nv-ingest-api 2025.8.18.dev20250818__tar.gz → 2025.8.20.dev20250820__tar.gz

{nv_ingest_api-2025.8.18.dev20250818/src/nv_ingest_api.egg-info → nv_ingest_api-2025.8.20.dev20250820}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest-api
-Version: 2025.8.18.dev20250818
+Version: 2025.8.20.dev20250820
 Summary: Python module with core document ingestion functions.
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

{nv_ingest_api-2025.8.18.dev20250818 → nv_ingest_api-2025.8.20.dev20250820}/src/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-2025.8.20.dev20250820/src/nv_ingest_api/internal/schemas/transform/transform_image_caption_schema.py

@@ -0,0 +1,33 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES.
+# All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+
+from pydantic import ConfigDict, BaseModel, model_validator, field_validator
+
+
+class ImageCaptionExtractionSchema(BaseModel):
+    api_key: str = ""
+    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-2025.8.18.dev20250818 → nv_ingest_api-2025.8.20.dev20250820}/src/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="")
     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-2025.8.18.dev20250818 → nv_ingest_api-2025.8.20.dev20250820}/src/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-2025.8.18.dev20250818 → nv_ingest_api-2025.8.20.dev20250820/src/nv_ingest_api.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest-api
-Version: 2025.8.18.dev20250818
+Version: 2025.8.20.dev20250820
 Summary: Python module with core document ingestion functions.
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

nv_ingest_api-2025.8.18.dev20250818/src/nv_ingest_api/internal/schemas/transform/transform_image_caption_schema.py

@@ -1,15 +0,0 @@
-# SPDX-FileCopyrightText: Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES.
-# All rights reserved.
-# SPDX-License-Identifier: Apache-2.0
-
-
-from pydantic import ConfigDict, BaseModel
-
-
-class ImageCaptionExtractionSchema(BaseModel):
-    api_key: str = "api_key"
-    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")