nucliadb 6.7.0.post4800__py3-none-any.whl → 6.7.0.post4805__py3-none-any.whl

nucliadb/ingest/fields/base.py

@@ -332,37 +332,54 @@ class Field(Generic[PbType]):
             self.question_answers = actual_payload
 
     async def set_extracted_text(self, payload: ExtractedTextWrapper) -> None:
+        actual_payload: Optional[ExtractedText] = None
         if self.type in SUBFIELDFIELDS:
+            # Try to get the previously extracted text protobuf if it exists so we can merge it with the new splits
+            # coming from the processing payload.
             try:
-                actual_payload: Optional[ExtractedText] = await self.get_extracted_text(force=True)
+                actual_payload = await self.get_extracted_text(force=True)
             except KeyError:
-                actual_payload = None
-        else:
-            actual_payload = None
-        sf = self.get_storage_field(FieldTypes.FIELD_TEXT)
+                # No previous extracted text found
+                pass
 
+        sf = self.get_storage_field(FieldTypes.FIELD_TEXT)
         if actual_payload is None:
-            # Its first extracted text
+            # No previous extracted text, this is the first time we set it so we can simply upload it to storage
             if payload.HasField("file"):
+                # Normalize the storage key if the payload is a reference to a file in storage.
+                # This is typically the case when the text is too large and we store it in a
+                # cloud file. Normalization is needed to ensure that the hybrid-onprem deployment stores
+                # the file in the correct bucket of its storage.
                 await self.storage.normalize_binary(payload.file, sf)
             else:
+                # Directly upload the ExtractedText protobuf to storage
                 await self.storage.upload_pb(sf, payload.body)
                 self.extracted_text = payload.body
         else:
             if payload.HasField("file"):
+                # The extracted text coming from processing has a reference to another storage key.
+                # Download it and copy it to its ExtractedText.body field. This is typically for cases
+                # when the text is too large.
                 raw_payload = await self.storage.downloadbytescf(payload.file)
                 pb = ExtractedText()
                 pb.ParseFromString(raw_payload.read())
                 raw_payload.flush()
                 payload.body.CopyFrom(pb)
-            # We know its payload.body
+
+            # Update or set the extracted text text for each split coming from the processing payload
             for key, value in payload.body.split_text.items():
                 actual_payload.split_text[key] = value
+
+            # Apply any split deletions that may come in the processing payload
             for key in payload.body.deleted_splits:
                 if key in actual_payload.split_text:
                     del actual_payload.split_text[key]
+
+            # Finally, handle the main text body (for the cases where the text is not split)
             if payload.body.text != "":
                 actual_payload.text = payload.body.text
+
+            # Upload the updated ExtractedText to storage
             await self.storage.upload_pb(sf, actual_payload)
             self.extracted_text = actual_payload
 

nucliadb/ingest/fields/conversation.py

@@ -26,8 +26,9 @@ from nucliadb_protos.resources_pb2 import Conversation as PBConversation
 from nucliadb_utils.storages.storage import StorageField
 
 PAGE_SIZE = 200
-KB_RESOURCE_FIELD = "/kbs/{kbid}/r/{uuid}/f/{type}/{field}/{page}"
-KB_RESOURCE_FIELD_METADATA = "/kbs/{kbid}/r/{uuid}/f/{type}/{field}"
+
+CONVERSATION_PAGE_VALUE = "/kbs/{kbid}/r/{uuid}/f/{type}/{field}/{page}"
+CONVERSATION_METADATA = "/kbs/{kbid}/r/{uuid}/f/{type}/{field}"
 
 
 class PageNotFound(Exception):
@@ -54,56 +55,65 @@ class Conversation(Field[PBConversation]):
         self.metadata = None
 
     async def set_value(self, payload: PBConversation):
-        last_page: Optional[PBConversation] = None
         metadata = await self.get_metadata()
         metadata.extract_strategy = payload.extract_strategy
         metadata.split_strategy = payload.split_strategy
 
+        # Get the last page if it exists
+        last_page: Optional[PBConversation] = None
         if self._created is False and metadata.pages > 0:
             try:
                 last_page = await self.db_get_value(page=metadata.pages)
             except PageNotFound:
                 pass
+        if last_page is None:
+            last_page = PBConversation()
+            metadata.pages += 1
 
-        # Make sure message attachment files are on our region
+        # Make sure message attachment files are on our region. This is needed
+        # to support the hybrid-onprem deployment as the attachments must be stored
+        # at the storage services of the client's premises.
         for message in payload.messages:
             new_message_files = []
-            for count, file in enumerate(message.content.attachments):
+            for idx, file in enumerate(message.content.attachments):
                 if self.storage.needs_move(file, self.kbid):
                     if message.ident == "":
                         message_ident = uuid.uuid4().hex
                     else:
                         message_ident = message.ident
-                    sf: StorageField = self.storage.conversation_field(
-                        self.kbid, self.uuid, self.id, message_ident, count
+                    sf: StorageField = self.storage.conversation_field_attachment(
+                        self.kbid, self.uuid, self.id, message_ident, attachment_index=idx
                     )
                     cf: CloudFile = await self.storage.normalize_binary(file, sf)
                     new_message_files.append(cf)
                 else:
                     new_message_files.append(file)
 
-            # Can be cleaned a list of PB
+            # Replace the attachments in the message with the new ones
             message.content.ClearField("attachments")
             for message_file in new_message_files:
                 message.content.attachments.append(message_file)
 
-        if last_page is None:
-            last_page = PBConversation()
-            metadata.pages += 1
-
-        # Merge on last page
+        # Increment the metadata total with the number of messages
         messages = list(payload.messages)
         metadata.total += len(messages)
+
+        # Store the messages in pages of PAGE_SIZE messages
         while len(messages) > 0:
-            count = metadata.size - len(last_page.messages)
-            last_page.messages.extend(messages[:count])
+            # Fit the messages in the last page
+            available_space = metadata.size - len(last_page.messages)
+            last_page.messages.extend(messages[:available_space])
+
+            # Save the last page
             await self.db_set_value(last_page, metadata.pages)
 
-            messages = messages[count:]
+            # If there are still messages, create a new page
+            messages = messages[available_space:]
             if len(messages) > 0:
                 metadata.pages += 1
                 last_page = PBConversation()
 
+        # Finally, set the metadata
         await self.db_set_metadata(metadata)
 
     async def get_value(self, page: Optional[int] = None) -> Optional[PBConversation]:
@@ -139,7 +149,7 @@ class Conversation(Field[PBConversation]):
     async def get_metadata(self) -> FieldConversation:
         if self.metadata is None:
             payload = await self.resource.txn.get(
-                KB_RESOURCE_FIELD_METADATA.format(
+                CONVERSATION_METADATA.format(
                     kbid=self.kbid, uuid=self.uuid, type=self.type, field=self.id
                 )
             )
@@ -158,7 +168,7 @@ class Conversation(Field[PBConversation]):
             raise ValueError(f"Conversation pages start at index 1")
 
         if self.value.get(page) is None:
-            field_key = KB_RESOURCE_FIELD.format(
+            field_key = CONVERSATION_PAGE_VALUE.format(
                 kbid=self.kbid,
                 uuid=self.uuid,
                 type=self.type,
@@ -174,7 +184,7 @@ class Conversation(Field[PBConversation]):
         return self.value[page]
 
     async def db_set_value(self, payload: PBConversation, page: int = 0):
-        field_key = KB_RESOURCE_FIELD.format(
+        field_key = CONVERSATION_PAGE_VALUE.format(
             kbid=self.kbid, uuid=self.uuid, type=self.type, field=self.id, page=page
         )
         await self.resource.txn.set(
@@ -186,9 +196,7 @@ class Conversation(Field[PBConversation]):
 
     async def db_set_metadata(self, payload: FieldConversation):
         await self.resource.txn.set(
-            KB_RESOURCE_FIELD_METADATA.format(
-                kbid=self.kbid, uuid=self.uuid, type=self.type, field=self.id
-            ),
+            CONVERSATION_METADATA.format(kbid=self.kbid, uuid=self.uuid, type=self.type, field=self.id),
             payload.SerializeToString(),
         )
         self.metadata = payload

nucliadb/reader/api/v1/download.py

@@ -174,7 +174,9 @@ async def download_field_conversation_rslug_prefix(
     message_id: str,
     file_num: int,
 ) -> Response:
-    return await _download_field_conversation(request, kbid, field_id, message_id, file_num, rslug=rslug)
+    return await _download_field_conversation_attachment(
+        request, kbid, field_id, message_id, file_num, rslug=rslug
+    )
 
 
 @api.get(
@@ -185,7 +187,7 @@ async def download_field_conversation_rslug_prefix(
 )
 @requires_one([NucliaDBRoles.READER])
 @version(1)
-async def download_field_conversation_rid_prefix(
+async def download_field_conversation_attachment_rid_prefix(
     request: Request,
     kbid: str,
     rid: str,
@@ -193,10 +195,12 @@ async def download_field_conversation_rid_prefix(
     message_id: str,
     file_num: int,
 ) -> Response:
-    return await _download_field_conversation(request, kbid, field_id, message_id, file_num, rid=rid)
+    return await _download_field_conversation_attachment(
+        request, kbid, field_id, message_id, file_num, rid=rid
+    )
 
 
-async def _download_field_conversation(
+async def _download_field_conversation_attachment(
     request: Request,
     kbid: str,
     field_id: str,
@@ -209,7 +213,9 @@ async def _download_field_conversation(
 
     storage = await get_storage(service_name=SERVICE_NAME)
 
-    sf = storage.conversation_field(kbid, rid, field_id, message_id, file_num)
+    sf = storage.conversation_field_attachment(
+        kbid, rid, field_id, message_id, attachment_index=file_num
+    )
 
     return await download_api(sf, request.headers)
 

nucliadb/writer/resource/field.py

@@ -466,9 +466,9 @@ async def parse_conversation_field(
             ]
         )
 
-        for count, file in enumerate(message.content.attachments):
-            sf_conv_field: StorageField = storage.conversation_field(
-                kbid, uuid, field=key, ident=message.ident, count=count
+        for idx, file in enumerate(message.content.attachments):
+            sf_conv_field: StorageField = storage.conversation_field_attachment(
+                kbid, uuid, field=key, ident=message.ident, attachment_index=idx
             )
             cf_conv_field = await storage.upload_b64file_to_cloudfile(
                 sf_conv_field,

{nucliadb-6.7.0.post4800.dist-info → nucliadb-6.7.0.post4805.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nucliadb
-Version: 6.7.0.post4800
+Version: 6.7.0.post4805
 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]>=6.7.0.post4800
-Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.7.0.post4800
-Requires-Dist: nucliadb-protos>=6.7.0.post4800
-Requires-Dist: nucliadb-models>=6.7.0.post4800
-Requires-Dist: nidx-protos>=6.7.0.post4800
+Requires-Dist: nucliadb-telemetry[all]>=6.7.0.post4805
+Requires-Dist: nucliadb-utils[cache,fastapi,storages]>=6.7.0.post4805
+Requires-Dist: nucliadb-protos>=6.7.0.post4805
+Requires-Dist: nucliadb-models>=6.7.0.post4805
+Requires-Dist: nidx-protos>=6.7.0.post4805
 Requires-Dist: nucliadb-admin-assets>=1.0.0.post1224
 Requires-Dist: nuclia-models>=0.46.0
 Requires-Dist: uvicorn[standard]

{nucliadb-6.7.0.post4800.dist-info → nucliadb-6.7.0.post4805.dist-info}/RECORD

@@ -148,8 +148,8 @@ nucliadb/ingest/consumer/service.py,sha256=8AD41mMN7EUeUtk4ZNy14zfvxzwmVjIX6Mwe0
 nucliadb/ingest/consumer/shard_creator.py,sha256=w0smEu01FU_2cjZnsfBRNqT_Ntho11X17zTMST-vKbc,4359
 nucliadb/ingest/consumer/utils.py,sha256=jpX8D4lKzuPCpArQLZeX_Zczq3pfen_zAf8sPJfOEZU,2642
 nucliadb/ingest/fields/__init__.py,sha256=cp15ZcFnHvpcu_5-aK2A4uUyvuZVV_MJn4bIXMa20ks,835
-nucliadb/ingest/fields/base.py,sha256=vYLGB-8SRYnFIHSZBSm20iXZDAzlwuBiJQC8s3BQv6w,22958
-nucliadb/ingest/fields/conversation.py,sha256=PGOr7l6cvg-l1HMr-5UIJq53R6W93F5uew0fOrSufPI,7140
+nucliadb/ingest/fields/base.py,sha256=O9rMKZmSdZmcxtuO6qUXjeoAkmUR5A-xga49Jyn6caI,24192
+nucliadb/ingest/fields/conversation.py,sha256=W68Y7riDXbwavpE4jKslLNAjdqCVaaR2a-BisZYayeo,7675
 nucliadb/ingest/fields/exceptions.py,sha256=sZBk21BSrXFdOdo1qUdCAyD-9YMYakSLdn4_WdIPCIQ,1217
 nucliadb/ingest/fields/file.py,sha256=1v4jLg3balUua2VmSV8hHkAwPFShTUCOzufZvIUQcQw,4740
 nucliadb/ingest/fields/generic.py,sha256=elgtqv15aJUq3zY7X_g0bli_2BpcwPArVvzhe54Y4Ig,1547
@@ -199,7 +199,7 @@ nucliadb/reader/run.py,sha256=AR-iCnON3YVXgI5-KEgg99G4KAPN1BKXDg7nr4dgoDA,1447
 nucliadb/reader/api/__init__.py,sha256=c-UD29C0FVzQDGEvslebDCKtvnyEcAbiDd-3Q_QgGN4,872
 nucliadb/reader/api/models.py,sha256=UHhOPmh8xcHhDjFm8_-8t66yEggXdxRBoY0xV-hI9to,2076
 nucliadb/reader/api/v1/__init__.py,sha256=ieP8lsCCwG66Jupv8II5MSTj6nh3eCtLcF4utH9JOcU,1102
-nucliadb/reader/api/v1/download.py,sha256=rGv1c5CjrJheDgGwAnNWy76A_4V2shqqHYvwmKGFlpk,10758
+nucliadb/reader/api/v1/download.py,sha256=A4ntoyPbLQskORrVSGnnsMAmCZiQ-Mufi5j2u4lXeQQ,10872
 nucliadb/reader/api/v1/export_import.py,sha256=x4VBNDFjnlY1nIt5kdq0eZTB_DeRzGzT8T7uB7wUhNU,6448
 nucliadb/reader/api/v1/knowledgebox.py,sha256=Uu-yPB8KKZt1VaFrFNMMaXOvLsclBJDK9dzZ9lF2ctI,3645
 nucliadb/reader/api/v1/learning_config.py,sha256=t_KqQBBbhpo0m6nQTkYmNdZsLVmW53SLcHMrCWiQMrk,6536
@@ -365,7 +365,7 @@ nucliadb/writer/api/v1/vectorsets.py,sha256=F3iMViL5G95_Tns4aO2SOA0DwAzxK2_P8MXx
 nucliadb/writer/resource/__init__.py,sha256=cp15ZcFnHvpcu_5-aK2A4uUyvuZVV_MJn4bIXMa20ks,835
 nucliadb/writer/resource/audit.py,sha256=FvxMZPzrNHtd31HgpZEvxzwAkbxJTZRhPLqRYYJi3tA,1426
 nucliadb/writer/resource/basic.py,sha256=44GK8M9EEVoAUfGiabdLrrpENqeFwNn7qwxF2AHhQGg,10504
-nucliadb/writer/resource/field.py,sha256=me3Ig-11evXEJl3UNJsmZyHbs_VZYMa-WIKOHZPsO3E,21175
+nucliadb/writer/resource/field.py,sha256=1fHm2N-3LcgQDE3BUDiJ0ToCiwY1Di7UbA6fcJs8BIk,21193
 nucliadb/writer/resource/origin.py,sha256=pvhUDdU0mlWPUcpoQi4LDUJaRtfjzVVrA8XcGVI_N8k,2021
 nucliadb/writer/tus/__init__.py,sha256=Kera0BtxoDX0ngPftXiMjNgjrhtQ3l2XFc5nJqSBOJY,5498
 nucliadb/writer/tus/azure.py,sha256=XhWAlWTM0vmXcXtuEPYjjeEhuZjiZXZu8q9WsJ7omFE,4107
@@ -376,8 +376,8 @@ nucliadb/writer/tus/local.py,sha256=7jYa_w9b-N90jWgN2sQKkNcomqn6JMVBOVeDOVYJHto,
 nucliadb/writer/tus/s3.py,sha256=vu1BGg4VqJ_x2P1u2BxqPKlSfw5orT_a3R-Ln5oPUpU,8483
 nucliadb/writer/tus/storage.py,sha256=ToqwjoYnjI4oIcwzkhha_MPxi-k4Jk3Lt55zRwaC1SM,2903
 nucliadb/writer/tus/utils.py,sha256=MSdVbRsRSZVdkaum69_0wku7X3p5wlZf4nr6E0GMKbw,2556
-nucliadb-6.7.0.post4800.dist-info/METADATA,sha256=p5YogXctzlcnUaCPjJowa0_MXNaVlxDYJdow5vfEqrs,4158
-nucliadb-6.7.0.post4800.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nucliadb-6.7.0.post4800.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
-nucliadb-6.7.0.post4800.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
-nucliadb-6.7.0.post4800.dist-info/RECORD,,
+nucliadb-6.7.0.post4805.dist-info/METADATA,sha256=uVCY9B5etvUvVS9QDwmPSGBJfJdGiMnOe0HRsHqh0JU,4158
+nucliadb-6.7.0.post4805.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nucliadb-6.7.0.post4805.dist-info/entry_points.txt,sha256=XqGfgFDuY3zXQc8ewXM2TRVjTModIq851zOsgrmaXx4,1268
+nucliadb-6.7.0.post4805.dist-info/top_level.txt,sha256=hwYhTVnX7jkQ9gJCkVrbqEG1M4lT2F_iPQND1fCzF80,20
+nucliadb-6.7.0.post4805.dist-info/RECORD,,