nucliadb 6.6.1.post4700__py3-none-any.whl → 6.6.1.post4708__py3-none-any.whl

nucliadb/ingest/orm/resource.py

@@ -146,6 +146,15 @@ class Resource:
         if basic_in_payload.HasField("metadata") and basic_in_payload.metadata.useful:
             current_basic.metadata.status = basic_in_payload.metadata.status
 
+    async def title_marked_for_reset(self) -> bool:
+        basic = await self.get_basic()
+        return basic is not None and basic.reset_title
+
+    async def unmark_title_for_reset(self):
+        basic = await self.get_basic()
+        if basic:
+            basic.reset_title = False
+
     @processor_observer.wrap({"type": "set_basic"})
     async def set_basic(
         self,
@@ -620,10 +629,20 @@ class Resource:
         assert self.basic is not None
         if not link_extracted_data.title:
             return
-        if not (self.basic.title.startswith("http") or self.basic.title == ""):
+        if not (
+            self.basic.title.startswith("http")
+            or self.basic.title == ""
+            or self.basic.title == self.uuid
+            or await self.title_marked_for_reset()
+        ):
             return
+        logger.info(
+            "Updating resource title from link extracted data",
+            extra={"kbid": self.kb.kbid, "field": link_extracted_data.field, "rid": self.uuid},
+        )
         title = link_extracted_data.title
         await self.update_resource_title(title)
+        await self.unmark_title_for_reset()
         self.modified = True
 
     async def update_resource_title(self, computed_title: str) -> None:
@@ -690,6 +709,9 @@ class Resource:
         if current_title in filenames:
             # If the title is equal to any of the file filenames, we should update it
             return True
+        if await self.title_marked_for_reset():
+            # If the title is marked for reset, we should update it
+            return True
         return False
 
     async def maybe_update_resource_title_from_file_extracted_data(self, message: BrokerMessage):
@@ -708,6 +730,7 @@ class Resource:
                 extra={"kbid": self.kb.kbid, "field": fid.full(), "new_title": fed.title},
             )
             await self.update_resource_title(fed.title)
+            await self.unmark_title_for_reset()
             # Break after the first file with a title is found
             break
 

nucliadb/writer/api/v1/field.py

@@ -20,7 +20,7 @@
 from inspect import iscoroutinefunction
 from typing import TYPE_CHECKING, Annotated, Callable, Optional, Type, Union
 
-from fastapi import HTTPException, Response
+from fastapi import HTTPException, Query, Response
 from fastapi_versioning import version
 from starlette.requests import Request
 
@@ -541,6 +541,10 @@ async def reprocess_file_field(
     field_id: FieldIdString,
     x_nucliadb_user: Annotated[str, X_NUCLIADB_USER] = "",
     x_file_password: Annotated[Optional[str], X_FILE_PASSWORD] = None,
+    reset_title: bool = Query(
+        default=False,
+        description="Reset the title of the resource so that the file or link computed titles are set after processing.",
+    ),
 ) -> ResourceUpdated:
     await maybe_back_pressure(kbid, resource_uuid=rid)
 
@@ -590,6 +594,8 @@ async def reprocess_file_field(
     writer.kbid = kbid
     writer.uuid = rid
     writer.source = BrokerMessage.MessageSource.WRITER
+    if reset_title:
+        writer.basic.reset_title = True
     writer.basic.metadata.useful = True
     writer.basic.metadata.status = Metadata.Status.PENDING
     writer.field_statuses.append(

nucliadb/writer/api/v1/resource.py

@@ -105,7 +105,6 @@ async def create_resource(
             status_code=422,
             detail="Cannot hide a resource: the KB does not have hidden resources enabled",
         )
-
     await maybe_back_pressure(kbid)
 
     partitioning = get_partitioning()
@@ -394,9 +393,15 @@ async def reprocess_resource_rslug_prefix(
     kbid: str,
     rslug: str,
     x_nucliadb_user: Annotated[str, X_NUCLIADB_USER] = "",
+    reset_title: bool = Query(
+        default=False,
+        description="Reset the title of the resource so that the file or link computed titles are set after processing.",
+    ),
 ):
     rid = await get_rid_from_slug_or_raise_error(kbid, rslug)
-    return await _reprocess_resource(request, kbid, rid, x_nucliadb_user=x_nucliadb_user)
+    return await _reprocess_resource(
+        request, kbid, rid, x_nucliadb_user=x_nucliadb_user, reset_title=reset_title
+    )
 
 
 @api.post(
@@ -413,8 +418,14 @@ async def reprocess_resource_rid_prefix(
     kbid: str,
     rid: str,
     x_nucliadb_user: Annotated[str, X_NUCLIADB_USER] = "",
+    reset_title: bool = Query(
+        default=False,
+        description="Reset the title of the resource so that the file or link computed titles are set after processing.",
+    ),
 ):
-    return await _reprocess_resource(request, kbid, rid, x_nucliadb_user=x_nucliadb_user)
+    return await _reprocess_resource(
+        request, kbid, rid, x_nucliadb_user=x_nucliadb_user, reset_title=reset_title
+    )
 
 
 async def _reprocess_resource(
@@ -422,6 +433,10 @@ async def _reprocess_resource(
     kbid: str,
     rid: str,
     x_nucliadb_user: str,
+    reset_title: bool = Query(
+        default=False,
+        description="Reset the title of the resource so that the file or link computed titles are set after processing.",
+    ),
 ):
     await validate_rid_exists_or_raise_error(kbid, rid)
     await maybe_back_pressure(kbid, resource_uuid=rid)
@@ -464,6 +479,8 @@ async def _reprocess_resource(
     writer.kbid = kbid
     writer.uuid = rid
     writer.source = BrokerMessage.MessageSource.WRITER
+    if reset_title:
+        writer.basic.reset_title = True
     writer.basic.metadata.useful = True
     writer.basic.metadata.status = Metadata.Status.PENDING
     await transaction.commit(writer, partition, wait=False)

{nucliadb-6.6.1.post4700.dist-info → nucliadb-6.6.1.post4708.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nucliadb
-Version: 6.6.1.post4700
+Version: 6.6.1.post4708
 Summary: NucliaDB
 Author-email: Nuclia <nucliadb@nuclia.com>
 License-Expression: AGPL-3.0-or-later
@@ -19,11 +19,11 @@ 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]>=6.6.1.post4700
-Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.6.1.post4700
-Requires-Dist: nucliadb-protos>=6.6.1.post4700
-Requires-Dist: nucliadb-models>=6.6.1.post4700
-Requires-Dist: nidx-protos>=6.6.1.post4700
+Requires-Dist: nucliadb-telemetry[all]>=6.6.1.post4708
+Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.6.1.post4708
+Requires-Dist: nucliadb-protos>=6.6.1.post4708
+Requires-Dist: nucliadb-models>=6.6.1.post4708
+Requires-Dist: nidx-protos>=6.6.1.post4708
 Requires-Dist: nucliadb-admin-assets>=1.0.0.post1224
 Requires-Dist: nuclia-models>=0.43.0
 Requires-Dist: uvicorn[standard]

{nucliadb-6.6.1.post4700.dist-info → nucliadb-6.6.1.post4708.dist-info}/RECORD

@@ -163,7 +163,7 @@ nucliadb/ingest/orm/exceptions.py,sha256=k4Esv4NtL4TrGTcsQpwrSfDhPQpiYcRbB1SpYmB
 nucliadb/ingest/orm/index_message.py,sha256=DWMTHJoVamUbK8opKl5csDvxfgz7c2j7phG1Ut4yIxk,15724
 nucliadb/ingest/orm/knowledgebox.py,sha256=_rkeTMIXMhR64gbYtZpFHoUHghV2DTJ2lUBqZsoqC_4,23898
 nucliadb/ingest/orm/metrics.py,sha256=OiuggTh-n3kZHA2G73NEUdIlh8c3yFrbusI88DK-Mko,1273
-nucliadb/ingest/orm/resource.py,sha256=yB0HWC3jc_1b-zXu-3FJCKOdAPPSb1aRBHpbZhsvyQk,37749
+nucliadb/ingest/orm/resource.py,sha256=G3efCiOnaIW0pdPtB28PkKgaOEE-7A5PnugBu3f9-OQ,38589
 nucliadb/ingest/orm/utils.py,sha256=fCQRuyecgqhaY7mcBG93oaXMkzkKb9BFjOcy4-ZiSNw,2693
 nucliadb/ingest/orm/processor/__init__.py,sha256=Aqd9wCNTvggkMkCY3WvoI8spdr94Jnqk-0iq9XpLs18,922
 nucliadb/ingest/orm/processor/auditing.py,sha256=TeYhXGJRyQ7ROytbb2u8R0fIh_FYi3HgTu3S1ribY3U,4623
@@ -352,10 +352,10 @@ nucliadb/writer/api/constants.py,sha256=qWEDjFUycrEZnSJyLnNK4PQNodU2oVmkO4NycaEZ
 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=v0sU55TtRSqDzwkDgcwv2uSaqKCuQTtGcMpYoHQYBQA,8192
-nucliadb/writer/api/v1/field.py,sha256=OicvLF1bnkJj1ixALFLuhvFX6NCMFpORROcFcS9nKpk,18505
+nucliadb/writer/api/v1/field.py,sha256=rF_x6fVczZKa4xBrHChUVau-70uFJT7srQpOLuqG2ic,18755
 nucliadb/writer/api/v1/knowledgebox.py,sha256=PHEYDFa-sN5JrI8-EiVVg5FDOsRuCLT43kyAB4xt-xA,9530
 nucliadb/writer/api/v1/learning_config.py,sha256=CKBjqcbewkfPwGUPLDWzZSpro6XkmCaVppe5Qtpu5Go,3117
-nucliadb/writer/api/v1/resource.py,sha256=IaKHwP4M4Pm3xXj_xcnQCnTzKtXj_xj-r7YOHdH-89I,19750
+nucliadb/writer/api/v1/resource.py,sha256=gr4VtYOtnIAL5zsDQsBDzeOOWWh6wN7PUh1XgUsQ9N4,20436
 nucliadb/writer/api/v1/router.py,sha256=RjuoWLpZer6Kl2BW_wznpNo6XL3BOpdTGqXZCn3QrrQ,1034
 nucliadb/writer/api/v1/services.py,sha256=3AUjk-SmvqJx76v7y89DZx6oyasojPliGYeniRQjpcU,13337
 nucliadb/writer/api/v1/slug.py,sha256=xlVBDBpRi9bNulpBHZwhyftVvulfE0zFm1XZIWl-AKY,2389
@@ -376,8 +376,8 @@ 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-6.6.1.post4700.dist-info/METADATA,sha256=A_aHp-CWMHPO0E19OROR6udFDABya3QCGlC8lihrLts,4158
-nucliadb-6.6.1.post4700.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nucliadb-6.6.1.post4700.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
-nucliadb-6.6.1.post4700.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
-nucliadb-6.6.1.post4700.dist-info/RECORD,,
+nucliadb-6.6.1.post4708.dist-info/METADATA,sha256=5ExGb5nsHOVIevztP91xdMRXg7XyF1I1uAJE8NB0OhU,4158
+nucliadb-6.6.1.post4708.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nucliadb-6.6.1.post4708.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
+nucliadb-6.6.1.post4708.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
+nucliadb-6.6.1.post4708.dist-info/RECORD,,