nv-ingest-api 2025.8.19.dev20250819__py3-none-any.whl → 2025.8.21.dev20250821__py3-none-any.whl

nv_ingest_api/interface/__init__.py

@@ -11,6 +11,7 @@ from typing import Dict, Any, Optional, List
 from pydantic import BaseModel
 
 from nv_ingest_api.internal.schemas.extract.extract_pdf_schema import PDFiumConfigSchema, NemoRetrieverParseConfigSchema
+from nv_ingest_api.util.logging.sanitize import sanitize_for_logging
 
 logger = logging.getLogger(__name__)
 
@@ -180,29 +181,31 @@ def extraction_interface_relay_constructor(api_fn, task_keys: Optional[List[str]
             if extractor_schema is None:
                 extractor_schema = {f"{extract_method}_config": extraction_config_dict}
 
-            # Log the task and extractor configurations for debugging
+            # Log the task and extractor configurations for debugging (sanitized)
             logger.debug("\n" + "=" * 80)
             logger.debug(f"DEBUG - API Function: {api_fn.__name__}")
             logger.debug(f"DEBUG - Extract Method: {extract_method}")
             logger.debug("-" * 80)
 
-            # Format the task config as a string and log it
-            task_config_str = pprint.pformat(task_config, width=100, sort_dicts=False)
-            logger.debug(f"DEBUG - Task Config:\n{task_config_str}")
+            # Sanitize and format the task config as a string and log it
+            sanitized_task_config = sanitize_for_logging(task_config)
+            task_config_str = pprint.pformat(sanitized_task_config, width=100, sort_dicts=False)
+            logger.debug(f"DEBUG - Task Config (sanitized):\n{task_config_str}")
             logger.debug("-" * 80)
 
-            # Format the extractor config as a string and log it
+            # Sanitize and format the extractor config as a string and log it
             if hasattr(extractor_schema, "model_dump"):
-                extractor_config_str = pprint.pformat(extractor_schema.model_dump(), width=100, sort_dicts=False)
+                sanitized_extractor_config = sanitize_for_logging(extractor_schema.model_dump())
             else:
-                extractor_config_str = pprint.pformat(extractor_schema, width=100, sort_dicts=False)
+                sanitized_extractor_config = sanitize_for_logging(extractor_schema)
+            extractor_config_str = pprint.pformat(sanitized_extractor_config, width=100, sort_dicts=False)
             logger.debug(f"DEBUG - Extractor Config Type: {type(extractor_schema)}")
-            logger.debug(f"DEBUG - Extractor Config:\n{extractor_config_str}")
+            logger.debug(f"DEBUG - Extractor Config (sanitized):\n{extractor_config_str}")
             logger.debug("=" * 80 + "\n")
 
-            # Call the backend API function.
-            pprint.pprint(task_config)
-            pprint.pprint(extractor_schema)
+            # Call the backend API function. Print sanitized configs for any debug consumers of stdout.
+            pprint.pprint(sanitized_task_config)
+            pprint.pprint(sanitized_extractor_config)
             result = api_fn(ledger, task_config, extractor_schema, execution_trace_log)
 
             # If the result is a tuple, return only the first element

nv_ingest_api/internal/extract/pdf/engines/llama.py

@@ -193,7 +193,10 @@ async def async_llama_parse(
         A string of extracted text.
     """
     base_url = "https://api.cloud.llamaindex.ai/api/parsing"
-    headers = {"Authorization": f"Bearer {api_key}"}
+    # Normalize in case api_key contains only whitespace; avoid sending an empty bearer token
+    _token = (api_key or "").strip()
+    _auth_value = f"Bearer {_token}" if _token else "Bearer <no key provided>"
+    headers = {"Authorization": _auth_value}
     mime_type = "application/pdf"
 
     try:

nv_ingest_api/internal/extract/pdf/engines/pdf_helpers/__init__.py

@@ -11,6 +11,7 @@ from typing import Any
 from typing import Dict
 from typing import List
 from typing import Optional
+from nv_ingest_api.util.logging.sanitize import sanitize_for_logging
 
 import pandas as pd
 from nv_ingest_api.internal.extract.pdf.engines import adobe_extractor
@@ -131,7 +132,7 @@ def _orchestrate_row_extraction(
         method_config = extractor_config[config_key]
     else:
         # If no matching config is found, log a warning but don't fail
-        logger.warning(f"No {config_key} found in extractor_config: {extractor_config}")
+        logger.warning(f"No {config_key} found in extractor_config: {sanitize_for_logging(extractor_config)}")
         method_config = None
 
     # Add the method-specific config to the parameters if available
@@ -141,7 +142,7 @@ def _orchestrate_row_extraction(
 
     # The resulting parameters constitute the complete extractor_config
     extractor_config = params
-    logger.debug(f"Final extractor_config: {extractor_config}")
+    logger.debug(f"Final extractor_config: {sanitize_for_logging(extractor_config)}")
 
     result = _work_extract_pdf(
         pdf_stream=pdf_stream,

nv_ingest_api/internal/schemas/extract/extract_audio_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 from pydantic import root_validator
 
 logger = logging.getLogger(__name__)
@@ -42,12 +42,12 @@ class AudioConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
     audio_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     audio_infer_protocol: Optional[str] = None
     function_id: Optional[str] = None
     use_ssl: Optional[bool] = None
-    ssl_cert: Optional[str] = None
+    ssl_cert: Optional[str] = Field(default=None, repr=False)
     segment_audio: Optional[bool] = None
 
     @root_validator(pre=True)

nv_ingest_api/internal/schemas/extract/extract_chart_schema.py

@@ -6,7 +6,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import field_validator, model_validator, ConfigDict, BaseModel
+from pydantic import field_validator, model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -44,7 +44,7 @@ class ChartExtractorConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_docx_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import model_validator, ConfigDict, BaseModel
+from pydantic import model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -41,7 +41,7 @@ class DocxConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_image_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import model_validator, ConfigDict, BaseModel
+from pydantic import model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -41,7 +41,7 @@ class ImageConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_infographic_schema.py

@@ -6,7 +6,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import field_validator, model_validator, ConfigDict, BaseModel
+from pydantic import field_validator, model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -40,7 +40,7 @@ class InfographicExtractorConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     ocr_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     ocr_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_pdf_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import model_validator, ConfigDict, BaseModel
+from pydantic import model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -41,7 +41,7 @@ class PDFiumConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""
@@ -123,7 +123,7 @@ class NemoRetrieverParseConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_pptx_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import model_validator, ConfigDict, BaseModel
+from pydantic import model_validator, ConfigDict, BaseModel, Field
 
 logger = logging.getLogger(__name__)
 
@@ -41,7 +41,7 @@ class PPTXConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/extract/extract_table_schema.py

@@ -7,7 +7,7 @@ import logging
 from typing import Optional
 from typing import Tuple
 
-from pydantic import field_validator, model_validator, ConfigDict, BaseModel
+from pydantic import field_validator, model_validator, ConfigDict, BaseModel, Field
 
 
 logger = logging.getLogger(__name__)
@@ -42,7 +42,7 @@ class TableExtractorConfigSchema(BaseModel):
         Pydantic config option to forbid extra fields.
     """
 
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
 
     yolox_endpoints: Tuple[Optional[str], Optional[str]] = (None, None)
     yolox_infer_protocol: str = ""

nv_ingest_api/internal/schemas/meta/ingest_job_schema.py

@@ -73,7 +73,7 @@ class IngestTaskStoreSchema(BaseModelNoExt):
 
 # Captioning: All fields are optional and override default parameters.
 class IngestTaskCaptionSchema(BaseModelNoExt):
-    api_key: Optional[str] = None
+    api_key: Optional[str] = Field(default=None, repr=False)
     endpoint_url: Optional[str] = None
     prompt: Optional[str] = None
     model_name: Optional[str] = None
@@ -105,7 +105,7 @@ class IngestTaskDedupSchema(BaseModelNoExt):
 class IngestTaskEmbedSchema(BaseModelNoExt):
     endpoint_url: Optional[str] = None
     model_name: Optional[str] = None
-    api_key: Optional[str] = None
+    api_key: Optional[str] = Field(default=None, repr=False)
     filter_errors: bool = False
     text_elements_modality: Optional[str] = None
     image_elements_modality: Optional[str] = None
@@ -121,13 +121,13 @@ class IngestTaskVdbUploadSchema(BaseModelNoExt):
 
 
 class IngestTaskAudioExtraction(BaseModelNoExt):
-    auth_token: Optional[str] = None
+    auth_token: Optional[str] = Field(default=None, repr=False)
     grpc_endpoint: Optional[str] = None
     http_endpoint: Optional[str] = None
     infer_protocol: Optional[str] = None
     function_id: Optional[str] = None
     use_ssl: Optional[bool] = None
-    ssl_cert: Optional[str] = None
+    ssl_cert: Optional[str] = Field(default=None, repr=False)
     segment_audio: Optional[bool] = None
 
 

nv_ingest_api/internal/schemas/transform/transform_image_caption_schema.py

@@ -3,13 +3,31 @@
 # SPDX-License-Identifier: Apache-2.0
 
 
-from pydantic import ConfigDict, BaseModel
+from pydantic import ConfigDict, BaseModel, model_validator, field_validator, Field
 
 
 class ImageCaptionExtractionSchema(BaseModel):
-    api_key: str = "api_key"
+    api_key: str = Field(default="", repr=False)
     endpoint_url: str = "https://integrate.api.nvidia.com/v1/chat/completions"
     prompt: str = "Caption the content of this image:"
     model_name: str = "nvidia/llama-3.1-nemotron-nano-vl-8b-v1"
     raise_on_failure: bool = False
     model_config = ConfigDict(extra="forbid")
+
+    @field_validator("api_key", mode="before")
+    @classmethod
+    def _coerce_api_key_none(cls, v):
+        return "" if v is None else v
+
+    @model_validator(mode="before")
+    @classmethod
+    def _coerce_none_to_empty(cls, values):
+        """Allow None for string fields where empty string is acceptable.
+
+        Specifically, convert api_key=None to api_key="" so validation passes
+        when no API key is supplied.
+        """
+        if isinstance(values, dict):
+            if values.get("api_key") is None:
+                values["api_key"] = ""
+        return values

nv_ingest_api/internal/schemas/transform/transform_text_embedding_schema.py

@@ -5,7 +5,7 @@
 
 import logging
 
-from pydantic import ConfigDict, BaseModel, Field
+from pydantic import ConfigDict, BaseModel, Field, model_validator, field_validator
 
 from nv_ingest_api.util.logging.configuration import LogLevel
 
@@ -13,7 +13,7 @@ logger = logging.getLogger(__name__)
 
 
 class TextEmbeddingSchema(BaseModel):
-    api_key: str = Field(default="api_key")
+    api_key: str = Field(default="", repr=False)
     batch_size: int = Field(default=4)
     embedding_model: str = Field(default="nvidia/llama-3.2-nv-embedqa-1b-v2")
     embedding_nim_endpoint: str = Field(default="http://embedding:8000/v1")
@@ -28,3 +28,16 @@ class TextEmbeddingSchema(BaseModel):
     audio_elements_modality: str = Field(default="text")
 
     model_config = ConfigDict(extra="forbid")
+
+    @field_validator("api_key", mode="before")
+    @classmethod
+    def _coerce_api_key_none(cls, v):
+        return "" if v is None else v
+
+    @model_validator(mode="before")
+    @classmethod
+    def _coerce_none_to_empty(cls, values):
+        """Convert api_key=None to empty string so validation passes when key is omitted."""
+        if isinstance(values, dict) and values.get("api_key") is None:
+            values["api_key"] = ""
+        return values

nv_ingest_api/internal/transform/embed_text.py

@@ -75,8 +75,11 @@ def _make_async_request(
     response = {}
 
     try:
+        # Normalize API key to avoid sending an empty bearer token via SDK internals
+        _token = (api_key or "").strip()
+        _api_key = _token if _token else "<no key provided>"
         client = OpenAI(
-            api_key=api_key,
+            api_key=_api_key,
             base_url=embedding_nim_endpoint,
         )
 

nv_ingest_api/util/logging/sanitize.py

@@ -0,0 +1,84 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-25, NVIDIA CORPORATION & AFFILIATES.
+# All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+from __future__ import annotations
+
+from typing import Any, Mapping, MutableMapping, Sequence, Set
+
+try:
+    # Pydantic is optional at runtime for this helper; import if available
+    from pydantic import BaseModel  # type: ignore
+except Exception:  # pragma: no cover - pydantic always present in this repo
+    BaseModel = None  # type: ignore
+
+
+_DEFAULT_SENSITIVE_KEYS: Set[str] = {
+    "access_token",
+    "api_key",
+    "authorization",
+    "auth_token",
+    "client_secret",
+    "hf_access_token",
+    "hugging_face_access_token",
+    "password",
+    "refresh_token",
+    "secret",
+    "ssl_cert",
+    "x-api-key",
+}
+
+_REDACTION = "***REDACTED***"
+
+
+def _is_mapping(obj: Any) -> bool:
+    try:
+        return isinstance(obj, Mapping)
+    except Exception:
+        return False
+
+
+def _is_sequence(obj: Any) -> bool:
+    # Exclude strings/bytes from sequences we want to traverse
+    return isinstance(obj, Sequence) and not isinstance(obj, (str, bytes, bytearray))
+
+
+def sanitize_for_logging(
+    data: Any,
+    sensitive_keys: Set[str] | None = None,
+    redaction: str = _REDACTION,
+) -> Any:
+    """
+    Recursively sanitize common secret fields from dicts, lists, tuples, and Pydantic models.
+
+    - Key comparison is case-insensitive and matches exact keys only.
+    - Does not mutate input; returns a sanitized deep copy.
+    - For Pydantic BaseModel instances, uses model_dump() before redaction.
+    """
+    keys = {k.lower() for k in (sensitive_keys or _DEFAULT_SENSITIVE_KEYS)}
+
+    # Handle Pydantic models without importing pydantic at module import time
+    if BaseModel is not None and isinstance(data, BaseModel):  # type: ignore[arg-type]
+        try:
+            return sanitize_for_logging(data.model_dump(), keys, redaction)
+        except Exception:
+            # Fall through and try generic handling below
+            pass
+
+    # Dict-like
+    if _is_mapping(data):
+        out: MutableMapping[str, Any] = type(data)()  # preserve mapping type where possible
+        for k, v in data.items():  # type: ignore[assignment]
+            key_lower = str(k).lower()
+            if key_lower in keys:
+                out[k] = redaction
+            else:
+                out[k] = sanitize_for_logging(v, keys, redaction)
+        return out
+
+    # List/Tuple/Sequence
+    if _is_sequence(data):
+        return type(data)(sanitize_for_logging(v, keys, redaction) for v in data)
+
+    # Fallback: return as-is
+    return data

{nv_ingest_api-2025.8.19.dev20250819.dist-info → nv_ingest_api-2025.8.21.dev20250821.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest-api
-Version: 2025.8.19.dev20250819
+Version: 2025.8.21.dev20250821
 Summary: Python module with core document ingestion functions.
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

{nv_ingest_api-2025.8.19.dev20250819.dist-info → nv_ingest_api-2025.8.21.dev20250821.dist-info}/RECORD

@@ -1,5 +1,5 @@
 nv_ingest_api/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest_api/interface/__init__.py,sha256=ltWlfmtCewHSRK4B7DF__QvlSUPuliz58JEcEIeIgI0,10134
+nv_ingest_api/interface/__init__.py,sha256=KPYjTT0p5I3aCGg6Bzs8igQQuXRY4mmUUVaK0F0uziI,10497
 nv_ingest_api/interface/extract.py,sha256=o9OdoWxYsj-O4HsDe6wWbyd69OAueb2rlMtKSzOrKZo,38743
 nv_ingest_api/interface/mutate.py,sha256=eZkd3sbHEJQiEPJyMbhewlPxQNMnL_Xur15icclnb-U,5934
 nv_ingest_api/interface/store.py,sha256=aR3Cf19lq9Yo9AHlAy1VVcrOP2dgyN01yYhwxyTprkQ,8207
@@ -30,12 +30,12 @@ nv_ingest_api/internal/extract/pdf/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusX
 nv_ingest_api/internal/extract/pdf/pdf_extractor.py,sha256=CxtWaD6mql9MEqSdk2CfSQ9T-Bn87beBkCOuGGjxGt8,2934
 nv_ingest_api/internal/extract/pdf/engines/__init__.py,sha256=u4GnAZmDKRl0RwYGIRiozIRw70Kybw3A72-lcKFeoTI,582
 nv_ingest_api/internal/extract/pdf/engines/adobe.py,sha256=VT0dEqkU-y2uGkaCqxtKYov_Q8R1028UQVBchgMLca4,17466
-nv_ingest_api/internal/extract/pdf/engines/llama.py,sha256=PpKTqS8jGHBV6mKLGZWwjpfT8ga6Fy8ffrvL-gPAf2c,8182
+nv_ingest_api/internal/extract/pdf/engines/llama.py,sha256=MwzM-n2tu0FHM0wDe_0mONLlzHrPte7EOTuPtzCh7Zs,8384
 nv_ingest_api/internal/extract/pdf/engines/nemoretriever.py,sha256=IVbNcH_phMiRSxnkZ04pGfQrPJ-x1zVR3hXyhxv7juc,22977
 nv_ingest_api/internal/extract/pdf/engines/pdfium.py,sha256=SKmias2iZmAE6Q8WXxmFEjvLOZy-vXRoaRIPpi7Tuhs,22962
 nv_ingest_api/internal/extract/pdf/engines/tika.py,sha256=6GyR2l6EsgNZl9jnYDXLeKNK9Fj2Mw9y2UWDq-eSkOc,3169
 nv_ingest_api/internal/extract/pdf/engines/unstructured_io.py,sha256=jrv2B4VZAH4PevAQrFz965qz8UyXq3rViiOTbGLejec,14908
-nv_ingest_api/internal/extract/pdf/engines/pdf_helpers/__init__.py,sha256=4bvN6LsPksLicI6jM0JqbJFiOZNHEcuc8MVVW4XfgV8,5875
+nv_ingest_api/internal/extract/pdf/engines/pdf_helpers/__init__.py,sha256=uTPTUTWQsGM1oeTUo49_hzwC5Yy9iEokrnS3z3WvtIo,5988
 nv_ingest_api/internal/extract/pptx/__init__.py,sha256=HIHfzSig66GT0Uk8qsGBm_f13fKYcPtItBicRUWOOVA,183
 nv_ingest_api/internal/extract/pptx/pptx_extractor.py,sha256=o-0P2dDyRFW37uQi_lKk6-eFozTcZvbq-2Y4I0EBMIY,7749
 nv_ingest_api/internal/extract/pptx/engines/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -69,22 +69,22 @@ nv_ingest_api/internal/primitives/tracing/logging.py,sha256=SSzIgS7afLH-e1C7VagY
 nv_ingest_api/internal/primitives/tracing/tagging.py,sha256=xU534rb94uKnsSu0_DzyZcCSkIpa5SWTMxX7NSA3HoE,11671
 nv_ingest_api/internal/schemas/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/internal/schemas/extract/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest_api/internal/schemas/extract/extract_audio_schema.py,sha256=W-nEBriqiNkjpaQ5AT_8LhtVXlW8AhlcftmoeQQtKAs,3812
-nv_ingest_api/internal/schemas/extract/extract_chart_schema.py,sha256=wDcvQ5XtOjIBGSWtNjQPiPtVKrSOYqbf2mnLrhfhue4,4283
-nv_ingest_api/internal/schemas/extract/extract_docx_schema.py,sha256=M2N7WjMNvSemHcJHWeNUD_kFG0wC5VE2W3K6SVrJqvA,3761
+nv_ingest_api/internal/schemas/extract/extract_audio_schema.py,sha256=3cjLcw5zFUb7WNbGPLDJN2KukF_hoOM7PE33UHici6w,3873
+nv_ingest_api/internal/schemas/extract/extract_chart_schema.py,sha256=PZFJPLrLs8k5I5ufnp0XWrBjmbQVkkaxjb-xq-2rn2Q,4317
+nv_ingest_api/internal/schemas/extract/extract_docx_schema.py,sha256=Bafw6lIXLS2PcEpU82D4Vb0OPD_FvGSr546IedsfR8o,3795
 nv_ingest_api/internal/schemas/extract/extract_html_schema.py,sha256=lazpONTGZ6Fl420BGBAr6rogFGtlzBiZTc1uA694OIs,841
-nv_ingest_api/internal/schemas/extract/extract_image_schema.py,sha256=GC4xV8Z9TPLOuxlEtf2fbklSSp8ETGMrDpZgMQ02UwA,3766
-nv_ingest_api/internal/schemas/extract/extract_infographic_schema.py,sha256=z42cs7w-U-IUCMGByp5e_iBUZ7KCl5vTIXkP64ty6gY,3973
-nv_ingest_api/internal/schemas/extract/extract_pdf_schema.py,sha256=G9g1lEORmryUWTzDyZ0vHAuPnVMK7VaRx0E4xzmAw3Q,6589
-nv_ingest_api/internal/schemas/extract/extract_pptx_schema.py,sha256=5dT0kv-Mmpe5KW-BZc1JOW3rUlgzVZI0rpB79NWytmw,3761
-nv_ingest_api/internal/schemas/extract/extract_table_schema.py,sha256=vd_1mf_LmQGvSTpQCuWr6ubsiav4TMhp_SpKGO-6RLc,3935
+nv_ingest_api/internal/schemas/extract/extract_image_schema.py,sha256=hXiHIKIZS2qb4u8g10m_S5CpeAzHvH3tSB_He5LYOfU,3800
+nv_ingest_api/internal/schemas/extract/extract_infographic_schema.py,sha256=Yb580-k1oYXanWMA17u0LGO5-AzUP80j6aTPGpjn920,4007
+nv_ingest_api/internal/schemas/extract/extract_pdf_schema.py,sha256=TTFzlPG5r1QFO2B5VWBPF69IXP1eQKrg16wo49hHDV0,6650
+nv_ingest_api/internal/schemas/extract/extract_pptx_schema.py,sha256=6KomEL6wdhpZVnE2SiOeZJNhyfHIQ2sARdfm0R16uCM,3795
+nv_ingest_api/internal/schemas/extract/extract_table_schema.py,sha256=npBEGXkVaShOiHrc8FQ-25kHhPRD8WB6dRn-T5TScKc,3969
 nv_ingest_api/internal/schemas/message_brokers/__init__.py,sha256=uLsBITo_XfgbwpzqXUm1IYX6XlZrTfx6T1cIhdILwG8,140
 nv_ingest_api/internal/schemas/message_brokers/message_broker_client_schema.py,sha256=4xTSFE_vH7yZE9RRJRflFAG9hNXIaF6K020M_xA7ylw,1351
 nv_ingest_api/internal/schemas/message_brokers/request_schema.py,sha256=LZX_wXDxTamVFqTQs2Yd8uvWyPE5mddHAWSU4PtfEIQ,966
 nv_ingest_api/internal/schemas/message_brokers/response_schema.py,sha256=4b275HlzBSzpmuE2wdoeaGKPCdKki3wuWldtRIfrj8w,727
 nv_ingest_api/internal/schemas/meta/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/internal/schemas/meta/base_model_noext.py,sha256=8hXU1uuiqZ6t8EsoZ8vlC5EFf2zSZrKEX133FcfZMwI,316
-nv_ingest_api/internal/schemas/meta/ingest_job_schema.py,sha256=En7wcLPB6hvfAXx2-xZM49wbJXmRX2Ckc0i6edqn21c,10145
+nv_ingest_api/internal/schemas/meta/ingest_job_schema.py,sha256=cIpoesvIs0dR6s8dGjGHL246k5kf7hDmdhA48i8Si7s,10253
 nv_ingest_api/internal/schemas/meta/metadata_schema.py,sha256=VnAzkSFat_ckI19mlwQTlFrvP6EZVCwyNl9bt51b8oU,7193
 nv_ingest_api/internal/schemas/meta/udf.py,sha256=GgzqbZOlipQgMpDhbXLqbF8xrHenj_hMNqhR_P-1ynw,779
 nv_ingest_api/internal/schemas/mutate/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
@@ -93,16 +93,16 @@ nv_ingest_api/internal/schemas/store/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQu
 nv_ingest_api/internal/schemas/store/store_embedding_schema.py,sha256=tdKeiraim9CDL9htgp4oUSCoPMoO5PrHBnlXqDyCpMw,956
 nv_ingest_api/internal/schemas/store/store_image_schema.py,sha256=p2LGij9i6sG6RYmsfdiQOiWIc2j-POjxYrNuMrp3ELU,1010
 nv_ingest_api/internal/schemas/transform/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest_api/internal/schemas/transform/transform_image_caption_schema.py,sha256=gRJCfEGrJXErMF_GsZDjyDR8HOrLrUNxz2BVS6LSPY8,542
+nv_ingest_api/internal/schemas/transform/transform_image_caption_schema.py,sha256=fRMRwcWP-L8sfv2enNDt_W_CL0eC2i3b_1VCCtmr1K8,1188
 nv_ingest_api/internal/schemas/transform/transform_image_filter_schema.py,sha256=31ThI5fr0yyENeJeE1xMAA-pxk1QVJLwM842zMate_k,429
-nv_ingest_api/internal/schemas/transform/transform_text_embedding_schema.py,sha256=RZCISA8CUqKiY8eJuk4uWxzo4PZ-fuYdzMO7_LYFkoM,1117
+nv_ingest_api/internal/schemas/transform/transform_text_embedding_schema.py,sha256=rzdhRANCqG9mOEoLargznuBwj1-MbEQUu2LDVi5vl50,1616
 nv_ingest_api/internal/schemas/transform/transform_text_splitter_schema.py,sha256=D9K8tvu-tkEBQkZo7uuRzgrHdGyM3ZcNycHbHy5HV2E,791
 nv_ingest_api/internal/store/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/internal/store/embed_text_upload.py,sha256=maxb4FPsBvWgvlrjAPEBlRZEFdJX5NxPG-p8kUbzV7I,9898
 nv_ingest_api/internal/store/image_upload.py,sha256=GNlY4k3pfcHv3lzXxkbmGLeHFsf9PI25bkBn6Xn9h3I,9654
 nv_ingest_api/internal/transform/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/internal/transform/caption_image.py,sha256=0ILCG2F8ESqKtZiPUM-6F1BHUflFZ76Dzi2GNzkE-lU,8517
-nv_ingest_api/internal/transform/embed_text.py,sha256=AdUXVosBHuZadaFLi9_RVB7vB_hjTziCLUeBZ1oWGZo,20003
+nv_ingest_api/internal/transform/embed_text.py,sha256=LB_2Zvw6plc7uOWT2QN13aDu2qFumXzl_RB3ZcZbLGs,20191
 nv_ingest_api/internal/transform/split_text.py,sha256=LAtInGVuydH43UwjNMQWFVC1A6NdhXP_dZup2xX4qEo,7745
 nv_ingest_api/util/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/util/control_message/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -135,6 +135,7 @@ nv_ingest_api/util/introspection/class_inspect.py,sha256=sEYe37ICHdhXxSbD0JTCDg3
 nv_ingest_api/util/introspection/function_inspect.py,sha256=_yCUUT1x3tLCsbXmUpTv_O8Qi6af4NiCpJ1qGfcksvw,2066
 nv_ingest_api/util/logging/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nv_ingest_api/util/logging/configuration.py,sha256=05KR3LOS-PCqU-Io__iiKG_Ds730eKxciklFfNeId3w,3126
+nv_ingest_api/util/logging/sanitize.py,sha256=-dIbmvLTevrTRd18QKUQQMV4hBk6pStWP_7_VtDDctg,2584
 nv_ingest_api/util/message_brokers/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest_api/util/message_brokers/simple_message_broker/__init__.py,sha256=WaQ3CWIpIKWEivT5kL-bkmzcSQKLGFNFHdXHUJjqZFs,325
 nv_ingest_api/util/message_brokers/simple_message_broker/broker.py,sha256=PekxaxVcAa9k1wgUtozlr04SW3sAeqYJE-wdVBZf9eo,17264
@@ -161,8 +162,8 @@ nv_ingest_api/util/string_processing/configuration.py,sha256=2HS08msccuPCT0fn_jf
 nv_ingest_api/util/string_processing/yaml.py,sha256=6SW2O6wbXRhGbhETMbtXjYCZn53HeCNOP6a96AaxlHs,1454
 nv_ingest_api/util/system/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nv_ingest_api/util/system/hardware_info.py,sha256=1UFM8XE6M3pgQcpbVsCsqDQ7Dj-zzptL-XRE-DEu9UA,27213
-nv_ingest_api-2025.8.19.dev20250819.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-nv_ingest_api-2025.8.19.dev20250819.dist-info/METADATA,sha256=xe42ZKv6dW8r2y2z5OU01FIQ0PRZF4MiK591993tbGk,13947
-nv_ingest_api-2025.8.19.dev20250819.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nv_ingest_api-2025.8.19.dev20250819.dist-info/top_level.txt,sha256=abjYMlTJGoG5tOdfIB-IWvLyKclw6HLaRSc8MxX4X6I,14
-nv_ingest_api-2025.8.19.dev20250819.dist-info/RECORD,,
+nv_ingest_api-2025.8.21.dev20250821.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+nv_ingest_api-2025.8.21.dev20250821.dist-info/METADATA,sha256=GO4G7PAWoOa3E0NQJKagm6Ir3ph9VmiSHl4MwpsnlYw,13947
+nv_ingest_api-2025.8.21.dev20250821.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nv_ingest_api-2025.8.21.dev20250821.dist-info/top_level.txt,sha256=abjYMlTJGoG5tOdfIB-IWvLyKclw6HLaRSc8MxX4X6I,14
+nv_ingest_api-2025.8.21.dev20250821.dist-info/RECORD,,