nv-ingest-api 2025.4.24.dev20250424__py3-none-any.whl → 2025.4.25.dev20250425__py3-none-any.whl

nv_ingest_api/internal/transform/embed_text.py

@@ -453,13 +453,16 @@ def transform_create_text_embeddings_internal(
 
     logger.debug("Generating text embeddings for supported content types: TEXT, STRUCTURED, IMAGE.")
 
+    def _content_type_getter(row):
+        return row["content_metadata"]["type"]
+
     # Process each supported content type.
     for content_type, content_getter in pandas_content_extractor.items():
         if not content_getter:
             logger.debug(f"Skipping unsupported content type: {content_type}")
             continue
 
-        content_mask = df_transform_ledger["document_type"] == content_type.value
+        content_mask = df_transform_ledger["metadata"].apply(_content_type_getter) == content_type.value
         if not content_mask.any():
             continue
 

nv_ingest_api/util/converters/type_mappings.py

@@ -7,7 +7,7 @@ from nv_ingest_api.internal.enums.common import ContentTypeEnum
 DOC_TO_CONTENT_MAP = {
     DocumentTypeEnum.BMP: ContentTypeEnum.IMAGE,
     DocumentTypeEnum.DOCX: ContentTypeEnum.STRUCTURED,
-    DocumentTypeEnum.HTML: ContentTypeEnum.STRUCTURED,
+    DocumentTypeEnum.HTML: ContentTypeEnum.TEXT,
     DocumentTypeEnum.JPEG: ContentTypeEnum.IMAGE,
     DocumentTypeEnum.MP3: ContentTypeEnum.AUDIO,
     DocumentTypeEnum.PDF: ContentTypeEnum.STRUCTURED,

{nv_ingest_api-2025.4.24.dev20250424.dist-info → nv_ingest_api-2025.4.25.dev20250425.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest-api
-Version: 2025.4.24.dev20250424
+Version: 2025.4.25.dev20250425
 Summary: Python module with core document ingestion functions.
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

{nv_ingest_api-2025.4.24.dev20250424.dist-info → nv_ingest_api-2025.4.25.dev20250425.dist-info}/RECORD

@@ -96,7 +96,7 @@ nv_ingest_api/internal/store/embed_text_upload.py,sha256=maxb4FPsBvWgvlrjAPEBlRZ
 nv_ingest_api/internal/store/image_upload.py,sha256=J5EHNng7Z5I6M4f3UcbniKQB29Scr3Qe05wsBpaVXds,9653
 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=MACFgVHUxK3aVlEmymF7F4pT_aKoCcOKxrmmHalk1f0,15622
+nv_ingest_api/internal/transform/embed_text.py,sha256=W-RKM9D9yQy8WAKiiqpkHESpLsL70SacvHwU9EdMeqs,15728
 nv_ingest_api/internal/transform/split_text.py,sha256=y6NYRkCEVpVsDu-AqrKx2D6JPp1vwxclw9obNZNJIIs,6561
 nv_ingest_api/util/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nv_ingest_api/util/control_message/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -107,7 +107,7 @@ nv_ingest_api/util/converters/containers.py,sha256=RV1ooujhq6dVujAzC0MIdZvpOZyQq
 nv_ingest_api/util/converters/datetools.py,sha256=9tskk4BkdLOMLI9ejXvRmri-otE1_Ast3oyX3HQoJZc,2579
 nv_ingest_api/util/converters/dftools.py,sha256=FjHjazIeiUd1LdFwWuummJmraqZe1a90YrWzSjZKzB4,3284
 nv_ingest_api/util/converters/formats.py,sha256=L11FtormO2SeHSebbwsGE_uuCv6Jk0D3VvVW2avU0vI,2258
-nv_ingest_api/util/converters/type_mappings.py,sha256=VFVK5IXfnXJjG4ijDTzKSEZQvJ7xK6iO8snsLgG2vv8,1108
+nv_ingest_api/util/converters/type_mappings.py,sha256=5TVXRyU6BlQvFOdqknEuQw3ss4PXeCvSUynJnjvgQpA,1102
 nv_ingest_api/util/detectors/__init__.py,sha256=HIHfzSig66GT0Uk8qsGBm_f13fKYcPtItBicRUWOOVA,183
 nv_ingest_api/util/detectors/language.py,sha256=TvzcESYY0bn0U4aLY6GjB4VaCWA6XrXxAGZbVzHTMuE,965
 nv_ingest_api/util/exception_handlers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -145,8 +145,8 @@ nv_ingest_api/util/service_clients/redis/redis_client.py,sha256=Xa9eeI3kfDBDlLsG
 nv_ingest_api/util/service_clients/rest/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nv_ingest_api/util/service_clients/rest/rest_client.py,sha256=tjh9BYPscMQ1-IubbI7O7MWwvga5ZYb7OJu65JtPgGw,21732
 nv_ingest_api/util/string_processing/__init__.py,sha256=mkwHthyS-IILcLcL1tJYeF6mpqX3pxEw5aUzDGjTSeU,1411
-nv_ingest_api-2025.4.24.dev20250424.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-nv_ingest_api-2025.4.24.dev20250424.dist-info/METADATA,sha256=MG9QcsfeRIGsKly5b9yFcwRO7VZf21vQwUfKgDQVBoc,13889
-nv_ingest_api-2025.4.24.dev20250424.dist-info/WHEEL,sha256=SmOxYU7pzNKBqASvQJ7DjX3XGUF92lrGhMb3R6_iiqI,91
-nv_ingest_api-2025.4.24.dev20250424.dist-info/top_level.txt,sha256=abjYMlTJGoG5tOdfIB-IWvLyKclw6HLaRSc8MxX4X6I,14
-nv_ingest_api-2025.4.24.dev20250424.dist-info/RECORD,,
+nv_ingest_api-2025.4.25.dev20250425.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+nv_ingest_api-2025.4.25.dev20250425.dist-info/METADATA,sha256=d314O-lCnJ7Hc1ZoIGT9AeQBDvAlYhFhUEVGFU0EzLQ,13889
+nv_ingest_api-2025.4.25.dev20250425.dist-info/WHEEL,sha256=SmOxYU7pzNKBqASvQJ7DjX3XGUF92lrGhMb3R6_iiqI,91
+nv_ingest_api-2025.4.25.dev20250425.dist-info/top_level.txt,sha256=abjYMlTJGoG5tOdfIB-IWvLyKclw6HLaRSc8MxX4X6I,14
+nv_ingest_api-2025.4.25.dev20250425.dist-info/RECORD,,