llama-stack 0.0.42__py3-none-any.whl → 0.3.4__py3-none-any.whl

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_entity_recall.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+context_entity_recall_fn_def = ScoringFn(
+    identifier="braintrust::context-entity-recall",
+    description=(
+        "Evaluates how well the context captures the named entities present in the "
+        "reference answer. See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="context-entity-recall",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_precision.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+context_precision_fn_def = ScoringFn(
+    identifier="braintrust::context-precision",
+    description=(
+        "Measures how much of the provided context is actually relevant to answering the "
+        "question. See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="context-precision",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_recall.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+context_recall_fn_def = ScoringFn(
+    identifier="braintrust::context-recall",
+    description=(
+        "Evaluates how well the context covers the information needed to answer the "
+        "question. See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="context-recall",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/context_relevancy.py

@@ -0,0 +1,23 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+context_relevancy_fn_def = ScoringFn(
+    identifier="braintrust::context-relevancy",
+    description=(
+        "Assesses how relevant the provided context is to the given question. See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="context-relevancy",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/factuality.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+factuality_fn_def = ScoringFn(
+    identifier="braintrust::factuality",
+    description=(
+        "Test output factuality against expected value using Braintrust LLM scorer. "
+        "See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="factuality",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/braintrust/scoring_fn/fn_defs/faithfulness.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    BasicScoringFnParams,
+    ScoringFn,
+)
+
+faithfulness_fn_def = ScoringFn(
+    identifier="braintrust::faithfulness",
+    description=(
+        "Test output faithfulness to the input query using Braintrust LLM scorer. "
+        "See: github.com/braintrustdata/autoevals"
+    ),
+    provider_id="braintrust",
+    provider_resource_id="faithfulness",
+    return_type=NumberType(),
+    params=BasicScoringFnParams(aggregation_functions=[AggregationFunctionType.average]),
+)

llama_stack/providers/inline/scoring/llm_as_judge/__init__.py

@@ -0,0 +1,21 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+from typing import Any
+
+from llama_stack.core.datatypes import Api
+
+from .config import LlmAsJudgeScoringConfig
+
+
+async def get_provider_impl(
+    config: LlmAsJudgeScoringConfig,
+    deps: dict[Api, Any],
+):
+    from .scoring import LlmAsJudgeScoringImpl
+
+    impl = LlmAsJudgeScoringImpl(config, deps[Api.datasetio], deps[Api.datasets], deps[Api.inference])
+    await impl.initialize()
+    return impl

llama_stack/providers/inline/scoring/llm_as_judge/config.py

@@ -0,0 +1,14 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+from typing import Any
+
+from pydantic import BaseModel
+
+
+class LlmAsJudgeScoringConfig(BaseModel):
+    @classmethod
+    def sample_run_config(cls, __distro_dir__: str, **kwargs: Any) -> dict[str, Any]:
+        return {}

llama_stack/providers/inline/scoring/llm_as_judge/scoring.py

@@ -0,0 +1,113 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+from typing import Any
+
+from llama_stack.apis.datasetio import DatasetIO
+from llama_stack.apis.datasets import Datasets
+from llama_stack.apis.inference import Inference
+from llama_stack.apis.scoring import (
+    ScoreBatchResponse,
+    ScoreResponse,
+    Scoring,
+    ScoringResult,
+)
+from llama_stack.apis.scoring_functions import ScoringFn, ScoringFnParams
+from llama_stack.core.datatypes import Api
+from llama_stack.providers.datatypes import ScoringFunctionsProtocolPrivate
+from llama_stack.providers.utils.common.data_schema_validator import (
+    get_valid_schemas,
+    validate_dataset_schema,
+)
+
+from .config import LlmAsJudgeScoringConfig
+from .scoring_fn.llm_as_judge_scoring_fn import LlmAsJudgeScoringFn
+
+LLM_JUDGE_FN = LlmAsJudgeScoringFn
+
+
+class LlmAsJudgeScoringImpl(
+    Scoring,
+    ScoringFunctionsProtocolPrivate,
+):
+    def __init__(
+        self,
+        config: LlmAsJudgeScoringConfig,
+        datasetio_api: DatasetIO,
+        datasets_api: Datasets,
+        inference_api: Inference,
+    ) -> None:
+        self.config = config
+        self.datasetio_api = datasetio_api
+        self.datasets_api = datasets_api
+        self.inference_api = inference_api
+
+    async def initialize(self) -> None:
+        impl = LLM_JUDGE_FN(inference_api=self.inference_api)
+        self.llm_as_judge_fn = impl
+
+    async def shutdown(self) -> None: ...
+
+    async def list_scoring_functions(self) -> list[ScoringFn]:
+        scoring_fn_defs_list = self.llm_as_judge_fn.get_supported_scoring_fn_defs()
+
+        for f in self.llm_as_judge_fn.get_supported_scoring_fn_defs():
+            assert f.identifier.startswith("llm-as-judge"), (
+                "All llm-as-judge scoring fn must have identifier prefixed with 'llm-as-judge'! "
+            )
+
+        return scoring_fn_defs_list
+
+    async def register_scoring_function(self, function_def: ScoringFn) -> None:
+        self.llm_as_judge_fn.register_scoring_fn_def(function_def)
+
+    async def unregister_scoring_function(self, scoring_fn_id: str) -> None:
+        self.llm_as_judge_fn.unregister_scoring_fn_def(scoring_fn_id)
+
+    async def score_batch(
+        self,
+        dataset_id: str,
+        scoring_functions: dict[str, ScoringFnParams | None] = None,
+        save_results_dataset: bool = False,
+    ) -> ScoreBatchResponse:
+        dataset_def = await self.datasets_api.get_dataset(dataset_id=dataset_id)
+        validate_dataset_schema(dataset_def.dataset_schema, get_valid_schemas(Api.scoring.value))
+
+        all_rows = await self.datasetio_api.iterrows(
+            dataset_id=dataset_id,
+            limit=-1,
+        )
+        res = await self.score(
+            input_rows=all_rows.data,
+            scoring_functions=scoring_functions,
+        )
+        if save_results_dataset:
+            # TODO: persist and register dataset on to server for reading
+            # self.datasets_api.register_dataset()
+            raise NotImplementedError("Save results dataset not implemented yet")
+
+        return ScoreBatchResponse(
+            results=res.results,
+        )
+
+    async def score(
+        self,
+        input_rows: list[dict[str, Any]],
+        scoring_functions: dict[str, ScoringFnParams | None] = None,
+    ) -> ScoreResponse:
+        res = {}
+        for scoring_fn_id in scoring_functions.keys():
+            scoring_fn = self.llm_as_judge_fn
+            scoring_fn_params = scoring_functions.get(scoring_fn_id, None)
+            score_results = await scoring_fn.score(input_rows, scoring_fn_id, scoring_fn_params)
+            agg_results = await scoring_fn.aggregate(score_results, scoring_fn_id, scoring_fn_params)
+            res[scoring_fn_id] = ScoringResult(
+                score_rows=score_results,
+                aggregated_results=agg_results,
+            )
+
+        return ScoreResponse(
+            results=res,
+        )

llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/__init__.py

@@ -0,0 +1,5 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.

llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/__init__.py

@@ -0,0 +1,5 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.

llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/llm_as_judge_405b_simpleqa.py

@@ -0,0 +1,96 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import (
+    AggregationFunctionType,
+    LLMAsJudgeScoringFnParams,
+    ScoringFn,
+)
+
+GRADER_TEMPLATE = """
+Your job is to look at a question, a gold target, and a predicted answer, and then assign a grade of either ["CORRECT", "INCORRECT", "NOT_ATTEMPTED"].
+First, I will give examples of each grade, and then you will grade a new example.
+The following are examples of CORRECT predicted answers.
+```
+Question: What are the names of Barack Obama's children?
+Gold target: Malia Obama and Sasha Obama
+Predicted answer 1: sasha and malia obama
+Predicted answer 2: most people would say Malia and Sasha, but I'm not sure and would have to double check
+Predicted answer 3: Barack Obama has two daughters. Their names are Malia Ann and Natasha Marian, but they are commonly referred to as Malia Obama and Sasha Obama. Malia was born on July 4, 1998, and Sasha was born on June 10, 2001.
+```
+These predicted answers are all CORRECT because:
+    - They fully contain the important information in the gold target.
+    - They do not contain any information that contradicts the gold target.
+    - Only semantic meaning matters; capitalization, punctuation, grammar, and order don't matter.
+    - Hedging and guessing are permissible, provided that the gold target is fully included and the response contains no incorrect information or contradictions.
+The following are examples of INCORRECT predicted answers.
+```
+Question: What are the names of Barack Obama's children?
+Gold target: Malia and Sasha
+Predicted answer 1: Malia.
+Predicted answer 2: Malia, Sasha, and Susan.
+Predicted answer 3: Barack Obama does not have any children.
+Predicted answer 4: I think it's either Malia and Sasha. Or it could be Malia and Jackie. Or it could be Joey and Malia.
+Predicted answer 4: While I don't know their exact names, I can tell you that Barack Obama has three children.
+Predicted answer 5: It's possible you may mean Betsy and Olivia. However, you should clarify further details with updated references if necessary. Is that the correct answer?
+Predicted answer 6: It may be the case that Obama's child is named James. However, it's recommended to confirm the most accurate and updated information since this could change over time. This model may not always reflect the most current information.
+```
+These predicted answers are all INCORRECT because:
+    - A factual statement in the answer contradicts the gold target. Incorrect statements that have some hedging (e.g., "it is possible that", "although i'm not sure, i think") are also considered incorrect.
+The following are examples of NOT_ATTEMPTED predicted answers.
+```
+Question: What are the names of Barack Obama's children?
+Gold target: Malia and Sasha
+Predicted answer 1: I don't know.
+Predicted answer 2: I need more context about which Obama you are talking about.
+Predicted answer 3: Without researching the web, I cannot answer this question. However, I can tell you that Barack Obama has two children.
+Predicted answer 4: Barack Obama has two children. I know that one of them is Malia, but I'm not sure about the other one.
+```
+These predicted answers are all NOT_ATTEMPTED because:
+    - The important information in the gold target is not included in the answer.
+    - No statements in the answer contradict the gold target.
+Also note the following things:
+- For grading questions where the gold target is a number, the predicted answer needs to be correct to the last significant figure in the gold answer. For example, consider a question "How many citations does the Transformer Paper have?" with gold target "120k".
+    - Predicted answers "120k", "124k", and 115k" are all CORRECT.
+    - Predicted answers "100k" and "113k" are INCORRECT.
+    - Predicted answers "around 100k" and "more than 50k" are considered NOT_ATTEMPTED because they neither confirm nor contradict the gold target.
+- The gold target may contain more information than the question. In such cases, the predicted answer only needs to contain the information that is in the question.
+    - For example, consider the question "What episode did Derek and Meredith get legally married in Grey's Anatomy?" with gold target "Season 7, Episode 20: White Wedding". Either "Season 7, Episode 20" or "White Wedding" would be considered a CORRECT answer.
+- Do not punish predicted answers if they omit information that would be clearly inferred from the question.
+    - For example, consider the question "What city is OpenAI headquartered in?" and the gold target "San Francisco, California". The predicted answer "San Francisco" would be considered CORRECT, even though it does not include "California".
+    - Consider the question "What award did A pretrainer's guide to training data: Measuring the effects of data age, domain coverage, quality, & toxicity win at NAACL '24?", the gold target is "Outstanding Paper Award". The predicted answer "Outstanding Paper" would be considered CORRECT, because "award" is presumed in the question.
+    - For the question "What is the height of Jason Wei in meters?", the gold target is "1.73 m". The predicted answer "1.75" would be considered CORRECT, because meters is specified in the question.
+    - For the question "What is the name of Barack Obama's wife?", the gold target is "Michelle Obama". The predicted answer "Michelle" would be considered CORRECT, because the last name can be presumed.
+- Do not punish for typos in people's name if it's clearly the same name.
+    - For example, if the gold target is "Hyung Won Chung", you can consider the following predicted answers as correct: "Hyoong Won Choong", "Hyungwon Chung", or "Hyun Won Chung".
+Here is a new example. Simply reply with either CORRECT, INCORRECT, NOT ATTEMPTED. Don't apologize or correct yourself if there was a mistake; we are just trying to grade the answer.
+```
+Question: {input_query}
+Gold target: {expected_answer}
+Predicted answer: {generated_answer}
+```
+Grade the predicted answer of this new question as one of:
+A: CORRECT
+B: INCORRECT
+C: NOT_ATTEMPTED
+Just return the letters "A", "B", or "C", with no text around it.
+""".strip()
+
+
+llm_as_judge_405b_simpleqa = ScoringFn(
+    identifier="llm-as-judge::405b-simpleqa",
+    description="Llm As Judge Scoring Function for SimpleQA Benchmark (https://github.com/openai/simple-evals/blob/main/simpleqa_eval.py)",
+    return_type=NumberType(),
+    provider_id="llm-as-judge",
+    provider_resource_id="llm-as-judge-405b-simpleqa",
+    params=LLMAsJudgeScoringFnParams(
+        judge_model="meta-llama/Llama-3.1-405B-Instruct",
+        prompt_template=GRADER_TEMPLATE,
+        judge_score_regexes=[r"(A|B|C)"],
+        aggregation_functions=[AggregationFunctionType.categorical_count.value],
+    ),
+)

llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/fn_defs/llm_as_judge_base.py

@@ -0,0 +1,20 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from llama_stack.apis.common.type_system import NumberType
+from llama_stack.apis.scoring_functions import LLMAsJudgeScoringFnParams, ScoringFn
+
+llm_as_judge_base = ScoringFn(
+    identifier="llm-as-judge::base",
+    description="Llm As Judge Scoring Function",
+    return_type=NumberType(),
+    provider_id="llm-as-judge",
+    provider_resource_id="llm-as-judge-base",
+    params=LLMAsJudgeScoringFnParams(
+        judge_model="meta-llama/Llama-3.1-405B-Instruct",
+        prompt_template="Enter custom LLM as Judge Prompt Template",
+    ),
+)

llama_stack/providers/inline/scoring/llm_as_judge/scoring_fn/llm_as_judge_scoring_fn.py

@@ -0,0 +1,81 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+import re
+from typing import Any
+
+from llama_stack.apis.inference import Inference, OpenAIChatCompletionRequestWithExtraBody
+from llama_stack.apis.scoring import ScoringResultRow
+from llama_stack.apis.scoring_functions import ScoringFnParams
+from llama_stack.providers.utils.scoring.base_scoring_fn import RegisteredBaseScoringFn
+
+from .fn_defs.llm_as_judge_405b_simpleqa import llm_as_judge_405b_simpleqa
+from .fn_defs.llm_as_judge_base import llm_as_judge_base
+
+
+class LlmAsJudgeScoringFn(RegisteredBaseScoringFn):
+    """
+    A scoring_fn that assigns
+    """
+
+    def __init__(self, inference_api: Inference, *arg, **kwargs) -> None:
+        super().__init__(*arg, **kwargs)
+        self.inference_api = inference_api
+        self.supported_fn_defs_registry = {
+            llm_as_judge_base.identifier: llm_as_judge_base,
+            llm_as_judge_405b_simpleqa.identifier: llm_as_judge_405b_simpleqa,
+        }
+
+    async def score_row(
+        self,
+        input_row: dict[str, Any],
+        scoring_fn_identifier: str | None = None,
+        scoring_params: ScoringFnParams | None = None,
+    ) -> ScoringResultRow:
+        assert scoring_fn_identifier is not None, "Scoring function identifier not found."
+        fn_def = self.supported_fn_defs_registry[scoring_fn_identifier]
+
+        # override params if scoring_params is provided
+        if scoring_params is not None:
+            fn_def.params = scoring_params
+
+        assert fn_def.params is not None, f"LLMAsJudgeparams not found for {fn_def}."
+        assert fn_def.params.prompt_template is not None, "LLM Judge prompt_template not found."
+        assert fn_def.params.judge_score_regexes is not None, "LLM Judge judge_score_regexes not found."
+
+        input_query = input_row["input_query"]
+        expected_answer = input_row["expected_answer"]
+        generated_answer = input_row["generated_answer"]
+
+        judge_input_msg = fn_def.params.prompt_template.format(
+            input_query=input_query,
+            expected_answer=expected_answer,
+            generated_answer=generated_answer,
+        )
+
+        params = OpenAIChatCompletionRequestWithExtraBody(
+            model=fn_def.params.judge_model,
+            messages=[
+                {
+                    "role": "user",
+                    "content": judge_input_msg,
+                }
+            ],
+        )
+        judge_response = await self.inference_api.openai_chat_completion(params)
+        content = judge_response.choices[0].message.content
+        rating_regexes = fn_def.params.judge_score_regexes
+
+        judge_rating = None
+        for regex in rating_regexes:
+            match = re.search(regex, content)
+            if match:
+                judge_rating = match.group(1)
+                break
+
+        return {
+            "score": judge_rating,
+            "judge_feedback": content,
+        }

llama_stack/providers/inline/telemetry/__init__.py

@@ -0,0 +1,5 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.

llama_stack/providers/inline/telemetry/meta_reference/__init__.py

@@ -0,0 +1,21 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from typing import Any
+
+from llama_stack.core.datatypes import Api
+
+from .config import TelemetryConfig, TelemetrySink
+
+__all__ = ["TelemetryConfig", "TelemetrySink"]
+
+
+async def get_provider_impl(config: TelemetryConfig, deps: dict[Api, Any]):
+    from .telemetry import TelemetryAdapter
+
+    impl = TelemetryAdapter(config, deps)
+    await impl.initialize()
+    return impl

llama_stack/providers/inline/telemetry/meta_reference/config.py

@@ -0,0 +1,47 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from enum import StrEnum
+from typing import Any
+
+from pydantic import BaseModel, Field, field_validator
+
+
+class TelemetrySink(StrEnum):
+    OTEL_TRACE = "otel_trace"
+    OTEL_METRIC = "otel_metric"
+    CONSOLE = "console"
+
+
+class TelemetryConfig(BaseModel):
+    otel_exporter_otlp_endpoint: str | None = Field(
+        default=None,
+        description="The OpenTelemetry collector endpoint URL (base URL for traces, metrics, and logs). If not set, the SDK will use OTEL_EXPORTER_OTLP_ENDPOINT environment variable.",
+    )
+    service_name: str = Field(
+        # service name is always the same, use zero-width space to avoid clutter
+        default="\u200b",
+        description="The service name to use for telemetry",
+    )
+    sinks: list[TelemetrySink] = Field(
+        default_factory=list,
+        description="List of telemetry sinks to enable (possible values: otel_trace, otel_metric, console)",
+    )
+
+    @field_validator("sinks", mode="before")
+    @classmethod
+    def validate_sinks(cls, v):
+        if isinstance(v, str):
+            return [TelemetrySink(sink.strip()) for sink in v.split(",")]
+        return v or []
+
+    @classmethod
+    def sample_run_config(cls, __distro_dir__: str) -> dict[str, Any]:
+        return {
+            "service_name": "${env.OTEL_SERVICE_NAME:=\u200b}",
+            "sinks": "${env.TELEMETRY_SINKS:=}",
+            "otel_exporter_otlp_endpoint": "${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}",
+        }