nemo-platform-plugin 0.1.0__py3-none-any.whl

nemo_platform/types/safe_synthesizer/evaluation_parameters_param.py

@@ -0,0 +1,66 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import TypedDict
+
+from ..._types import SequenceNotStr
+
+__all__ = ["EvaluationParametersParam"]
+
+
+class EvaluationParametersParam(TypedDict, total=False):
+    """Configuration for evaluating synthetic data quality and privacy.
+
+    This class controls which evaluation metrics are computed and how they are configured.
+    It includes privacy attack evaluations, statistical quality metrics, and downstream
+    machine learning performance assessments.
+    """
+
+    aia_enabled: bool
+    """Enable attribute inference attack evaluation for privacy assessment."""
+
+    enabled: bool
+    """Enable or disable evaluation."""
+
+    mandatory_columns: int
+    """Number of mandatory columns that must be used in evaluation."""
+
+    mia_enabled: bool
+    """Enable membership inference attack evaluation for privacy assessment."""
+
+    pii_replay_columns: SequenceNotStr[str]
+    """List of columns for PII Replay. If not provided, only entities will be used."""
+
+    pii_replay_enabled: bool
+    """Enable PII Replay detection."""
+
+    pii_replay_entities: SequenceNotStr[str]
+    """List of entities for PII Replay.
+
+    If not provided, default entities will be used.
+    """
+
+    quasi_identifier_count: int
+    """Number of quasi-identifiers to sample for privacy attacks."""
+
+    sqs_report_columns: int
+    """Number of columns to include in statistical quality reports."""
+
+    sqs_report_rows: int
+    """Number of rows to include in statistical quality reports."""

nemo_platform/types/safe_synthesizer/generate_parameters.py

@@ -0,0 +1,104 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from typing import Optional
+from typing_extensions import Literal
+
+from ..._models import BaseModel
+from .validation_parameters import ValidationParameters
+
+__all__ = ["GenerateParameters"]
+
+
+class GenerateParameters(BaseModel):
+    """Configuration parameters for synthetic data generation.
+
+    These parameters control how synthetic data is generated after the model is trained.
+    They affect the quality, diversity, and validity of the generated synthetic records.
+    """
+
+    attention_backend: Optional[str] = None
+    """The attention backend for the vLLM engine.
+
+    Common values: 'FLASHINFER', 'FLASH_ATTN', 'TRITON_ATTN', 'FLEX_ATTENTION'. If
+    `None` or 'auto', vLLM will auto-select the best available backend.
+    """
+
+    enforce_timeseries_fidelity: Optional[bool] = None
+    """
+    Enforce time-series fidelity by enforcing order, intervals, start and end times
+    of the records.
+    """
+
+    invalid_fraction_threshold: Optional[float] = None
+    """
+    The fraction of invalid records that will stop generation after the `patience`
+    limit is reached. Must be in [0, 1].
+    """
+
+    num_records: Optional[int] = None
+    """Number of records to generate."""
+
+    patience: Optional[int] = None
+    """
+    Number of consecutive generations where the `invalid_fraction_threshold` is
+    reached before stopping generation. Must be >= 1.
+    """
+
+    repetition_penalty: Optional[float] = None
+    """The value used to control the likelihood of the model repeating the same token.
+
+    Must be > 0.
+    """
+
+    structured_generation_backend: Optional[
+        Literal["auto", "xgrammar", "guidance", "outlines", "lm-format-enforcer"]
+    ] = None
+    """The backend used by vLLM when `use_structured_generation` is `True`.
+
+    Supported backends: 'outlines', 'guidance', 'xgrammar', 'lm-format-enforcer'.
+    'auto' will allow vLLM to choose the backend.
+    """
+
+    structured_generation_schema_method: Optional[Literal["regex", "json_schema"]] = None
+    """
+    The method used to generate the schema from your dataset and pass it to the
+    generation backend. 'regex' uses a custom regex construction method that tends
+    to be more comprehensive than 'json_schema' at the cost of speed.
+    """
+
+    structured_generation_use_single_sequence: Optional[bool] = None
+    """
+    Whether to use a regex that matches exactly one sequence or record if
+    `max_sequences_per_example` is 1.
+    """
+
+    temperature: Optional[float] = None
+    """Sampling temperature for controlling randomness (higher = more random)."""
+
+    top_p: Optional[float] = None
+    """Nucleus sampling probability for token selection. Must be in (0, 1]."""
+
+    use_structured_generation: Optional[bool] = None
+    """Whether to use structured generation for better format control."""
+
+    validation: Optional[ValidationParameters] = None
+    """Configuration for record and sequence validation.
+
+    These parameters control the validation and automatic fixes when going from LLM
+    output to tabular data.
+    """

nemo_platform/types/safe_synthesizer/generate_parameters_param.py

@@ -0,0 +1,102 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import Literal, TypedDict
+
+from .validation_parameters_param import ValidationParametersParam
+
+__all__ = ["GenerateParametersParam"]
+
+
+class GenerateParametersParam(TypedDict, total=False):
+    """Configuration parameters for synthetic data generation.
+
+    These parameters control how synthetic data is generated after the model is trained.
+    They affect the quality, diversity, and validity of the generated synthetic records.
+    """
+
+    attention_backend: str
+    """The attention backend for the vLLM engine.
+
+    Common values: 'FLASHINFER', 'FLASH_ATTN', 'TRITON_ATTN', 'FLEX_ATTENTION'. If
+    `None` or 'auto', vLLM will auto-select the best available backend.
+    """
+
+    enforce_timeseries_fidelity: bool
+    """
+    Enforce time-series fidelity by enforcing order, intervals, start and end times
+    of the records.
+    """
+
+    invalid_fraction_threshold: float
+    """
+    The fraction of invalid records that will stop generation after the `patience`
+    limit is reached. Must be in [0, 1].
+    """
+
+    num_records: int
+    """Number of records to generate."""
+
+    patience: int
+    """
+    Number of consecutive generations where the `invalid_fraction_threshold` is
+    reached before stopping generation. Must be >= 1.
+    """
+
+    repetition_penalty: float
+    """The value used to control the likelihood of the model repeating the same token.
+
+    Must be > 0.
+    """
+
+    structured_generation_backend: Literal["auto", "xgrammar", "guidance", "outlines", "lm-format-enforcer"]
+    """The backend used by vLLM when `use_structured_generation` is `True`.
+
+    Supported backends: 'outlines', 'guidance', 'xgrammar', 'lm-format-enforcer'.
+    'auto' will allow vLLM to choose the backend.
+    """
+
+    structured_generation_schema_method: Literal["regex", "json_schema"]
+    """
+    The method used to generate the schema from your dataset and pass it to the
+    generation backend. 'regex' uses a custom regex construction method that tends
+    to be more comprehensive than 'json_schema' at the cost of speed.
+    """
+
+    structured_generation_use_single_sequence: bool
+    """
+    Whether to use a regex that matches exactly one sequence or record if
+    `max_sequences_per_example` is 1.
+    """
+
+    temperature: float
+    """Sampling temperature for controlling randomness (higher = more random)."""
+
+    top_p: float
+    """Nucleus sampling probability for token selection. Must be in (0, 1]."""
+
+    use_structured_generation: bool
+    """Whether to use structured generation for better format control."""
+
+    validation: ValidationParametersParam
+    """Configuration for record and sequence validation.
+
+    These parameters control the validation and automatic fixes when going from LLM
+    output to tabular data.
+    """

nemo_platform/types/safe_synthesizer/gliner_config.py

@@ -0,0 +1,41 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from typing import Optional
+
+from ..._models import BaseModel
+
+__all__ = ["GlinerConfig"]
+
+
+class GlinerConfig(BaseModel):
+    """Configuration for the GLiNER named-entity recognition model."""
+
+    batch_size: Optional[int] = None
+    """GLiNER batch size."""
+
+    chunk_length: Optional[int] = None
+    """GLiNER batch chunk length in characters."""
+
+    enable_batch_mode: Optional[bool] = None
+    """Enable GLiNER batch mode."""
+
+    enable_gliner: Optional[bool] = None
+    """Enable GLiNER NER module."""
+
+    gliner_model: Optional[str] = None
+    """GLiNER model name."""

nemo_platform/types/safe_synthesizer/gliner_config_param.py

@@ -0,0 +1,41 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import TypedDict
+
+__all__ = ["GlinerConfigParam"]
+
+
+class GlinerConfigParam(TypedDict, total=False):
+    """Configuration for the GLiNER named-entity recognition model."""
+
+    batch_size: int
+    """GLiNER batch size."""
+
+    chunk_length: int
+    """GLiNER batch chunk length in characters."""
+
+    enable_batch_mode: bool
+    """Enable GLiNER batch mode."""
+
+    enable_gliner: bool
+    """Enable GLiNER NER module."""
+
+    gliner_model: str
+    """GLiNER model name."""

nemo_platform/types/safe_synthesizer/globals.py

@@ -0,0 +1,45 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from typing import List, Optional
+
+from ..._models import BaseModel
+from .ner_config import NERConfig
+from .classify_config import ClassifyConfig
+
+__all__ = ["Globals"]
+
+
+class Globals(BaseModel):
+    """
+    Global settings for the PII replacer including locales, seed, NER, and classification.
+    """
+
+    classify: Optional[ClassifyConfig] = None
+    """Configuration for column classification using an LLM."""
+
+    locales: Optional[List[str]] = None
+    """List of locales."""
+
+    lock_columns: Optional[List[str]] = None
+    """List of columns to preserve as immutable across all transformations."""
+
+    ner: Optional[NERConfig] = None
+    """Configuration for Named Entity Recognition."""
+
+    seed: Optional[int] = None
+    """Optional random seed."""

nemo_platform/types/safe_synthesizer/globals_param.py

@@ -0,0 +1,47 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import TypedDict
+
+from ..._types import SequenceNotStr
+from .ner_config_param import NERConfigParam
+from .classify_config_param import ClassifyConfigParam
+
+__all__ = ["GlobalsParam"]
+
+
+class GlobalsParam(TypedDict, total=False):
+    """
+    Global settings for the PII replacer including locales, seed, NER, and classification.
+    """
+
+    classify: ClassifyConfigParam
+    """Configuration for column classification using an LLM."""
+
+    locales: SequenceNotStr[str]
+    """List of locales."""
+
+    lock_columns: SequenceNotStr[str]
+    """List of columns to preserve as immutable across all transformations."""
+
+    ner: NERConfigParam
+    """Configuration for Named Entity Recognition."""
+
+    seed: int
+    """Optional random seed."""

nemo_platform/types/safe_synthesizer/job_create_params.py

@@ -0,0 +1,46 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing import Dict
+from typing_extensions import Required, TypedDict
+
+from .safe_synthesizer_job_config_param import SafeSynthesizerJobConfigParam
+
+__all__ = ["JobCreateParams"]
+
+
+class JobCreateParams(TypedDict, total=False):
+    workspace: str
+
+    spec: Required[SafeSynthesizerJobConfigParam]
+    """Configuration model for Safe Synthesizer jobs.
+
+    Used primarily internally to configure a run submitted to the NeMo Jobs
+    Microservice.
+    """
+
+    custom_fields: Dict[str, object]
+
+    description: str
+
+    name: str
+
+    ownership: Dict[str, object]
+
+    project: str

nemo_platform/types/safe_synthesizer/job_get_logs_params.py

@@ -0,0 +1,30 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import TypedDict
+
+__all__ = ["JobGetLogsParams"]
+
+
+class JobGetLogsParams(TypedDict, total=False):
+    workspace: str
+
+    limit: int
+
+    page_cursor: str

nemo_platform/types/safe_synthesizer/job_list_params.py

@@ -0,0 +1,44 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing_extensions import TypedDict
+
+from .safe_synthesizer_jobs_sort_field import SafeSynthesizerJobsSortField
+from .safe_synthesizer_jobs_list_filter_param import SafeSynthesizerJobsListFilterParam
+
+__all__ = ["JobListParams"]
+
+
+class JobListParams(TypedDict, total=False):
+    workspace: str
+
+    filter: SafeSynthesizerJobsListFilterParam
+    """Filter jobs on various criteria."""
+
+    page: int
+    """Page number."""
+
+    page_size: int
+    """Page size."""
+
+    sort: SafeSynthesizerJobsSortField
+    """The field to sort by.
+
+    To sort in decreasing order, use `-` in front of the field name.
+    """

nemo_platform/types/safe_synthesizer/jobs/__init__.py

@@ -0,0 +1,21 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from .safe_synthesizer_timing import SafeSynthesizerTiming as SafeSynthesizerTiming
+from .safe_synthesizer_summary import SafeSynthesizerSummary as SafeSynthesizerSummary