nv-ingest 2025.7.16.dev20250716__py3-none-any.whl → 2025.7.17.dev20250717__py3-none-any.whl

nv_ingest/api/v1/health.py

@@ -18,7 +18,7 @@ router = APIRouter()
 
 # List of ALL of the HTTP environment variable endpoints that should be checked
 READY_CHECK_ENV_VAR_MAP = {
-    "paddle": "PADDLE_HTTP_ENDPOINT",
+    "ocr": "OCR_HTTP_ENDPOINT",
     "yolox_graphic_elements": "YOLOX_GRAPHIC_ELEMENTS_HTTP_ENDPOINT",
     "yolox_page_elements": "YOLOX_HTTP_ENDPOINT",
     "yolox_table_structure": "YOLOX_TABLE_STRUCTURE_HTTP_ENDPOINT",

nv_ingest/framework/orchestration/ray/examples/pipeline_test_harness.py

@@ -147,8 +147,9 @@ if __name__ == "__main__":
     os.environ["YOLOX_GRAPHIC_ELEMENTS_GRPC_ENDPOINT"] = "127.0.0.1:8004"
     os.environ["YOLOX_GRAPHIC_ELEMENTS_HTTP_ENDPOINT"] = "http://localhost:8003/v1/infer"
     os.environ["YOLOX_GRAPHIC_ELEMENTS_INFER_PROTOCOL"] = "http"
-    os.environ["PADDLE_GRPC_ENDPOINT"] = "localhost:8010"
-    os.environ["PADDLE_INFER_PROTOCOL"] = "grpc"
+    os.environ["OCR_GRPC_ENDPOINT"] = "localhost:8010"
+    os.environ["OCR_INFER_PROTOCOL"] = "grpc"
+    os.environ["OCR_MODEL_NAME"] = "paddle"
     os.environ["NEMORETRIEVER_PARSE_HTTP_ENDPOINT"] = "https://integrate.api.nvidia.com/v1/chat/completions"
     os.environ["VLM_CAPTION_ENDPOINT"] = "https://integrate.api.nvidia.com/v1/chat/completions"
     os.environ["VLM_CAPTION_MODEL_NAME"] = "nvidia/llama-3.1-nemotron-nano-vl-8b-v1"
@@ -172,7 +173,7 @@ if __name__ == "__main__":
     nemoretriever_parse_grpc, nemoretriever_parse_http, nemoretriever_parse_auth, nemoretriever_parse_protocol = (
         get_nim_service("nemoretriever_parse")
     )
-    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+    ocr_grpc, ocr_http, ocr_auth, ocr_protocol = get_nim_service("ocr")
 
     model_name = os.environ.get("NEMORETRIEVER_PARSE_MODEL_NAME", "nvidia/nemoretriever-parse")
     pdf_extractor_config = {
@@ -201,8 +202,8 @@ if __name__ == "__main__":
         "endpoint_config": {
             "yolox_endpoints": (yolox_graphic_elements_grpc, yolox_graphic_elements_http),
             "yolox_infer_protocol": yolox_graphic_elements_protocol,
-            "paddle_endpoints": (paddle_grpc, paddle_http),
-            "paddle_infer_protocol": paddle_protocol,
+            "ocr_endpoints": (ocr_grpc, ocr_http),
+            "ocr_infer_protocol": ocr_protocol,
             "auth_token": yolox_auth,
         }
     }
@@ -210,8 +211,8 @@ if __name__ == "__main__":
         "endpoint_config": {
             "yolox_endpoints": (yolox_table_structure_grpc, yolox_table_structure_http),
             "yolox_infer_protocol": yolox_table_structure_protocol,
-            "paddle_endpoints": (paddle_grpc, paddle_http),
-            "paddle_infer_protocol": paddle_protocol,
+            "ocr_endpoints": (ocr_grpc, ocr_http),
+            "ocr_infer_protocol": ocr_protocol,
             "auth_token": yolox_auth,
         }
     }

nv_ingest/framework/orchestration/ray/stages/sources/message_broker_task_source.py

@@ -269,8 +269,11 @@ class MessageBrokerTaskSourceStage(RayActorSourceStage):
             self._logger.debug("Received message type: %s", type(job))
             if isinstance(job, BaseModel):
                 self._logger.debug("Message is a BaseModel with response_code: %s", job.response_code)
-                if job.response_code != 0:
-                    self._logger.debug("Message response_code != 0, returning None")
+                if job.response_code not in (0, 2):
+                    self._logger.debug("Message received with unhandled response_code, returning None")
+                    return None
+                if job.response_code == 2:
+                    self._logger.debug("Message response_code == 2, returning None")
                     return None
                 job = json.loads(job.response)
             self._logger.debug("Successfully fetched message with job_id: %s", job.get("job_id", "unknown"))

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

@@ -78,8 +78,9 @@ class PipelineCreationSchema(BaseModel):
     otel_exporter_otlp_endpoint: str = os.getenv("OTEL_EXPORTER_OTLP_ENDPOINT", "localhost:4317")
 
     # OCR settings
-    paddle_http_endpoint: str = os.getenv("PADDLE_HTTP_ENDPOINT", "https://ai.api.nvidia.com/v1/cv/baidu/paddleocr")
-    paddle_infer_protocol: str = os.getenv("PADDLE_INFER_PROTOCOL", "http")
+    ocr_http_endpoint: str = os.getenv("OCR_HTTP_ENDPOINT", "https://ai.api.nvidia.com/v1/cv/baidu/paddleocr")
+    ocr_infer_protocol: str = os.getenv("OCR_INFER_PROTOCOL", "http")
+    ocr_model_name: str = os.getenv("OCR_MODEL_NAME", "paddle")
 
     # Task queue settings
     REDIS_INGEST_TASK_QUEUE: str = "ingest_task_queue"

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

@@ -223,15 +223,15 @@ def add_table_extractor_stage(pipeline, default_cpu_count, stage_name="table_ext
     yolox_table_structure_grpc, yolox_table_structure_http, yolox_auth, yolox_table_structure_protocol = (
         get_nim_service("yolox_table_structure")
     )
-    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+    ocr_grpc, ocr_http, ocr_auth, ocr_protocol = get_nim_service("ocr")
 
     table_extractor_config = TableExtractorSchema(
         **{
             "endpoint_config": {
                 "yolox_endpoints": (yolox_table_structure_grpc, yolox_table_structure_http),
                 "yolox_infer_protocol": yolox_table_structure_protocol,
-                "paddle_endpoints": (paddle_grpc, paddle_http),
-                "paddle_infer_protocol": paddle_protocol,
+                "ocr_endpoints": (ocr_grpc, ocr_http),
+                "ocr_infer_protocol": ocr_protocol,
                 "auth_token": yolox_auth,
             }
         }
@@ -252,15 +252,15 @@ def add_chart_extractor_stage(pipeline, default_cpu_count, stage_name="chart_ext
     yolox_graphic_elements_grpc, yolox_graphic_elements_http, yolox_auth, yolox_graphic_elements_protocol = (
         get_nim_service("yolox_graphic_elements")
     )
-    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+    ocr_grpc, ocr_http, ocr_auth, ocr_protocol = get_nim_service("ocr")
 
     chart_extractor_config = ChartExtractorSchema(
         **{
             "endpoint_config": {
                 "yolox_endpoints": (yolox_graphic_elements_grpc, yolox_graphic_elements_http),
                 "yolox_infer_protocol": yolox_graphic_elements_protocol,
-                "paddle_endpoints": (paddle_grpc, paddle_http),
-                "paddle_infer_protocol": paddle_protocol,
+                "ocr_endpoints": (ocr_grpc, ocr_http),
+                "ocr_infer_protocol": ocr_protocol,
                 "auth_token": yolox_auth,
             }
         }
@@ -278,14 +278,14 @@ def add_chart_extractor_stage(pipeline, default_cpu_count, stage_name="chart_ext
 
 
 def add_infographic_extractor_stage(pipeline, default_cpu_count, stage_name="infographic_extractor"):
-    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+    ocr_grpc, ocr_http, ocr_auth, ocr_protocol = get_nim_service("ocr")
 
     infographic_content_extractor_config = InfographicExtractorSchema(
         **{
             "endpoint_config": {
-                "paddle_endpoints": (paddle_grpc, paddle_http),
-                "paddle_infer_protocol": paddle_protocol,
-                "auth_token": paddle_auth,
+                "ocr_endpoints": (ocr_grpc, ocr_http),
+                "ocr_infer_protocol": ocr_protocol,
+                "auth_token": ocr_auth,
             }
         }
     )

{nv_ingest-2025.7.16.dev20250716.dist-info → nv_ingest-2025.7.17.dev20250717.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nv-ingest
-Version: 2025.7.16.dev20250716
+Version: 2025.7.17.dev20250717
 Summary: Python module for multimodal document ingestion
 Author-email: Jeremy Dyer <jdyer@nvidia.com>
 License:                                  Apache License

{nv_ingest-2025.7.16.dev20250716.dist-info → nv_ingest-2025.7.17.dev20250717.dist-info}/RECORD

@@ -3,7 +3,7 @@ nv_ingest/version.py,sha256=MG7DxlzpnoJI56vqxwzs9WeMAEI3uPhfDiNLs6GN6wI,986
 nv_ingest/api/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/api/main.py,sha256=XE-p4lJp1E7CCDOB8ENtYFrf63Dtq2bzQiGxpRfL2LA,1603
 nv_ingest/api/v1/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest/api/v1/health.py,sha256=zqu-isMRjh4NveS4XWh5FaAZGPIlBVxpCOg3Uu8nUHQ,4746
+nv_ingest/api/v1/health.py,sha256=pV-RoVq5y0iBPp0qZoLzd1xKpd0JiHAi0UMyMj99LqU,4740
 nv_ingest/api/v1/ingest.py,sha256=LWk3LN4lBd3uO8h30EN42g3LHCVcO00avVd5ohVK7NI,19392
 nv_ingest/api/v1/metrics.py,sha256=ZGVRApYLnzc2f2C7wRgGd7deqiXan-jxfA-33a16clY,981
 nv_ingest/framework/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
@@ -14,7 +14,7 @@ nv_ingest/framework/orchestration/ray/edges/async_queue_edge.py,sha256=PQliU_kyG
 nv_ingest/framework/orchestration/ray/edges/ray_queue_edge.py,sha256=VFii2yxJuikimOxie3edKq5JN06g78AF8bdHSHVX8p8,2677
 nv_ingest/framework/orchestration/ray/edges/threaded_queue_edge.py,sha256=N6NH4KgZJ60e_JkGRcSmfQtX37qtX4TMcavOR-n3heE,2549
 nv_ingest/framework/orchestration/ray/examples/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest/framework/orchestration/ray/examples/pipeline_test_harness.py,sha256=qUNvWiNBUEEzuCySY3usWmHQz9qMgTGVZuKmLWqTsi4,16412
+nv_ingest/framework/orchestration/ray/examples/pipeline_test_harness.py,sha256=DufjmNm-05uTkq_Mz0QQB6fHw_Rl9eX3PRtnH4sntGs,16405
 nv_ingest/framework/orchestration/ray/examples/task_source_harness.py,sha256=Yt7uxThg7s8WuMiaHLKC8r1XAG7QixegfkT-juE5oNw,1953
 nv_ingest/framework/orchestration/ray/examples/task_source_sink_harness.py,sha256=XkvsoIzH5ftXvAZ4ox7mxbx7ESVx6D8Xupcwbqgd52w,3277
 nv_ingest/framework/orchestration/ray/primitives/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
@@ -48,7 +48,7 @@ nv_ingest/framework/orchestration/ray/stages/sinks/__init__.py,sha256=wQSlVx3T14
 nv_ingest/framework/orchestration/ray/stages/sinks/default_drain.py,sha256=0SQHJlFuXlP16YRWduX1fMKgjhUd7UhDAWQ8XZh4_0I,1471
 nv_ingest/framework/orchestration/ray/stages/sinks/message_broker_task_sink.py,sha256=enylryvcPmzirpOjCahqYJbNSLsNvv1KpMnOzGqNZQQ,11509
 nv_ingest/framework/orchestration/ray/stages/sources/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
-nv_ingest/framework/orchestration/ray/stages/sources/message_broker_task_source.py,sha256=KzRil999sHGK4jV-EBU8LUuPp_e3W-Vc_feFEAvG2-E,20995
+nv_ingest/framework/orchestration/ray/stages/sources/message_broker_task_source.py,sha256=b9ndnQBB1paR0iRe3NdzQ7BZ2S65LG2jbtjXvvDc_s4,21183
 nv_ingest/framework/orchestration/ray/stages/storage/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 nv_ingest/framework/orchestration/ray/stages/storage/image_storage.py,sha256=6NkwQzseAnaj0Ptpr3oKvab2EnJdMwTjI2p4dS_HzsI,3901
 nv_ingest/framework/orchestration/ray/stages/storage/store_embeddings.py,sha256=SMLHQElZkKldnjy0_VHIKS65DBAAtOhwhdoaFe1yb9I,3337
@@ -66,8 +66,8 @@ nv_ingest/framework/orchestration/ray/util/__init__.py,sha256=wQSlVx3T14ZgQAt-EP
 nv_ingest/framework/orchestration/ray/util/pipeline/__init__.py,sha256=wQSlVx3T14ZgQAt-EPzEczQusXVW0W8yynnUaFFGE3s,143
 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=5I7N-nmGXaYqyPEtPZbFhgvog2b2c3eagWc69naMc9s,14340
-nv_ingest/framework/orchestration/ray/util/pipeline/stage_builders.py,sha256=-IoaRYO7tDA9JXJ7J_j-8pVcU4dYWXrKDM3vo392XZA,21229
+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/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-2025.7.16.dev20250716.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-nv_ingest-2025.7.16.dev20250716.dist-info/METADATA,sha256=PmbdHa6yAUfDi83pFlWjbUkrAr4Lf8hMHbVhNq5DBiY,15142
-nv_ingest-2025.7.16.dev20250716.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nv_ingest-2025.7.16.dev20250716.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
-nv_ingest-2025.7.16.dev20250716.dist-info/RECORD,,
+nv_ingest-2025.7.17.dev20250717.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+nv_ingest-2025.7.17.dev20250717.dist-info/METADATA,sha256=Kn17uqN2NU6FWJzEU46y6-M_zZ8v-l3vFNjnP2BdF-U,15142
+nv_ingest-2025.7.17.dev20250717.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nv_ingest-2025.7.17.dev20250717.dist-info/top_level.txt,sha256=sjb0ajIsgn3YgftSjZHlYO0HjYAIIhNuXG_AmywCvaU,10
+nv_ingest-2025.7.17.dev20250717.dist-info/RECORD,,