nv-ingest 2025.7.19.dev20250719__py3-none-any.whl → 2025.7.21.dev20250721__py3-none-any.whl

nv_ingest/framework/orchestration/ray/util/pipeline/stage_builders.py

@@ -215,7 +215,6 @@ def add_pdf_extractor_stage(pipeline, default_cpu_count, stage_name="pdf_extract
         min_replicas=0,
         max_replicas=max_replicas,
     )
-
     return stage_name
 
 
@@ -242,7 +241,7 @@ def add_table_extractor_stage(pipeline, default_cpu_count, stage_name="table_ext
         stage_actor=TableExtractorStage,
         config=table_extractor_config,
         min_replicas=0,
-        max_replicas=2,
+        max_replicas=_get_max_replicas(default_cpu_count, percentage_of_cpu=0.20),
     )
 
     return stage_name
@@ -271,7 +270,7 @@ def add_chart_extractor_stage(pipeline, default_cpu_count, stage_name="chart_ext
         stage_actor=ChartExtractorStage,
         config=chart_extractor_config,
         min_replicas=0,
-        max_replicas=2,
+        max_replicas=_get_max_replicas(default_cpu_count, percentage_of_cpu=0.20),
     )
 
     return stage_name
@@ -295,7 +294,7 @@ def add_infographic_extractor_stage(pipeline, default_cpu_count, stage_name="inf
         stage_actor=InfographicExtractorStage,
         config=infographic_content_extractor_config,
         min_replicas=0,
-        max_replicas=1,
+        max_replicas=2,
     )
 
     return stage_name
@@ -317,7 +316,7 @@ def add_image_extractor_stage(pipeline, default_cpu_count, stage_name="image_ext
         stage_actor=ImageExtractorStage,
         config=image_extractor_config,
         min_replicas=0,
-        max_replicas=1,
+        max_replicas=2,
     )
 
     return stage_name
@@ -383,11 +382,7 @@ def add_audio_extractor_stage(pipeline, default_cpu_count, stage_name="audio_ext
     )
 
     pipeline.add_stage(
-        name=stage_name,
-        stage_actor=AudioExtractorStage,
-        config=audio_extractor_config,
-        min_replicas=0,
-        max_replicas=1,
+        name=stage_name, stage_actor=AudioExtractorStage, config=audio_extractor_config, min_replicas=0, max_replicas=2
     )
 
     return stage_name
@@ -400,7 +395,7 @@ def add_html_extractor_stage(pipeline, default_cpu_count, stage_name="html_extra
         stage_actor=HtmlExtractorStage,
         config=HtmlExtractorSchema(),
         min_replicas=0,
-        max_replicas=1,
+        max_replicas=2,
     )
 
     return stage_name
@@ -630,3 +625,10 @@ def add_source_stage(pipeline, default_cpu_count, source_name="pipeline_source")
         start_simple_message_broker(source_config.broker_client.model_dump())
 
     return source_name
+
+
+def _get_max_replicas(default_cpu_count=None, percentage_of_cpu=0.14):
+    if default_cpu_count is None:
+        default_cpu_count = _system_resource_probe.get_cpu_count()
+
+    return int(max(1, (default_cpu_count * percentage_of_cpu)))

{nv_ingest-2025.7.19.dev20250719.dist-info → nv_ingest-2025.7.21.dev20250721.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest
-Version: 2025.7.19.dev20250719
+Version: 2025.7.21.dev20250721
 Summary: Python module for multimodal document ingestion
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

{nv_ingest-2025.7.19.dev20250719.dist-info → nv_ingest-2025.7.21.dev20250721.dist-info}/RECORD

@@ -67,7 +67,7 @@ nv_ingest/framework/orchestration/ray/util/pipeline/__init__.py,sha256=wQSlVx3T1
 nv_ingest/framework/orchestration/ray/util/pipeline/pid_controller.py,sha256=flRLS7yc5n6gheykayuL3prC7O-ZhcVY2s9Wc14SGWE,47377
 nv_ingest/framework/orchestration/ray/util/pipeline/pipeline_builders.py,sha256=d2-GS2tqk6JOFdw65CL1AwfjdUbkC_XxUuJH8Dy-aQ0,10456
 nv_ingest/framework/orchestration/ray/util/pipeline/pipeline_runners.py,sha256=dlz83vEFKcvrwsPFP1M0Md1lOYbuOVX2MeyCPq43RGg,14392
-nv_ingest/framework/orchestration/ray/util/pipeline/stage_builders.py,sha256=rlMqLtuaI-VdVlAT_7-9HSIgX6YfsxWsBzKe3fBvYl0,21136
+nv_ingest/framework/orchestration/ray/util/pipeline/stage_builders.py,sha256=bnhnrl6Yp39txWoLCB9os9PEzHs5WVJlbW4aE6_yNFo,21459
 nv_ingest/framework/orchestration/ray/util/pipeline/tools.py,sha256=LQVb8k9jURaxh2Ga44Js_XuYFCbeN4_nLgDmtExovQg,8026
 nv_ingest/framework/orchestration/ray/util/system_tools/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/framework/orchestration/ray/util/system_tools/memory.py,sha256=ICqY0LLB3hFTZk03iX5yffMSKFH2q_aQomtDVzS_mKw,2228
@@ -96,8 +96,8 @@ nv_ingest/framework/util/service/meta/ingest/__init__.py,sha256=wQSlVx3T14ZgQAt-
 nv_ingest/framework/util/service/meta/ingest/ingest_service_meta.py,sha256=QS3uNxWBl5dIcmIpJKNe8_TLcTUuN2vcKyHeAwa-eSo,1589
 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-2025.7.19.dev20250719.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-nv_ingest-2025.7.19.dev20250719.dist-info/METADATA,sha256=mWq6HM-znacym6NTDg4UsUXcxQ8OYKNMlrSQsweU3C0,15142
-nv_ingest-2025.7.19.dev20250719.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nv_ingest-2025.7.19.dev20250719.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
-nv_ingest-2025.7.19.dev20250719.dist-info/RECORD,,
+nv_ingest-2025.7.21.dev20250721.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+nv_ingest-2025.7.21.dev20250721.dist-info/METADATA,sha256=5pQ5u5URCsxzHsroY3_BNcc9ziLzke-HsWH1gOuqT9A,15142
+nv_ingest-2025.7.21.dev20250721.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nv_ingest-2025.7.21.dev20250721.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
+nv_ingest-2025.7.21.dev20250721.dist-info/RECORD,,