petitRADTRANS.sbi.task ====================== .. py:module:: petitRADTRANS.sbi.task .. autoapi-nested-parse:: Task definitions for simulation-based inference workflows. These classes bridge the user-facing retrieval configuration objects and the runtime-native simulation interfaces used by amortized inference. Classes ------- .. autoapisummary:: petitRADTRANS.sbi.task.NoiseModelConfig petitRADTRANS.sbi.task.ObservationValueConstraint petitRADTRANS.sbi.task.ObservationSchema petitRADTRANS.sbi.task.SimulationConfig petitRADTRANS.sbi.task.SBITask Module Contents --------------- .. py:class:: NoiseModelConfig Describe how observational noise is injected during simulation. Attributes: mode: Short identifier of the noise model implementation. parameters: Backend-specific parameters used to instantiate the noise model. seed: Optional deterministic seed for repeatable simulation pipelines. .. py:attribute:: mode :type: str :value: 'observational' .. py:attribute:: parameters :type: Mapping[str, Any] .. py:attribute:: seed :type: int | None :value: None .. py:class:: ObservationValueConstraint Admissible range for simulated observation values. Attributes: min_value: Optional lower bound for valid simulated values. max_value: Optional upper bound for valid simulated values. min_inclusive: Whether ``min_value`` is included in the valid interval. max_inclusive: Whether ``max_value`` is included in the valid interval. .. py:attribute:: min_value :type: float | None :value: None .. py:attribute:: max_value :type: float | None :value: None .. py:attribute:: min_inclusive :type: bool :value: True .. py:attribute:: max_inclusive :type: bool :value: True .. py:method:: to_payload() -> dict[str, Any] .. py:class:: ObservationSchema Capture the supported observation family for an amortized task. Attributes: dataset_names: Dataset identifiers included in the task. modalities: Mapping from dataset name to a short modality label such as ``'spectrum'``, ``'photometry'``, or ``'time_series'``. metadata: Additional task-level information required by encoders or benchmarks, such as instrument names or wavelength coverage. .. py:attribute:: dataset_names :type: tuple[str, Ellipsis] .. py:attribute:: modalities :type: Mapping[str, str] .. py:attribute:: metadata :type: Mapping[str, Any] .. py:class:: SimulationConfig Control how a task generates prior-predictive simulations. Attributes: batch_size: Number of parameter points produced per simulator call. n_simulations: Optional target number of simulations for dataset generation jobs. use_vectorized_runtime: Whether to prefer the JAX-vectorized runtime path when available. store_per_datapoint_log_likelihood: Persist log-likelihood components alongside simulated observations. noise_model: Noise model configuration applied after forward-model evaluation. observation_value_constraints: Optional per-dataset admissible value ranges enforced on deterministic projected observations before simulations are accepted. .. py:attribute:: batch_size :type: int :value: 256 .. py:attribute:: n_simulations :type: int | None :value: None .. py:attribute:: use_vectorized_runtime :type: bool :value: True .. py:attribute:: store_per_datapoint_log_likelihood :type: bool :value: False .. py:attribute:: noise_model :type: NoiseModelConfig .. py:attribute:: observation_value_constraints :type: Mapping[str, ObservationValueConstraint] .. py:class:: SBITask Bundle the immutable ingredients required for an SBI problem. The task owns the retrieval configuration, canonical parameter layout, observation schema, and simulation policy needed to generate training data and evaluate amortized posteriors. .. py:attribute:: name :type: str .. py:attribute:: retrieval_config :type: petitRADTRANS.retrieval.retrieval_config.RetrievalConfig .. py:attribute:: parameter_layout :type: petitRADTRANS.retrieval.runtime.ParameterLayout .. py:attribute:: observation_schema :type: ObservationSchema .. py:attribute:: simulation_config :type: SimulationConfig .. py:attribute:: task_version :type: str :value: '0.1.0' .. py:attribute:: preprocessing_version :type: str :value: '0.5.0' .. py:attribute:: forward_model_family :type: str :value: 'unknown' .. py:attribute:: petitradtrans_version :type: str .. py:attribute:: metadata :type: Mapping[str, Any] .. py:method:: from_retrieval_config(retrieval_config: petitRADTRANS.retrieval.retrieval_config.RetrievalConfig, simulation_config: SimulationConfig | None = None, metadata: Mapping[str, Any] | None = None) -> SBITask :classmethod: Create an SBI task from an existing retrieval configuration. Parameters ---------- retrieval_config: Retrieval configuration describing the free parameters, observation datasets, and runtime-native forward model family. simulation_config: Optional simulation policy overriding the default prior-predictive batch configuration. metadata: Optional task-level metadata. Reserved keys such as ``task_version``, ``preprocessing_version``, and ``forward_model_family`` override the default inferred values. Returns ------- SBITask Immutable SBI task description with parameter layout, observation schema, and deterministic task fingerprint payload. Notes ----- The task inspects the configured retrieval datasets to infer modalities and a forward-model-family string that later participates in artifact compatibility checks. .. py:property:: parameter_names :type: tuple[str, Ellipsis] Return the free parameter names in canonical task order. .. py:property:: line_opacity_modes :type: Mapping[str, str] Return the line-opacity mode for each configured dataset. .. py:property:: fingerprint_payload :type: Mapping[str, Any] Return the deterministic payload used to fingerprint the task family. .. py:property:: task_fingerprint :type: petitRADTRANS.sbi.compatibility.TaskFingerprint Return the deterministic fingerprint of the task family. .. py:method:: validate_observation_schema(observation_schema: ObservationSchema) -> petitRADTRANS.sbi.compatibility.CompatibilityReport Compare an external observation schema with the task expectation. Parameters ---------- observation_schema: Candidate schema to compare against the task's required datasets and modalities. Returns ------- CompatibilityReport Structured compatibility report containing mismatches, if any. .. py:method:: validate_artifact_metadata(artifact_metadata: Any) -> petitRADTRANS.sbi.compatibility.CompatibilityReport Compare a task with persisted artifact metadata. Parameters ---------- artifact_metadata: Artifact-like object exposing task and preprocessing provenance fields. ``Any`` is accepted to avoid a hard dependency cycle. Returns ------- CompatibilityReport Structured report describing whether the artifact is compatible with the current task definition. The method accepts ``Any`` to avoid a hard import dependency back from the task module to the artifact module. .. py:method:: build_observation_states(model_contract: str = 'differentiable') -> dict[str, petitRADTRANS.retrieval.runtime.ObservationState] Materialize immutable runtime observations for the task datasets. Parameters ---------- model_contract: Runtime contract requested from each retrieval dataset. Returns ------- dict[str, ObservationState] Mapping from dataset name to immutable runtime observation state. .. py:method:: build_runtime() -> petitRADTRANS.retrieval.runtime.RetrievalRuntime Construct a retrieval runtime matching the task definition. Returns ------- RetrievalRuntime Runtime object configured with the task's parameter layout and retrieval datasets. The default implementation mirrors the retrieval-runtime grouping logic used by the exact retrieval package and is sufficient for initial SBI simulation backends. .. py:method:: _infer_modality(data: petitRADTRANS.retrieval.data.Data) -> str :staticmethod: .. py:method:: _infer_forward_model_family(retrieval_config: petitRADTRANS.retrieval.retrieval_config.RetrievalConfig) -> str :staticmethod: