
-
|
|
98
|
-
return None
|
|
99
|
-
|
|
100
|
-
def _run_stack_build_command_from_build_config(
|
|
101
|
-
self, build_config: BuildConfig
|
|
102
|
-
) -> None:
|
|
103
|
-
import json
|
|
104
|
-
import os
|
|
105
|
-
|
|
106
|
-
import yaml
|
|
107
|
-
|
|
108
|
-
from llama_stack.distribution.build import build_image, ImageType
|
|
109
|
-
from llama_stack.distribution.utils.config_dirs import DISTRIBS_BASE_DIR
|
|
110
|
-
from llama_stack.distribution.utils.serialize import EnumEncoder
|
|
111
|
-
from termcolor import cprint
|
|
112
|
-
|
|
113
|
-
# save build.yaml spec for building same distribution again
|
|
114
|
-
if build_config.image_type == ImageType.docker.value:
|
|
115
|
-
# docker needs build file to be in the llama-stack repo dir to be able to copy over to the image
|
|
116
|
-
llama_stack_path = Path(
|
|
117
|
-
os.path.abspath(__file__)
|
|
118
|
-
).parent.parent.parent.parent
|
|
119
|
-
build_dir = llama_stack_path / "tmp/configs/"
|
|
120
|
-
else:
|
|
121
|
-
build_dir = DISTRIBS_BASE_DIR / f"llamastack-{build_config.name}"
|
|
122
|
-
|
|
123
|
-
os.makedirs(build_dir, exist_ok=True)
|
|
124
|
-
build_file_path = build_dir / f"{build_config.name}-build.yaml"
|
|
125
|
-
|
|
126
|
-
with open(build_file_path, "w") as f:
|
|
127
|
-
to_write = json.loads(json.dumps(build_config.dict(), cls=EnumEncoder))
|
|
128
|
-
f.write(yaml.dump(to_write, sort_keys=False))
|
|
129
|
-
|
|
130
|
-
return_code = build_image(build_config, build_file_path)
|
|
131
|
-
if return_code != 0:
|
|
132
|
-
return
|
|
133
|
-
|
|
134
|
-
configure_name = (
|
|
135
|
-
build_config.name
|
|
136
|
-
if build_config.image_type == "conda"
|
|
137
|
-
else (f"llamastack-{build_config.name}")
|
|
138
|
-
)
|
|
139
|
-
if build_config.image_type == "conda":
|
|
140
|
-
cprint(
|
|
141
|
-
f"You can now run `llama stack configure {configure_name}`",
|
|
142
|
-
color="green",
|
|
143
|
-
)
|
|
144
|
-
else:
|
|
145
|
-
cprint(
|
|
146
|
-
f"You can now run `llama stack run {build_config.name}`",
|
|
147
|
-
color="green",
|
|
148
|
-
)
|
|
149
|
-
|
|
150
|
-
def _run_template_list_cmd(self, args: argparse.Namespace) -> None:
|
|
151
|
-
import json
|
|
152
|
-
|
|
153
|
-
from llama_stack.cli.table import print_table
|
|
154
|
-
|
|
155
|
-
# eventually, this should query a registry at llama.meta.com/llamastack/distributions
|
|
156
|
-
headers = [
|
|
157
|
-
"Template Name",
|
|
158
|
-
"Providers",
|
|
159
|
-
"Description",
|
|
160
|
-
]
|
|
161
|
-
|
|
162
|
-
rows = []
|
|
163
|
-
for spec in available_templates_specs():
|
|
164
|
-
rows.append(
|
|
165
|
-
[
|
|
166
|
-
spec.name,
|
|
167
|
-
json.dumps(spec.distribution_spec.providers, indent=2),
|
|
168
|
-
spec.distribution_spec.description,
|
|
169
|
-
]
|
|
170
|
-
)
|
|
171
|
-
print_table(
|
|
172
|
-
rows,
|
|
173
|
-
headers,
|
|
174
|
-
separate_rows=True,
|
|
175
|
-
)
|
|
176
|
-
|
|
177
|
-
def _run_stack_build_command(self, args: argparse.Namespace) -> None:
|
|
178
|
-
import textwrap
|
|
179
|
-
|
|
180
|
-
import yaml
|
|
181
|
-
from llama_stack.distribution.distribution import get_provider_registry
|
|
182
|
-
from prompt_toolkit import prompt
|
|
183
|
-
from prompt_toolkit.completion import WordCompleter
|
|
184
|
-
from prompt_toolkit.validation import Validator
|
|
185
|
-
from termcolor import cprint
|
|
186
|
-
|
|
187
|
-
if args.list_templates:
|
|
188
|
-
self._run_template_list_cmd(args)
|
|
189
|
-
return
|
|
190
|
-
|
|
191
|
-
if args.template:
|
|
192
|
-
if not args.name:
|
|
193
|
-
self.parser.error(
|
|
194
|
-
"You must specify a name for the build using --name when using a template"
|
|
195
|
-
)
|
|
196
|
-
return
|
|
197
|
-
build_path = TEMPLATES_PATH / f"{args.template}-build.yaml"
|
|
198
|
-
if not build_path.exists():
|
|
199
|
-
self.parser.error(
|
|
200
|
-
f"Could not find template {args.template}. Please run `llama stack build --list-templates` to check out the available templates"
|
|
201
|
-
)
|
|
202
|
-
return
|
|
203
|
-
with open(build_path, "r") as f:
|
|
204
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
205
|
-
build_config.name = args.name
|
|
206
|
-
if args.image_type:
|
|
207
|
-
build_config.image_type = args.image_type
|
|
208
|
-
self._run_stack_build_command_from_build_config(build_config)
|
|
209
|
-
|
|
210
|
-
return
|
|
211
|
-
|
|
212
|
-
# try to see if we can find a pre-existing build config file through name
|
|
213
|
-
if args.name:
|
|
214
|
-
maybe_build_config = self._get_build_config_from_name(args)
|
|
215
|
-
if maybe_build_config:
|
|
216
|
-
cprint(
|
|
217
|
-
f"Building from existing build config for {args.name} in {str(maybe_build_config)}...",
|
|
218
|
-
"green",
|
|
219
|
-
)
|
|
220
|
-
with open(maybe_build_config, "r") as f:
|
|
221
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
222
|
-
self._run_stack_build_command_from_build_config(build_config)
|
|
223
|
-
return
|
|
224
|
-
|
|
225
|
-
if not args.config and not args.template:
|
|
226
|
-
if not args.name:
|
|
227
|
-
name = prompt(
|
|
228
|
-
"> Enter a name for your Llama Stack (e.g. my-local-stack): ",
|
|
229
|
-
validator=Validator.from_callable(
|
|
230
|
-
lambda x: len(x) > 0,
|
|
231
|
-
error_message="Name cannot be empty, please enter a name",
|
|
232
|
-
),
|
|
233
|
-
)
|
|
234
|
-
else:
|
|
235
|
-
name = args.name
|
|
236
|
-
|
|
237
|
-
image_type = prompt(
|
|
238
|
-
"> Enter the image type you want your Llama Stack to be built as (docker or conda): ",
|
|
239
|
-
validator=Validator.from_callable(
|
|
240
|
-
lambda x: x in ["docker", "conda"],
|
|
241
|
-
error_message="Invalid image type, please enter conda or docker",
|
|
242
|
-
),
|
|
243
|
-
default="conda",
|
|
244
|
-
)
|
|
245
|
-
|
|
246
|
-
cprint(
|
|
247
|
-
textwrap.dedent(
|
|
248
|
-
"""
|
|
249
|
-
Llama Stack is composed of several APIs working together. Let's select
|
|
250
|
-
the provider types (implementations) you want to use for these APIs.
|
|
251
|
-
""",
|
|
252
|
-
),
|
|
253
|
-
color="green",
|
|
254
|
-
)
|
|
255
|
-
|
|
256
|
-
print("Tip: use <TAB> to see options for the providers.\n")
|
|
257
|
-
|
|
258
|
-
providers = dict()
|
|
259
|
-
for api, providers_for_api in get_provider_registry().items():
|
|
260
|
-
available_providers = [
|
|
261
|
-
x
|
|
262
|
-
for x in providers_for_api.keys()
|
|
263
|
-
if x not in ("remote", "remote::sample")
|
|
264
|
-
]
|
|
265
|
-
api_provider = prompt(
|
|
266
|
-
"> Enter provider for API {}: ".format(api.value),
|
|
267
|
-
completer=WordCompleter(available_providers),
|
|
268
|
-
complete_while_typing=True,
|
|
269
|
-
validator=Validator.from_callable(
|
|
270
|
-
lambda x: x in available_providers,
|
|
271
|
-
error_message="Invalid provider, use <TAB> to see options",
|
|
272
|
-
),
|
|
273
|
-
)
|
|
274
|
-
|
|
275
|
-
providers[api.value] = api_provider
|
|
276
|
-
|
|
277
|
-
description = prompt(
|
|
278
|
-
"\n > (Optional) Enter a short description for your Llama Stack: ",
|
|
279
|
-
default="",
|
|
280
|
-
)
|
|
281
|
-
|
|
282
|
-
distribution_spec = DistributionSpec(
|
|
283
|
-
providers=providers,
|
|
284
|
-
description=description,
|
|
285
|
-
)
|
|
286
|
-
|
|
287
|
-
build_config = BuildConfig(
|
|
288
|
-
name=name, image_type=image_type, distribution_spec=distribution_spec
|
|
289
|
-
)
|
|
290
|
-
self._run_stack_build_command_from_build_config(build_config)
|
|
291
|
-
return
|
|
292
|
-
|
|
293
|
-
with open(args.config, "r") as f:
|
|
294
|
-
try:
|
|
295
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
296
|
-
except Exception as e:
|
|
297
|
-
self.parser.error(f"Could not parse config file {args.config}: {e}")
|
|
298
|
-
return
|
|
299
|
-
self._run_stack_build_command_from_build_config(build_config)
|
|
@@ -1,178 +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 argparse
|
|
8
|
-
|
|
9
|
-
from llama_stack.cli.subcommand import Subcommand
|
|
10
|
-
from llama_stack.distribution.utils.config_dirs import BUILDS_BASE_DIR
|
|
11
|
-
from llama_stack.distribution.datatypes import * # noqa: F403
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
class StackConfigure(Subcommand):
|
|
15
|
-
"""Llama cli for configuring llama toolchain configs"""
|
|
16
|
-
|
|
17
|
-
def __init__(self, subparsers: argparse._SubParsersAction):
|
|
18
|
-
super().__init__()
|
|
19
|
-
self.parser = subparsers.add_parser(
|
|
20
|
-
"configure",
|
|
21
|
-
prog="llama stack configure",
|
|
22
|
-
description="configure a llama stack distribution",
|
|
23
|
-
formatter_class=argparse.RawTextHelpFormatter,
|
|
24
|
-
)
|
|
25
|
-
self._add_arguments()
|
|
26
|
-
self.parser.set_defaults(func=self._run_stack_configure_cmd)
|
|
27
|
-
|
|
28
|
-
def _add_arguments(self):
|
|
29
|
-
self.parser.add_argument(
|
|
30
|
-
"config",
|
|
31
|
-
type=str,
|
|
32
|
-
help="Path to the build config file (e.g. ~/.llama/builds/<image_type>/<name>-build.yaml). For docker, this could also be the name of the docker image. ",
|
|
33
|
-
)
|
|
34
|
-
|
|
35
|
-
self.parser.add_argument(
|
|
36
|
-
"--output-dir",
|
|
37
|
-
type=str,
|
|
38
|
-
help="Path to the output directory to store generated run.yaml config file. If not specified, will use ~/.llama/build/<image_type>/<name>-run.yaml",
|
|
39
|
-
)
|
|
40
|
-
|
|
41
|
-
def _run_stack_configure_cmd(self, args: argparse.Namespace) -> None:
|
|
42
|
-
import json
|
|
43
|
-
import os
|
|
44
|
-
import subprocess
|
|
45
|
-
from pathlib import Path
|
|
46
|
-
|
|
47
|
-
import pkg_resources
|
|
48
|
-
|
|
49
|
-
import yaml
|
|
50
|
-
from termcolor import cprint
|
|
51
|
-
|
|
52
|
-
from llama_stack.distribution.build import ImageType
|
|
53
|
-
from llama_stack.distribution.utils.exec import run_with_pty
|
|
54
|
-
|
|
55
|
-
docker_image = None
|
|
56
|
-
|
|
57
|
-
build_config_file = Path(args.config)
|
|
58
|
-
|
|
59
|
-
if build_config_file.exists():
|
|
60
|
-
with open(build_config_file, "r") as f:
|
|
61
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
62
|
-
self._configure_llama_distribution(build_config, args.output_dir)
|
|
63
|
-
return
|
|
64
|
-
|
|
65
|
-
# if we get here, we need to try to find the conda build config file
|
|
66
|
-
cprint(
|
|
67
|
-
f"Could not find {build_config_file}. Trying conda build name instead...",
|
|
68
|
-
color="green",
|
|
69
|
-
)
|
|
70
|
-
|
|
71
|
-
conda_dir = (
|
|
72
|
-
Path(os.path.expanduser("~/.conda/envs")) / f"llamastack-{args.config}"
|
|
73
|
-
)
|
|
74
|
-
output = subprocess.check_output(["bash", "-c", "conda info --json"])
|
|
75
|
-
conda_envs = json.loads(output.decode("utf-8"))["envs"]
|
|
76
|
-
|
|
77
|
-
for x in conda_envs:
|
|
78
|
-
if x.endswith(f"/llamastack-{args.config}"):
|
|
79
|
-
conda_dir = Path(x)
|
|
80
|
-
break
|
|
81
|
-
|
|
82
|
-
build_config_file = Path(conda_dir) / f"{args.config}-build.yaml"
|
|
83
|
-
|
|
84
|
-
if build_config_file.exists():
|
|
85
|
-
with open(build_config_file, "r") as f:
|
|
86
|
-
build_config = BuildConfig(**yaml.safe_load(f))
|
|
87
|
-
|
|
88
|
-
self._configure_llama_distribution(build_config, args.output_dir)
|
|
89
|
-
return
|
|
90
|
-
|
|
91
|
-
# if we get here, we need to try to find the docker image
|
|
92
|
-
cprint(
|
|
93
|
-
f"Could not find {build_config_file}. Trying docker image name instead...",
|
|
94
|
-
color="green",
|
|
95
|
-
)
|
|
96
|
-
docker_image = args.config
|
|
97
|
-
builds_dir = BUILDS_BASE_DIR / ImageType.docker.value
|
|
98
|
-
if args.output_dir:
|
|
99
|
-
builds_dir = Path(output_dir)
|
|
100
|
-
os.makedirs(builds_dir, exist_ok=True)
|
|
101
|
-
|
|
102
|
-
script = pkg_resources.resource_filename(
|
|
103
|
-
"llama_stack", "distribution/configure_container.sh"
|
|
104
|
-
)
|
|
105
|
-
script_args = [script, docker_image, str(builds_dir)]
|
|
106
|
-
|
|
107
|
-
return_code = run_with_pty(script_args)
|
|
108
|
-
|
|
109
|
-
# we have regenerated the build config file with script, now check if it exists
|
|
110
|
-
if return_code != 0:
|
|
111
|
-
self.parser.error(
|
|
112
|
-
f"Failed to configure container {docker_image} with return code {return_code}. Please run `llama stack build` first. "
|
|
113
|
-
)
|
|
114
|
-
return
|
|
115
|
-
|
|
116
|
-
return
|
|
117
|
-
|
|
118
|
-
def _configure_llama_distribution(
|
|
119
|
-
self,
|
|
120
|
-
build_config: BuildConfig,
|
|
121
|
-
output_dir: Optional[str] = None,
|
|
122
|
-
):
|
|
123
|
-
import json
|
|
124
|
-
import os
|
|
125
|
-
from pathlib import Path
|
|
126
|
-
|
|
127
|
-
import yaml
|
|
128
|
-
from termcolor import cprint
|
|
129
|
-
|
|
130
|
-
from llama_stack.distribution.configure import (
|
|
131
|
-
configure_api_providers,
|
|
132
|
-
parse_and_maybe_upgrade_config,
|
|
133
|
-
)
|
|
134
|
-
from llama_stack.distribution.utils.serialize import EnumEncoder
|
|
135
|
-
|
|
136
|
-
builds_dir = BUILDS_BASE_DIR / build_config.image_type
|
|
137
|
-
if output_dir:
|
|
138
|
-
builds_dir = Path(output_dir)
|
|
139
|
-
os.makedirs(builds_dir, exist_ok=True)
|
|
140
|
-
image_name = build_config.name.replace("::", "-")
|
|
141
|
-
run_config_file = builds_dir / f"{image_name}-run.yaml"
|
|
142
|
-
|
|
143
|
-
if run_config_file.exists():
|
|
144
|
-
cprint(
|
|
145
|
-
f"Configuration already exists at `{str(run_config_file)}`. Will overwrite...",
|
|
146
|
-
"yellow",
|
|
147
|
-
attrs=["bold"],
|
|
148
|
-
)
|
|
149
|
-
config_dict = yaml.safe_load(run_config_file.read_text())
|
|
150
|
-
config = parse_and_maybe_upgrade_config(config_dict)
|
|
151
|
-
else:
|
|
152
|
-
config = StackRunConfig(
|
|
153
|
-
built_at=datetime.now(),
|
|
154
|
-
image_name=image_name,
|
|
155
|
-
apis=list(build_config.distribution_spec.providers.keys()),
|
|
156
|
-
providers={},
|
|
157
|
-
)
|
|
158
|
-
|
|
159
|
-
config = configure_api_providers(config, build_config.distribution_spec)
|
|
160
|
-
|
|
161
|
-
config.docker_image = (
|
|
162
|
-
image_name if build_config.image_type == "docker" else None
|
|
163
|
-
)
|
|
164
|
-
config.conda_env = image_name if build_config.image_type == "conda" else None
|
|
165
|
-
|
|
166
|
-
with open(run_config_file, "w") as f:
|
|
167
|
-
to_write = json.loads(json.dumps(config.dict(), cls=EnumEncoder))
|
|
168
|
-
f.write(yaml.dump(to_write, sort_keys=False))
|
|
169
|
-
|
|
170
|
-
cprint(
|
|
171
|
-
f"> YAML configuration has been written to `{run_config_file}`.",
|
|
172
|
-
color="blue",
|
|
173
|
-
)
|
|
174
|
-
|
|
175
|
-
cprint(
|
|
176
|
-
f"You can now run `llama stack run {image_name} --port PORT`",
|
|
177
|
-
color="green",
|
|
178
|
-
)
|
|
@@ -1,123 +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 enum import Enum
|
|
8
|
-
from typing import List, Optional
|
|
9
|
-
|
|
10
|
-
import pkg_resources
|
|
11
|
-
|
|
12
|
-
from llama_stack.distribution.utils.exec import run_with_pty
|
|
13
|
-
from pydantic import BaseModel
|
|
14
|
-
|
|
15
|
-
from termcolor import cprint
|
|
16
|
-
|
|
17
|
-
from llama_stack.distribution.datatypes import * # noqa: F403
|
|
18
|
-
from pathlib import Path
|
|
19
|
-
|
|
20
|
-
from llama_stack.distribution.utils.config_dirs import BUILDS_BASE_DIR
|
|
21
|
-
from llama_stack.distribution.distribution import get_provider_registry
|
|
22
|
-
|
|
23
|
-
|
|
24
|
-
# These are the dependencies needed by the distribution server.
|
|
25
|
-
# `llama-stack` is automatically installed by the installation script.
|
|
26
|
-
SERVER_DEPENDENCIES = [
|
|
27
|
-
"fastapi",
|
|
28
|
-
"fire",
|
|
29
|
-
"httpx",
|
|
30
|
-
"uvicorn",
|
|
31
|
-
]
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
class ImageType(Enum):
|
|
35
|
-
docker = "docker"
|
|
36
|
-
conda = "conda"
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
class Dependencies(BaseModel):
|
|
40
|
-
pip_packages: List[str]
|
|
41
|
-
docker_image: Optional[str] = None
|
|
42
|
-
|
|
43
|
-
|
|
44
|
-
class ApiInput(BaseModel):
|
|
45
|
-
api: Api
|
|
46
|
-
provider: str
|
|
47
|
-
|
|
48
|
-
|
|
49
|
-
def build_image(build_config: BuildConfig, build_file_path: Path):
|
|
50
|
-
package_deps = Dependencies(
|
|
51
|
-
docker_image=build_config.distribution_spec.docker_image or "python:3.10-slim",
|
|
52
|
-
pip_packages=SERVER_DEPENDENCIES,
|
|
53
|
-
)
|
|
54
|
-
|
|
55
|
-
# extend package dependencies based on providers spec
|
|
56
|
-
all_providers = get_provider_registry()
|
|
57
|
-
for (
|
|
58
|
-
api_str,
|
|
59
|
-
provider_or_providers,
|
|
60
|
-
) in build_config.distribution_spec.providers.items():
|
|
61
|
-
providers_for_api = all_providers[Api(api_str)]
|
|
62
|
-
|
|
63
|
-
providers = (
|
|
64
|
-
provider_or_providers
|
|
65
|
-
if isinstance(provider_or_providers, list)
|
|
66
|
-
else [provider_or_providers]
|
|
67
|
-
)
|
|
68
|
-
|
|
69
|
-
for provider in providers:
|
|
70
|
-
if provider not in providers_for_api:
|
|
71
|
-
raise ValueError(
|
|
72
|
-
f"Provider `{provider}` is not available for API `{api_str}`"
|
|
73
|
-
)
|
|
74
|
-
|
|
75
|
-
provider_spec = providers_for_api[provider]
|
|
76
|
-
package_deps.pip_packages.extend(provider_spec.pip_packages)
|
|
77
|
-
if provider_spec.docker_image:
|
|
78
|
-
raise ValueError("A stack's dependencies cannot have a docker image")
|
|
79
|
-
|
|
80
|
-
special_deps = []
|
|
81
|
-
deps = []
|
|
82
|
-
for package in package_deps.pip_packages:
|
|
83
|
-
if "--no-deps" in package or "--index-url" in package:
|
|
84
|
-
special_deps.append(package)
|
|
85
|
-
else:
|
|
86
|
-
deps.append(package)
|
|
87
|
-
deps = list(set(deps))
|
|
88
|
-
special_deps = list(set(special_deps))
|
|
89
|
-
|
|
90
|
-
if build_config.image_type == ImageType.docker.value:
|
|
91
|
-
script = pkg_resources.resource_filename(
|
|
92
|
-
"llama_stack", "distribution/build_container.sh"
|
|
93
|
-
)
|
|
94
|
-
args = [
|
|
95
|
-
script,
|
|
96
|
-
build_config.name,
|
|
97
|
-
package_deps.docker_image,
|
|
98
|
-
str(build_file_path),
|
|
99
|
-
str(BUILDS_BASE_DIR / ImageType.docker.value),
|
|
100
|
-
" ".join(deps),
|
|
101
|
-
]
|
|
102
|
-
else:
|
|
103
|
-
script = pkg_resources.resource_filename(
|
|
104
|
-
"llama_stack", "distribution/build_conda_env.sh"
|
|
105
|
-
)
|
|
106
|
-
args = [
|
|
107
|
-
script,
|
|
108
|
-
build_config.name,
|
|
109
|
-
str(build_file_path),
|
|
110
|
-
" ".join(deps),
|
|
111
|
-
]
|
|
112
|
-
|
|
113
|
-
if special_deps:
|
|
114
|
-
args.append("#".join(special_deps))
|
|
115
|
-
|
|
116
|
-
return_code = run_with_pty(args)
|
|
117
|
-
if return_code != 0:
|
|
118
|
-
cprint(
|
|
119
|
-
f"Failed to build target {build_config.name} with return code {return_code}",
|
|
120
|
-
color="red",
|
|
121
|
-
)
|
|
122
|
-
|
|
123
|
-
return return_code
|