nv-ingest 2025.12.27.dev20251227__py3-none-any.whl → 2026.1.14.dev20260114__py3-none-any.whl

nv_ingest/api/v2/ingest.py

@@ -122,11 +122,16 @@ def get_pdf_split_page_count(client_override: Optional[int] = None) -> int:
         )
         return DEFAULT_PDF_SPLIT_PAGE_COUNT
 
-    if parsed <= 0:
-        logger.warning("PDF_SPLIT_PAGE_COUNT must be >= 1; received %s. Using 1.", parsed)
-        return 1
-
-    return parsed
+    clamped = max(MIN_PAGES, min(parsed, MAX_PAGES))
+    if clamped != parsed:
+        logger.warning(
+            "Env PDF_SPLIT_PAGE_COUNT=%s clamped to %s (min=%s, max=%s)",
+            parsed,
+            clamped,
+            MIN_PAGES,
+            MAX_PAGES,
+        )
+    return clamped
 
 
 def split_pdf_to_chunks(pdf_content: bytes, pages_per_chunk: int) -> List[Dict[str, Any]]:
@@ -955,7 +960,7 @@ async def submit_job_v2(
                         "subjob_order": subjob_ids,
                     }
                 )
-        elif document_types and payloads and document_types[0].lower() in ["mp4", "mov", "avi", "mp3", "wav"]:
+        elif document_types and payloads and document_types[0].lower() in ["mp4", "mov", "avi", "mp3", "wav", "mkv"]:
             document_type = document_types[0]
             upload_path = f"./{Path(original_source_id).name}"
             # dump the payload to a file, just came from client

nv_ingest/pipeline/default_libmode_pipeline_impl.py

@@ -68,7 +68,7 @@ stages:
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
       nemotron_parse_config:
@@ -81,7 +81,7 @@ stages:
         nemotron_parse_model_name: $NEMOTRON_PARSE_MODEL_NAME|"nvidia/nemotron-parse"
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
     replicas:
@@ -124,14 +124,14 @@ stages:
       docx_extraction_config:
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
       pdfium_config:
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
@@ -152,14 +152,14 @@ stages:
       pptx_extraction_config:
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
       pdfium_config:
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
@@ -180,7 +180,7 @@ stages:
       image_extraction_config:
         yolox_endpoints: [
           $YOLOX_GRPC_ENDPOINT|"",
-          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v2"
+          $YOLOX_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-page-elements-v3"
         ]
         yolox_infer_protocol: $YOLOX_INFER_PROTOCOL|http
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
@@ -243,7 +243,7 @@ stages:
           $OCR_GRPC_ENDPOINT|"",
           $OCR_HTTP_ENDPOINT|"https://ai.api.nvidia.com/v1/cv/nvidia/nemoretriever-ocr-v1"
         ]
-        ocr_infer_protocol: $PADDLE_INFER_PROTOCOL|"http"
+        ocr_infer_protocol: $OCR_INFER_PROTOCOL|"http"
         auth_token: $NGC_API_KEY|$NVIDIA_API_KEY
     replicas:
       min_replicas: 0
@@ -332,7 +332,7 @@ stages:
     actor: "nv_ingest.framework.orchestration.ray.stages.transforms.image_caption:ImageCaptionTransformStage"
     config:
       api_key: $NGC_API_KEY|$NVIDIA_API_KEY
-      endpoint_url: $VLM_CAPTION_ENDPOINT|"http://vlm:8000/v1/chat/completions"
+      endpoint_url: $VLM_CAPTION_ENDPOINT|"https://integrate.api.nvidia.com/v1/chat/completions"
       model_name: $VLM_CAPTION_MODEL_NAME|"nvidia/nemotron-nano-12b-v2-vl"
       prompt: $VLM_CAPTION_PROMPT|"Caption the content of this image:"
       system_prompt: $VLM_CAPTION_SYSTEM_PROMPT|"/no_think"

{nv_ingest-2025.12.27.dev20251227.dist-info → nv_ingest-2026.1.14.dev20260114.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest
-Version: 2025.12.27.dev20251227
+Version: 2026.1.14.dev20260114
 Summary: Python module for multimodal document ingestion
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License
@@ -227,7 +227,7 @@ Requires-Dist: httpx>=0.28.1
 Requires-Dist: isodate>=0.7.2
 Requires-Dist: langdetect>=1.0.9
 Requires-Dist: minio>=7.2.12
-Requires-Dist: librosa>=0.10.2
+Requires-Dist: librosa==0.10.2
 Requires-Dist: opentelemetry-api>=1.27.0
 Requires-Dist: opentelemetry-exporter-otlp>=1.27.0
 Requires-Dist: opentelemetry-sdk>=1.27.0
@@ -254,7 +254,6 @@ Requires-Dist: transformers>=4.47.0
 Requires-Dist: tqdm>=4.67.1
 Requires-Dist: uvicorn
 Requires-Dist: pip
-Requires-Dist: llama-index-embeddings-nvidia
 Requires-Dist: opencv-python
 Requires-Dist: pymilvus>=2.5.10
 Requires-Dist: pymilvus[bulk_writer,model]

{nv_ingest-2025.12.27.dev20251227.dist-info → nv_ingest-2026.1.14.dev20260114.dist-info}/RECORD

@@ -9,7 +9,7 @@ nv_ingest/api/v1/ingest.py,sha256=LWk3LN4lBd3uO8h30EN42g3LHCVcO00avVd5ohVK7NI,19
 nv_ingest/api/v1/metrics.py,sha256=ZGVRApYLnzc2f2C7wRgGd7deqiXan-jxfA-33a16clY,981
 nv_ingest/api/v2/README.md,sha256=VhpdjEmCyr3qIOhwqISFx9C5WezJFcxYc-NB9S98HMg,7562
 nv_ingest/api/v2/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest/api/v2/ingest.py,sha256=vjjb2xOOtlTVoTMc4rNdUI6yKYdEeR-umA_pwP_Rt64,53103
+nv_ingest/api/v2/ingest.py,sha256=DjPqw1SwQqwqBraQ7n1WajejnGeHbqGmXpzzyfRohH0,53256
 nv_ingest/framework/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/framework/orchestration/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/framework/orchestration/execution/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
@@ -112,15 +112,15 @@ nv_ingest/framework/util/service/meta/ingest/ingest_service_meta.py,sha256=QS3uN
 nv_ingest/framework/util/telemetry/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/framework/util/telemetry/global_stats.py,sha256=nq65pEEdiwjAfGiqsxG1CeQMC96O3CfQxsZuGFCY-ds,4554
 nv_ingest/pipeline/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest/pipeline/default_libmode_pipeline_impl.py,sha256=YYASfM68qNhGL5PcK0Fv72qmRZfE2TtY3cq2Oz-L478,16267
+nv_ingest/pipeline/default_libmode_pipeline_impl.py,sha256=PSFWMwrKCEBlo6e8nCdL5mg84SawoExwrUbc4nOHu5M,16281
 nv_ingest/pipeline/default_pipeline_impl.py,sha256=6SykgH_LJ8uuE2jrWGIT7OkJP6EjPyB8Ju6LMDu5IK0,16800
 nv_ingest/pipeline/ingest_pipeline.py,sha256=wHAJhqAM2s8nbY-8itVogmSU-yVN4PZONGWcKnhzgfg,17794
 nv_ingest/pipeline/pipeline_schema.py,sha256=rLZZz2It2o2hVNWrZUJU8CarrqRei1fho3ZEMkkoBcg,17940
 nv_ingest/pipeline/config/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/pipeline/config/loaders.py,sha256=75Yr9WYO7j7ghvKTnYLfZXQZEH3J3VEZo5J4TunC_Us,7590
 nv_ingest/pipeline/config/replica_resolver.py,sha256=dEwqMXNttfw0QeisTGGkp24785jqzVCDAEFyQIffeGc,9369
-nv_ingest-2025.12.27.dev20251227.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-nv_ingest-2025.12.27.dev20251227.dist-info/METADATA,sha256=y2hPzoDUslthurcySM2XsuE0fiNG9hShORDJxrH0vm8,15163
-nv_ingest-2025.12.27.dev20251227.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nv_ingest-2025.12.27.dev20251227.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
-nv_ingest-2025.12.27.dev20251227.dist-info/RECORD,,
+nv_ingest-2026.1.14.dev20260114.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+nv_ingest-2026.1.14.dev20260114.dist-info/METADATA,sha256=UZHSNIHvkqd7dlc7kZaaMN5L5vK2ZKiJAyvrIubfGpk,15117
+nv_ingest-2026.1.14.dev20260114.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nv_ingest-2026.1.14.dev20260114.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
+nv_ingest-2026.1.14.dev20260114.dist-info/RECORD,,