nucliadb 6.3.1.post3570__py3-none-any.whl → 6.3.1.post3574__py3-none-any.whl

nucliadb/export_import/tasks.py

@@ -17,7 +17,6 @@
 # You should have received a copy of the GNU Affero General Public License
 # along with this program. If not, see <http://www.gnu.org/licenses/>.
 #
-from nucliadb.common.context import ApplicationContext
 from nucliadb.export_import.exporter import export_kb_to_blob_storage
 from nucliadb.export_import.importer import import_kb_from_blob_storage
 from nucliadb.export_import.models import NatsTaskMessage
@@ -60,14 +59,13 @@ def get_exports_consumer() -> NatsTaskConsumer[NatsTaskMessage]:
     )
 
 
-async def get_exports_producer(context: ApplicationContext) -> NatsTaskProducer[NatsTaskMessage]:
+def get_exports_producer() -> NatsTaskProducer[NatsTaskMessage]:
     producer = create_producer(
         name="exports_producer",
         stream=ExportsNatsConfig.stream,
         producer_subject=ExportsNatsConfig.consumer.subject,
         msg_type=NatsTaskMessage,
     )
-    await producer.initialize(context)
     return producer
 
 
@@ -82,12 +80,11 @@ def get_imports_consumer() -> NatsTaskConsumer[NatsTaskMessage]:
     )
 
 
-async def get_imports_producer(context: ApplicationContext) -> NatsTaskProducer[NatsTaskMessage]:
+def get_imports_producer() -> NatsTaskProducer[NatsTaskMessage]:
     producer = create_producer(
         name="imports_producer",
         stream=ImportsNatsConfig.stream,
         producer_subject=ImportsNatsConfig.consumer.subject,
         msg_type=NatsTaskMessage,
     )
-    await producer.initialize(context)
     return producer

nucliadb/search/search/chat/query.py

@@ -272,6 +272,7 @@ async def get_relations_results_from_entities(
     timeout: Optional[float] = None,
     only_with_metadata: bool = False,
     only_agentic_relations: bool = False,
+    only_entity_to_entity: bool = False,
     deleted_entities: set[str] = set(),
 ) -> Relations:
     request = SearchRequest()
@@ -295,7 +296,11 @@ async def get_relations_results_from_entities(
     )
     relations_results: list[RelationSearchResponse] = [result.relation for result in results]
     return await merge_relations_results(
-        relations_results, request.relation_subgraph, only_with_metadata, only_agentic_relations
+        relations_results,
+        request.relation_subgraph,
+        only_with_metadata,
+        only_agentic_relations,
+        only_entity_to_entity,
     )
 
 

nucliadb/search/search/graph_strategy.py

@@ -369,8 +369,10 @@ async def get_graph_results(
                     kbid=kbid,
                     entities=entities_to_explore,
                     timeout=5.0,
-                    only_with_metadata=True,
+                    only_with_metadata=not graph_strategy.relation_text_as_paragraphs,
                     only_agentic_relations=graph_strategy.agentic_graph_only,
+                    # We only want entity to entity relations (skip resource/labels/collaborators/etc.)
+                    only_entity_to_entity=True,
                     deleted_entities=explored_entities,
                 )
             except Exception as e:
@@ -683,6 +685,7 @@ def build_text_blocks_from_relations(
     triplets: dict[tuple[str, str, str], tuple[float, Relations, Optional[ParagraphId]]] = defaultdict(
         lambda: (0.0, Relations(entities={}), None)
     )
+    paragraph_count = 0
     for ent, subgraph in relations.entities.items():
         for rel, score in zip(subgraph.related_to, scores[ent]):
             key = (
@@ -702,6 +705,14 @@ def build_text_blocks_from_relations(
             # we keep the first one, but we lose the other ones
             if p_id is None and rel.metadata and rel.metadata.paragraph_id:
                 p_id = ParagraphId.from_string(rel.metadata.paragraph_id)
+            else:
+                # No paragraph ID set, fake it so we can hydrate the resource
+                p_id = ParagraphId(
+                    field_id=FieldId(rel.resource_id, "a", "usermetadata"),
+                    paragraph_start=paragraph_count,
+                    paragraph_end=paragraph_count + 1,
+                )
+                paragraph_count += 1
             existing_relations.entities[ent].related_to.append(rel)
             # XXX: Here we use the max even though all relations with same triplet should have same score
             triplets[key] = (max(existing_score, score), existing_relations, p_id)

nucliadb/search/search/merge.py

@@ -35,6 +35,7 @@ from nucliadb.search.search.fetch import (
 )
 from nucliadb_models.common import FieldTypeName
 from nucliadb_models.labels import translate_system_to_alias_label
+from nucliadb_models.metadata import RelationType
 from nucliadb_models.resource import ExtractedDataTypeName
 from nucliadb_models.search import (
     DirectionalRelation,
@@ -445,6 +446,7 @@ async def merge_relations_results(
     query: EntitiesSubgraphRequest,
     only_with_metadata: bool = False,
     only_agentic: bool = False,
+    only_entity_to_entity: bool = False,
 ) -> Relations:
     loop = asyncio.get_event_loop()
     return await loop.run_in_executor(
@@ -454,6 +456,7 @@ async def merge_relations_results(
         query,
         only_with_metadata,
         only_agentic,
+        only_entity_to_entity,
     )
 
 
@@ -462,6 +465,7 @@ def _merge_relations_results(
     query: EntitiesSubgraphRequest,
     only_with_metadata: bool,
     only_agentic: bool,
+    only_entity_to_entity: bool,
 ) -> Relations:
     """
     Merge relation search responses into a single Relations object while applying filters.
@@ -490,33 +494,41 @@ def _merge_relations_results(
             # If only_with_metadata is True, we check that metadata for the relation is not None
             # If only_agentic is True, we check that metadata for the relation is not None and that it has a data_augmentation_task_id
             # TODO: This is suboptimal, we should be able to filter this in the query to the index,
-            if (not only_with_metadata or metadata) and (
-                not only_agentic or (metadata and metadata.data_augmentation_task_id)
-            ):
-                if origin.value in relations.entities:
-                    relations.entities[origin.value].related_to.append(
-                        DirectionalRelation(
-                            entity=destination.value,
-                            entity_type=relation_node_type_to_entity_type(destination.ntype),
-                            entity_subtype=destination.subtype,
-                            relation=relation_type,
-                            relation_label=relation_label,
-                            direction=RelationDirection.OUT,
-                            metadata=from_proto.relation_metadata(metadata) if metadata else None,
-                        )
+            if only_with_metadata and not metadata:
+                continue
+
+            if only_agentic and (not metadata or not metadata.data_augmentation_task_id):
+                continue
+
+            if only_entity_to_entity and relation_type != RelationType.ENTITY:
+                continue
+
+            if origin.value in relations.entities:
+                relations.entities[origin.value].related_to.append(
+                    DirectionalRelation(
+                        entity=destination.value,
+                        entity_type=relation_node_type_to_entity_type(destination.ntype),
+                        entity_subtype=destination.subtype,
+                        relation=relation_type,
+                        relation_label=relation_label,
+                        direction=RelationDirection.OUT,
+                        metadata=from_proto.relation_metadata(metadata) if metadata else None,
+                        resource_id=relation.resource_id,
                     )
-                elif destination.value in relations.entities:
-                    relations.entities[destination.value].related_to.append(
-                        DirectionalRelation(
-                            entity=origin.value,
-                            entity_type=relation_node_type_to_entity_type(origin.ntype),
-                            entity_subtype=origin.subtype,
-                            relation=relation_type,
-                            relation_label=relation_label,
-                            direction=RelationDirection.IN,
-                            metadata=from_proto.relation_metadata(metadata) if metadata else None,
-                        )
+                )
+            elif destination.value in relations.entities:
+                relations.entities[destination.value].related_to.append(
+                    DirectionalRelation(
+                        entity=origin.value,
+                        entity_type=relation_node_type_to_entity_type(origin.ntype),
+                        entity_subtype=origin.subtype,
+                        relation=relation_type,
+                        relation_label=relation_label,
+                        direction=RelationDirection.IN,
+                        metadata=from_proto.relation_metadata(metadata) if metadata else None,
+                        resource_id=relation.resource_id,
                     )
+                )
 
     return relations
 

nucliadb/tasks/producer.py

@@ -17,13 +17,13 @@
 # You should have received a copy of the GNU Affero General Public License
 # along with this program. If not, see <http://www.gnu.org/licenses/>.
 #
-from typing import Generic, Optional, Type
+from typing import Generic, Type
 
-from nucliadb.common.context import ApplicationContext
 from nucliadb.tasks.logger import logger
 from nucliadb.tasks.models import MsgType
-from nucliadb.tasks.utils import NatsStream, create_nats_stream_if_not_exists
+from nucliadb.tasks.utils import NatsStream
 from nucliadb_telemetry import errors
+from nucliadb_utils.utilities import get_nats_manager
 
 
 class NatsTaskProducer(Generic[MsgType]):
@@ -38,29 +38,17 @@ class NatsTaskProducer(Generic[MsgType]):
         self.stream = stream
         self.producer_subject = producer_subject
         self.msg_type = msg_type
-        self.context: Optional[ApplicationContext] = None
-        self.initialized = False
-
-    async def initialize(self, context: ApplicationContext):
-        self.context = context
-        await create_nats_stream_if_not_exists(
-            self.context,
-            self.stream.name,
-            subjects=self.stream.subjects,
-        )
-        self.initialized = True
+        self.nats_manager = get_nats_manager()
 
     async def send(self, msg: MsgType) -> int:
         """
         Publish message to the producer's nats stream.
         Returns the sequence number of the published message.
         """
-        if not self.initialized:
-            raise RuntimeError("NatsTaskProducer not initialized")
         try:
-            pub_ack = await self.context.nats_manager.js.publish(  # type: ignore
+            pub_ack = await self.nats_manager.js.publish(
                 self.producer_subject,
-                msg.model_dump_json().encode("utf-8"),  # type: ignore
+                msg.model_dump_json().encode("utf-8"),
             )
             logger.info(
                 "Message sent to Nats",

nucliadb/writer/api/v1/export_import.py

@@ -195,7 +195,7 @@ async def start_export_task(context: ApplicationContext, kbid: str, export_id: s
     metadata.task.status = Status.SCHEDULED
     await dm.set_metadata("export", metadata)
     try:
-        producer = await get_exports_producer(context)
+        producer = get_exports_producer()
         msg = NatsTaskMessage(kbid=kbid, id=export_id)
         seqid = await producer.send(msg)
         logger.info(f"Export task produced. seqid={seqid} kbid={kbid} export_id={export_id}")
@@ -212,7 +212,7 @@ async def start_import_task(context: ApplicationContext, kbid: str, import_id: s
     metadata.total = import_size or 0
     await dm.set_metadata("import", metadata)
     try:
-        producer = await get_imports_producer(context)
+        producer = get_imports_producer()
         msg = NatsTaskMessage(kbid=kbid, id=import_id)
         seqid = await producer.send(msg)
         logger.info(f"Import task produced. seqid={seqid} kbid={kbid} import_id={import_id}")

{nucliadb-6.3.1.post3570.dist-info → nucliadb-6.3.1.post3574.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: nucliadb
-Version: 6.3.1.post3570
+Version: 6.3.1.post3574
 Summary: NucliaDB
 Author-email: Nuclia <nucliadb@nuclia.com>
 License: AGPL
@@ -20,11 +20,11 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3 :: Only
 Requires-Python: <4,>=3.9
 Description-Content-Type: text/markdown
-Requires-Dist: nucliadb-telemetry[all]>=6.3.1.post3570
-Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.3.1.post3570
-Requires-Dist: nucliadb-protos>=6.3.1.post3570
-Requires-Dist: nucliadb-models>=6.3.1.post3570
-Requires-Dist: nidx-protos>=6.3.1.post3570
+Requires-Dist: nucliadb-telemetry[all]>=6.3.1.post3574
+Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.3.1.post3574
+Requires-Dist: nucliadb-protos>=6.3.1.post3574
+Requires-Dist: nucliadb-models>=6.3.1.post3574
+Requires-Dist: nidx-protos>=6.3.1.post3574
 Requires-Dist: nucliadb-admin-assets>=1.0.0.post1224
 Requires-Dist: nuclia-models>=0.24.2
 Requires-Dist: uvicorn

{nucliadb-6.3.1.post3570.dist-info → nucliadb-6.3.1.post3574.dist-info}/RECORD

@@ -108,7 +108,7 @@ nucliadb/export_import/exceptions.py,sha256=Dw8WqfG4r6MPJc5TPfbjMvCgXXWTcTOecGHR
 nucliadb/export_import/exporter.py,sha256=k2QVx1EjqFlDYiggriWiEJzwtMXzHbldsqWdpGQM3_U,7074
 nucliadb/export_import/importer.py,sha256=v5cq9Nn8c2zrY_K_00mydR52f8mdFxR7tLdtNLQ0qvk,4229
 nucliadb/export_import/models.py,sha256=dbjScNkiMRv4X3Ktudy1JRliD25bfoDTy3JmEZgQSCc,2121
-nucliadb/export_import/tasks.py,sha256=4JX3bygyLCLSuGxMCStYyoclh_CL8rPxrVVWuGqvcmM,3146
+nucliadb/export_import/tasks.py,sha256=DWbdqY97ffoyfipelGXz3Jqz1iam6JCjQSh367Fc3NA,2947
 nucliadb/export_import/utils.py,sha256=iAQAjYuNx0dhM2b5-1A0NEs8tSRsznuT-izysUrTwS0,19986
 nucliadb/ingest/__init__.py,sha256=fsw3C38VP50km3R-nHL775LNGPpJ4JxqXJ2Ib1f5SqE,1011
 nucliadb/ingest/app.py,sha256=rX1KE5vsAzG9hlArBk8WE2SOlvdYylcb-jNkMQNPJdQ,7407
@@ -224,10 +224,10 @@ nucliadb/search/search/fetch.py,sha256=XJHIFnZmXM_8Kb37lb4lg1GYG7cZ1plT-qAIb_Qzi
 nucliadb/search/search/filters.py,sha256=1MkHlJjAQqoRCj7e5cEzK2HvBxGLE17I_omsjiklbtw,6476
 nucliadb/search/search/find.py,sha256=jQZOqu8VeX8k3ELV8bLK4TwUUjGrvmubouxvO1IvJV0,10236
 nucliadb/search/search/find_merge.py,sha256=3FnzKFEnVemg6FO_6zveulbAU7klvsiPEBvLrpBBMg8,17450
-nucliadb/search/search/graph_strategy.py,sha256=ahwcUTQZ0Ll-rnS285DO9PmRyiM-1p4BM3UvmOYVwhM,31750
+nucliadb/search/search/graph_strategy.py,sha256=gisL2GpbSIa_SucyOwEt7TWdqURyAQqxvD_-PkXQct8,32339
 nucliadb/search/search/hydrator.py,sha256=-R37gCrGxkyaiHQalnTWHNG_FCx11Zucd7qA1vQCxuw,6985
 nucliadb/search/search/ingestion_agents.py,sha256=NeJr4EEX-bvFFMGvXOOwLv8uU7NuQ-ntJnnrhnKfMzY,3174
-nucliadb/search/search/merge.py,sha256=i_PTBFRqC5iTTziOMEltxLIlmokIou5hjjgR4BnoLBE,22635
+nucliadb/search/search/merge.py,sha256=aUn6f5XnwWzUFhVC6uBqHE8NKdlfgw_xcTo57rS23U8,22950
 nucliadb/search/search/metrics.py,sha256=GGGtXHLhK79_ESV277xkBVjcaMURXHCxYG0EdGamUd8,2886
 nucliadb/search/search/paragraphs.py,sha256=pNAEiYqJGGUVcEf7xf-PFMVqz0PX4Qb-WNG-_zPGN2o,7799
 nucliadb/search/search/pgcatalog.py,sha256=V1NYLEUSXHpWmgcPIo1HS2riK_HDXSi-uykJjSoOOrE,9033
@@ -243,7 +243,7 @@ nucliadb/search/search/chat/ask.py,sha256=olZT08JVo3ZGDsDXkjvI2JTlqQln_o91HJzv0T
 nucliadb/search/search/chat/exceptions.py,sha256=Siy4GXW2L7oPhIR86H3WHBhE9lkV4A4YaAszuGGUf54,1356
 nucliadb/search/search/chat/images.py,sha256=PA8VWxT5_HUGfW1ULhKTK46UBsVyINtWWqEM1ulzX1E,3095
 nucliadb/search/search/chat/prompt.py,sha256=Jnja-Ss7skgnnDY8BymVfdeYsFPnIQFL8tEvcRXTKUE,47356
-nucliadb/search/search/chat/query.py,sha256=2QhVzvX12zLHOpVZ5MlBflqAauyCBl6dojhRGdm_6qU,16388
+nucliadb/search/search/chat/query.py,sha256=0IoeW-JNaRBe2d9C3bXNfkYpzmsN_IIg3U4Vqb8eOEk,16485
 nucliadb/search/search/query_parser/__init__.py,sha256=cp15ZcFnHvpcu_5-aK2A4uUyvuZVV_MJn4bIXMa20ks,835
 nucliadb/search/search/query_parser/catalog.py,sha256=PtH5nb6UTzH8l7Lmdd1RgLVFsn9CN5M5-JkVq9YeR4k,7116
 nucliadb/search/search/query_parser/exceptions.py,sha256=szAOXUZ27oNY-OSa9t2hQ5HHkQQC0EX1FZz_LluJHJE,1224
@@ -271,7 +271,7 @@ nucliadb/tasks/__init__.py,sha256=oFJ3A8HD7w11mBu-IixYE_KxA7juMGlYQb7YD_y6WPM,97
 nucliadb/tasks/consumer.py,sha256=xc0Ql3N1Iq52dJ3t4YYGJFj1NCQAly0J5W_brfLa_F8,6894
 nucliadb/tasks/logger.py,sha256=C7keOEO_mjLVp5VbqAZ2QXfqVB2Hot7NgBlUP_SDSMw,924
 nucliadb/tasks/models.py,sha256=qrZKi5DNDQ07waMsp5L4_Fi7WRs57YiO-kmXlrBzEAA,1168
-nucliadb/tasks/producer.py,sha256=JRGnATkALyr_iLHq0OAjzVbfxZ_SOUa6sx-smU5p6SQ,3136
+nucliadb/tasks/producer.py,sha256=UnpJAzhj_GElsCoO5G6T4m6MshsgOaqR2tVzJmEta64,2625
 nucliadb/tasks/retries.py,sha256=Zv-3Hys-SKayG9VQ7_7EIflkegE5j-xPGrf-nwaxsfY,5075
 nucliadb/tasks/utils.py,sha256=tV1AbWdFc3qfIULX44Veqj41FCD1B6XYjG6brULBeiw,1459
 nucliadb/tests/__init__.py,sha256=cp15ZcFnHvpcu_5-aK2A4uUyvuZVV_MJn4bIXMa20ks,835
@@ -322,7 +322,7 @@ nucliadb/writer/api/__init__.py,sha256=cp15ZcFnHvpcu_5-aK2A4uUyvuZVV_MJn4bIXMa20
 nucliadb/writer/api/constants.py,sha256=qWEDjFUycrEZnSJyLnNK4PQNodU2oVmkO4NycaEZtio,1738
 nucliadb/writer/api/utils.py,sha256=wIQHlU8RQiIGVLI72suvyVIKlCU44Unh0Ae0IiN6Qwo,1313
 nucliadb/writer/api/v1/__init__.py,sha256=akI9A_jloNLb0dU4T5zjfdyvmSAiDeIdjAlzNx74FlU,1128
-nucliadb/writer/api/v1/export_import.py,sha256=FxGY_rofj3AH-HQReoMDOYfEpACeWB5bsmxrT03QUtQ,8227
+nucliadb/writer/api/v1/export_import.py,sha256=elf-EQY5DD3mhw8kWb9tQpDcbrF9sY6VFYqxQOjuVP0,8201
 nucliadb/writer/api/v1/field.py,sha256=OsWOYA0WQ6onE5Rkl20QIEdtrSi7Jgnu62fUt90Ziy8,17503
 nucliadb/writer/api/v1/knowledgebox.py,sha256=MLeIuym4jPrJgfy1NTcN9CpUGwuBiqDHMcx0hY9DR7g,9530
 nucliadb/writer/api/v1/learning_config.py,sha256=CKBjqcbewkfPwGUPLDWzZSpro6XkmCaVppe5Qtpu5Go,3117
@@ -347,8 +347,8 @@ nucliadb/writer/tus/local.py,sha256=7jYa_w9b-N90jWgN2sQKkNcomqn6JMVBOVeDOVYJHto,
 nucliadb/writer/tus/s3.py,sha256=vF0NkFTXiXhXq3bCVXXVV-ED38ECVoUeeYViP8uMqcU,8357
 nucliadb/writer/tus/storage.py,sha256=ToqwjoYnjI4oIcwzkhha_MPxi-k4Jk3Lt55zRwaC1SM,2903
 nucliadb/writer/tus/utils.py,sha256=MSdVbRsRSZVdkaum69_0wku7X3p5wlZf4nr6E0GMKbw,2556
-nucliadb-6.3.1.post3570.dist-info/METADATA,sha256=a5FFqexlq9oEflkrfSYCAW8j3sV62OJnY1LCgPtpbJg,4291
-nucliadb-6.3.1.post3570.dist-info/WHEEL,sha256=52BFRY2Up02UkjOa29eZOS2VxUrpPORXg1pkohGGUS8,91
-nucliadb-6.3.1.post3570.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
-nucliadb-6.3.1.post3570.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
-nucliadb-6.3.1.post3570.dist-info/RECORD,,
+nucliadb-6.3.1.post3574.dist-info/METADATA,sha256=JiB_eKqj0pPsq4nf9VvqU9gm6CEEQkYvElBlLSISd7c,4291
+nucliadb-6.3.1.post3574.dist-info/WHEEL,sha256=52BFRY2Up02UkjOa29eZOS2VxUrpPORXg1pkohGGUS8,91
+nucliadb-6.3.1.post3574.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
+nucliadb-6.3.1.post3574.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
+nucliadb-6.3.1.post3574.dist-info/RECORD,,