test1

backend/alembic/versions/f71470ba9274_add_prompt_length_limit.py

@@ -1,50 +0,0 @@
-"""add prompt length limit
-
-Revision ID: f71470ba9274
-Revises: 6a804aeb4830
-Create Date: 2025-04-01 15:07:14.977435
-
-"""
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision = "f71470ba9274"
-down_revision = "6a804aeb4830"
-branch_labels = None
-depends_on = None
-
-
-def upgrade() -> None:
-    op.alter_column(
-        "prompt",
-        "system_prompt",
-        existing_type=sa.TEXT(),
-        type_=sa.String(length=8000),
-        existing_nullable=False,
-    )
-    op.alter_column(
-        "prompt",
-        "task_prompt",
-        existing_type=sa.TEXT(),
-        type_=sa.String(length=8000),
-        existing_nullable=False,
-    )
-
-
-def downgrade() -> None:
-    op.alter_column(
-        "prompt",
-        "system_prompt",
-        existing_type=sa.String(length=8000),
-        type_=sa.TEXT(),
-        existing_nullable=False,
-    )
-    op.alter_column(
-        "prompt",
-        "task_prompt",
-        existing_type=sa.String(length=8000),
-        type_=sa.TEXT(),
-        existing_nullable=False,
-    )

backend/alembic/versions/f7505c5b0284_updated_constraints_for_ccpairs.py

@@ -1,77 +0,0 @@
-"""updated constraints for ccpairs
-
-Revision ID: f7505c5b0284
-Revises: f71470ba9274
-Create Date: 2025-04-01 17:50:42.504818
-
-"""
-from alembic import op
-
-
-# revision identifiers, used by Alembic.
-revision = "f7505c5b0284"
-down_revision = "f71470ba9274"
-branch_labels = None
-depends_on = None
-
-
-def upgrade() -> None:
-    # 1) Drop the old foreign-key constraints
-    op.drop_constraint(
-        "document_by_connector_credential_pair_connector_id_fkey",
-        "document_by_connector_credential_pair",
-        type_="foreignkey",
-    )
-    op.drop_constraint(
-        "document_by_connector_credential_pair_credential_id_fkey",
-        "document_by_connector_credential_pair",
-        type_="foreignkey",
-    )
-
-    # 2) Re-add them with ondelete='CASCADE'
-    op.create_foreign_key(
-        "document_by_connector_credential_pair_connector_id_fkey",
-        source_table="document_by_connector_credential_pair",
-        referent_table="connector",
-        local_cols=["connector_id"],
-        remote_cols=["id"],
-        ondelete="CASCADE",
-    )
-    op.create_foreign_key(
-        "document_by_connector_credential_pair_credential_id_fkey",
-        source_table="document_by_connector_credential_pair",
-        referent_table="credential",
-        local_cols=["credential_id"],
-        remote_cols=["id"],
-        ondelete="CASCADE",
-    )
-
-
-def downgrade() -> None:
-    # Reverse the changes for rollback
-    op.drop_constraint(
-        "document_by_connector_credential_pair_connector_id_fkey",
-        "document_by_connector_credential_pair",
-        type_="foreignkey",
-    )
-    op.drop_constraint(
-        "document_by_connector_credential_pair_credential_id_fkey",
-        "document_by_connector_credential_pair",
-        type_="foreignkey",
-    )
-
-    # Recreate without CASCADE
-    op.create_foreign_key(
-        "document_by_connector_credential_pair_connector_id_fkey",
-        "document_by_connector_credential_pair",
-        "connector",
-        ["connector_id"],
-        ["id"],
-    )
-    op.create_foreign_key(
-        "document_by_connector_credential_pair_credential_id_fkey",
-        "document_by_connector_credential_pair",
-        "credential",
-        ["credential_id"],
-        ["id"],
-    )

backend/onyx/background/celery/apps/app_base.py

@@ -1,6 +1,5 @@
 import logging
 import multiprocessing
-import os
 import time
 from typing import Any
 from typing import cast
@@ -306,7 +305,7 @@ def wait_for_db(sender: Any, **kwargs: Any) -> None:
 
 
 def on_secondary_worker_init(sender: Any, **kwargs: Any) -> None:
-    logger.info(f"Running as a secondary celery worker: pid={os.getpid()}")
+    logger.info("Running as a secondary celery worker.")
 
     # Set up variables for waiting on primary worker
     WAIT_INTERVAL = 5

backend/onyx/background/celery/apps/client.py

@@ -1,7 +0,0 @@
-from celery import Celery
-
-import onyx.background.celery.apps.app_base as app_base
-
-celery_app = Celery(__name__)
-celery_app.config_from_object("onyx.background.celery.configs.client")
-celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]

backend/onyx/background/celery/apps/primary.py

@@ -1,5 +1,4 @@
 import logging
-import os
 from typing import Any
 from typing import cast
 
@@ -96,7 +95,7 @@ def on_worker_init(sender: Worker, **kwargs: Any) -> None:
     app_base.wait_for_db(sender, **kwargs)
     app_base.wait_for_vespa_or_shutdown(sender, **kwargs)
 
-    logger.info(f"Running as the primary celery worker: pid={os.getpid()}")
+    logger.info("Running as the primary celery worker.")
 
     # Less startup checks in multi-tenant case
     if MULTI_TENANT:

backend/onyx/background/celery/configs/client.py

@@ -1,16 +0,0 @@
-import onyx.background.celery.configs.base as shared_config
-
-broker_url = shared_config.broker_url
-broker_connection_retry_on_startup = shared_config.broker_connection_retry_on_startup
-broker_pool_limit = shared_config.broker_pool_limit
-broker_transport_options = shared_config.broker_transport_options
-
-redis_socket_keepalive = shared_config.redis_socket_keepalive
-redis_retry_on_timeout = shared_config.redis_retry_on_timeout
-redis_backend_health_check_interval = shared_config.redis_backend_health_check_interval
-
-result_backend = shared_config.result_backend
-result_expires = shared_config.result_expires  # 86400 seconds is the default
-
-task_default_priority = shared_config.task_default_priority
-task_acks_late = shared_config.task_acks_late

backend/onyx/background/celery/versioned_apps/client.py

@@ -1,20 +0,0 @@
-"""Factory stub for running celery worker / celery beat.
-This code is different from the primary/beat stubs because there is no EE version to
-fetch. Port over the code in those files if we add an EE version of this worker.
-
-This is an app stub purely for sending tasks as a client.
-"""
-from celery import Celery
-
-from onyx.utils.variable_functionality import set_is_ee_based_on_env_variable
-
-set_is_ee_based_on_env_variable()
-
-
-def get_app() -> Celery:
-    from onyx.background.celery.apps.client import celery_app
-
-    return celery_app
-
-
-app = get_app()

backend/onyx/connectors/confluence/utils.py

@@ -13,7 +13,6 @@ from typing import TYPE_CHECKING
 from typing import TypeVar
 from urllib.parse import parse_qs
 from urllib.parse import quote
-from urllib.parse import urljoin
 from urllib.parse import urlparse
 
 import requests
@@ -343,14 +342,9 @@ def build_confluence_document_id(
     Returns:
         str: The document id
     """
-
-    # NOTE: urljoin is tricky and will drop the last segment of the base if it doesn't
-    # end with "/" because it believes that makes it a file.
-    final_url = base_url.rstrip("/") + "/"
-    if is_cloud and not final_url.endswith("/wiki/"):
-        final_url = urljoin(final_url, "wiki") + "/"
-    final_url = urljoin(final_url, content_url.lstrip("/"))
-    return final_url
+    if is_cloud and not base_url.endswith("/wiki"):
+        base_url += "/wiki"
+    return f"{base_url}{content_url}"
 
 
 def datetime_from_string(datetime_string: str) -> datetime:
@@ -460,19 +454,6 @@ def _handle_http_error(e: requests.HTTPError, attempt: int) -> int:
         logger.warning("HTTPError with `None` as response or as headers")
         raise e
 
-    # Confluence Server returns 403 when rate limited
-    if e.response.status_code == 403:
-        FORBIDDEN_MAX_RETRY_ATTEMPTS = 7
-        FORBIDDEN_RETRY_DELAY = 10
-        if attempt < FORBIDDEN_MAX_RETRY_ATTEMPTS:
-            logger.warning(
-                "403 error. This sometimes happens when we hit "
-                f"Confluence rate limits. Retrying in {FORBIDDEN_RETRY_DELAY} seconds..."
-            )
-            return FORBIDDEN_RETRY_DELAY
-
-        raise e
-
     if (
         e.response.status_code != 429
         and RATE_LIMIT_MESSAGE_LOWERCASE not in e.response.text.lower()

backend/onyx/connectors/salesforce/utils.py

@@ -25,6 +25,8 @@ class SalesforceObject:
         )
 
 
+# te
+
 # This defines the base path for all data files relative to this file
 # AKA BE CAREFUL WHEN MOVING THIS FILE
 BASE_DATA_PATH = os.path.join(os.path.dirname(__file__), "data")

backend/onyx/context/search/pipeline.py

@@ -5,13 +5,11 @@ from typing import cast
 
 from sqlalchemy.orm import Session
 
-from onyx.chat.models import ContextualPruningConfig
 from onyx.chat.models import PromptConfig
 from onyx.chat.models import SectionRelevancePiece
 from onyx.chat.prune_and_merge import _merge_sections
 from onyx.chat.prune_and_merge import ChunkRange
 from onyx.chat.prune_and_merge import merge_chunk_intervals
-from onyx.chat.prune_and_merge import prune_and_merge_sections
 from onyx.configs.chat_configs import DISABLE_LLM_DOC_RELEVANCE
 from onyx.context.search.enums import LLMEvaluationType
 from onyx.context.search.enums import QueryFlow
@@ -63,7 +61,6 @@ class SearchPipeline:
         | None = None,
         rerank_metrics_callback: Callable[[RerankMetricsContainer], None] | None = None,
         prompt_config: PromptConfig | None = None,
-        contextual_pruning_config: ContextualPruningConfig | None = None,
     ):
         # NOTE: The Search Request contains a lot of fields that are overrides, many of them can be None
         # and typically are None. The preprocessing will fetch default values to replace these empty overrides.
@@ -80,9 +77,6 @@ class SearchPipeline:
         self.search_settings = get_current_search_settings(db_session)
         self.document_index = get_default_document_index(self.search_settings, None)
         self.prompt_config: PromptConfig | None = prompt_config
-        self.contextual_pruning_config: ContextualPruningConfig | None = (
-            contextual_pruning_config
-        )
 
         # Preprocessing steps generate this
         self._search_query: SearchQuery | None = None
@@ -426,26 +420,7 @@ class SearchPipeline:
         if self._final_context_sections is not None:
             return self._final_context_sections
 
-        if (
-            self.contextual_pruning_config is not None
-            and self.prompt_config is not None
-        ):
-            self._final_context_sections = prune_and_merge_sections(
-                sections=self.reranked_sections,
-                section_relevance_list=None,
-                prompt_config=self.prompt_config,
-                llm_config=self.llm.config,
-                question=self.search_query.query,
-                contextual_pruning_config=self.contextual_pruning_config,
-            )
-
-        else:
-            logger.error(
-                "Contextual pruning or prompt config not set, using default merge"
-            )
-            self._final_context_sections = _merge_sections(
-                sections=self.reranked_sections
-            )
+        self._final_context_sections = _merge_sections(sections=self.reranked_sections)
         return self._final_context_sections
 
     @property

backend/onyx/db/models.py

@@ -703,11 +703,7 @@ class Connector(Base):
     )
     documents_by_connector: Mapped[
         list["DocumentByConnectorCredentialPair"]
-    ] = relationship(
-        "DocumentByConnectorCredentialPair",
-        back_populates="connector",
-        passive_deletes=True,
-    )
+    ] = relationship("DocumentByConnectorCredentialPair", back_populates="connector")
 
     # synchronize this validation logic with RefreshFrequencySchema etc on front end
     # until we have a centralized validation schema
@@ -761,11 +757,7 @@ class Credential(Base):
     )
     documents_by_credential: Mapped[
         list["DocumentByConnectorCredentialPair"]
-    ] = relationship(
-        "DocumentByConnectorCredentialPair",
-        back_populates="credential",
-        passive_deletes=True,
-    )
+    ] = relationship("DocumentByConnectorCredentialPair", back_populates="credential")
 
     user: Mapped[User | None] = relationship("User", back_populates="credentials")
 
@@ -1118,10 +1110,10 @@ class DocumentByConnectorCredentialPair(Base):
     id: Mapped[str] = mapped_column(ForeignKey("document.id"), primary_key=True)
     # TODO: transition this to use the ConnectorCredentialPair id directly
     connector_id: Mapped[int] = mapped_column(
-        ForeignKey("connector.id", ondelete="CASCADE"), primary_key=True
+        ForeignKey("connector.id"), primary_key=True
     )
     credential_id: Mapped[int] = mapped_column(
-        ForeignKey("credential.id", ondelete="CASCADE"), primary_key=True
+        ForeignKey("credential.id"), primary_key=True
     )
 
     # used to better keep track of document counts at a connector level
@@ -1131,10 +1123,10 @@ class DocumentByConnectorCredentialPair(Base):
     has_been_indexed: Mapped[bool] = mapped_column(Boolean)
 
     connector: Mapped[Connector] = relationship(
-        "Connector", back_populates="documents_by_connector", passive_deletes=True
+        "Connector", back_populates="documents_by_connector"
     )
     credential: Mapped[Credential] = relationship(
-        "Credential", back_populates="documents_by_credential", passive_deletes=True
+        "Credential", back_populates="documents_by_credential"
     )
 
     __table_args__ = (
@@ -1658,8 +1650,8 @@ class Prompt(Base):
     )
     name: Mapped[str] = mapped_column(String)
     description: Mapped[str] = mapped_column(String)
-    system_prompt: Mapped[str] = mapped_column(String(length=8000))
-    task_prompt: Mapped[str] = mapped_column(String(length=8000))
+    system_prompt: Mapped[str] = mapped_column(Text)
+    task_prompt: Mapped[str] = mapped_column(Text)
     include_citations: Mapped[bool] = mapped_column(Boolean, default=True)
     datetime_aware: Mapped[bool] = mapped_column(Boolean, default=True)
     # Default prompts are configured via backend during deployment

backend/onyx/llm/utils.py

@@ -602,7 +602,7 @@ def get_max_input_tokens(
     )
 
     if input_toks <= 0:
-        return GEN_AI_MODEL_FALLBACK_MAX_TOKENS
+        raise RuntimeError("No tokens for input for the LLM given settings")
 
     return input_toks
 

backend/onyx/server/documents/cc_pair.py

@@ -21,7 +21,7 @@ from onyx.background.celery.tasks.external_group_syncing.tasks import (
 from onyx.background.celery.tasks.pruning.tasks import (
     try_creating_prune_generator_task,
 )
-from onyx.background.celery.versioned_apps.client import app as client_app
+from onyx.background.celery.versioned_apps.primary import app as primary_app
 from onyx.background.indexing.models import IndexAttemptErrorPydantic
 from onyx.configs.constants import OnyxCeleryPriority
 from onyx.configs.constants import OnyxCeleryTask
@@ -219,7 +219,7 @@ def update_cc_pair_status(
                     continue
 
                 # Revoke the task to prevent it from running
-                client_app.control.revoke(index_payload.celery_task_id)
+                primary_app.control.revoke(index_payload.celery_task_id)
 
                 # If it is running, then signaling for termination will get the
                 # watchdog thread to kill the spawned task
@@ -238,7 +238,7 @@ def update_cc_pair_status(
     db_session.commit()
 
     # this speeds up the start of indexing by firing the check immediately
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_INDEXING,
         kwargs=dict(tenant_id=tenant_id),
         priority=OnyxCeleryPriority.HIGH,
@@ -376,7 +376,7 @@ def prune_cc_pair(
         f"{cc_pair.connector.name} connector."
     )
     payload_id = try_creating_prune_generator_task(
-        client_app, cc_pair, db_session, r, tenant_id
+        primary_app, cc_pair, db_session, r, tenant_id
     )
     if not payload_id:
         raise HTTPException(
@@ -450,7 +450,7 @@ def sync_cc_pair(
         f"{cc_pair.connector.name} connector."
     )
     payload_id = try_creating_permissions_sync_task(
-        client_app, cc_pair_id, r, tenant_id
+        primary_app, cc_pair_id, r, tenant_id
     )
     if not payload_id:
         raise HTTPException(
@@ -524,7 +524,7 @@ def sync_cc_pair_groups(
         f"{cc_pair.connector.name} connector."
     )
     payload_id = try_creating_external_group_sync_task(
-        client_app, cc_pair_id, r, tenant_id
+        primary_app, cc_pair_id, r, tenant_id
     )
     if not payload_id:
         raise HTTPException(
@@ -634,7 +634,7 @@ def associate_credential_to_connector(
         )
 
         # trigger indexing immediately
-        client_app.send_task(
+        primary_app.send_task(
             OnyxCeleryTask.CHECK_FOR_INDEXING,
             priority=OnyxCeleryPriority.HIGH,
             kwargs={"tenant_id": tenant_id},

backend/onyx/server/documents/connector.py

@@ -20,7 +20,7 @@ from onyx.auth.users import current_admin_user
 from onyx.auth.users import current_chat_accessible_user
 from onyx.auth.users import current_curator_or_admin_user
 from onyx.auth.users import current_user
-from onyx.background.celery.versioned_apps.client import app as client_app
+from onyx.background.celery.versioned_apps.primary import app as primary_app
 from onyx.configs.app_configs import ENABLED_CONNECTOR_TYPES
 from onyx.configs.app_configs import MOCK_CONNECTOR_FILE_PATH
 from onyx.configs.constants import DocumentSource
@@ -928,7 +928,7 @@ def create_connector_with_mock_credential(
         )
 
         # trigger indexing immediately
-        client_app.send_task(
+        primary_app.send_task(
             OnyxCeleryTask.CHECK_FOR_INDEXING,
             priority=OnyxCeleryPriority.HIGH,
             kwargs={"tenant_id": tenant_id},
@@ -1314,7 +1314,7 @@ def trigger_indexing_for_cc_pair(
     # run the beat task to pick up the triggers immediately
     priority = OnyxCeleryPriority.HIGHEST if is_user_file else OnyxCeleryPriority.HIGH
     logger.info(f"Sending indexing check task with priority {priority}")
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_INDEXING,
         priority=priority,
         kwargs={"tenant_id": tenant_id},

backend/onyx/server/features/document_set/api.py

@@ -6,7 +6,7 @@ from sqlalchemy.orm import Session
 
 from onyx.auth.users import current_curator_or_admin_user
 from onyx.auth.users import current_user
-from onyx.background.celery.versioned_apps.client import app as client_app
+from onyx.background.celery.versioned_apps.primary import app as primary_app
 from onyx.configs.constants import OnyxCeleryPriority
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.db.document_set import check_document_sets_are_public
@@ -52,7 +52,7 @@ def create_document_set(
     except Exception as e:
         raise HTTPException(status_code=400, detail=str(e))
 
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_VESPA_SYNC_TASK,
         kwargs={"tenant_id": tenant_id},
         priority=OnyxCeleryPriority.HIGH,
@@ -85,7 +85,7 @@ def patch_document_set(
     except Exception as e:
         raise HTTPException(status_code=400, detail=str(e))
 
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_VESPA_SYNC_TASK,
         kwargs={"tenant_id": tenant_id},
         priority=OnyxCeleryPriority.HIGH,
@@ -108,7 +108,7 @@ def delete_document_set(
     except Exception as e:
         raise HTTPException(status_code=400, detail=str(e))
 
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_VESPA_SYNC_TASK,
         kwargs={"tenant_id": tenant_id},
         priority=OnyxCeleryPriority.HIGH,

backend/onyx/server/manage/administrative.py

@@ -10,7 +10,7 @@ from sqlalchemy.orm import Session
 
 from onyx.auth.users import current_admin_user
 from onyx.auth.users import current_curator_or_admin_user
-from onyx.background.celery.versioned_apps.client import app as client_app
+from onyx.background.celery.versioned_apps.primary import app as primary_app
 from onyx.configs.app_configs import GENERATIVE_MODEL_ACCESS_CHECK_FREQ
 from onyx.configs.constants import DocumentSource
 from onyx.configs.constants import KV_GEN_AI_KEY_CHECK_TIME
@@ -192,7 +192,7 @@ def create_deletion_attempt_for_connector_id(
     db_session.commit()
 
     # run the beat task to pick up this deletion from the db immediately
-    client_app.send_task(
+    primary_app.send_task(
         OnyxCeleryTask.CHECK_FOR_CONNECTOR_DELETION,
         priority=OnyxCeleryPriority.HIGH,
         kwargs={"tenant_id": tenant_id},

backend/onyx/tools/tool_implementations/search/search_tool.py

@@ -376,7 +376,6 @@ class SearchTool(Tool[SearchToolOverrideKwargs]):
             db_session=alternate_db_session or self.db_session,
             prompt_config=self.prompt_config,
             retrieved_sections_callback=retrieved_sections_callback,
-            contextual_pruning_config=self.contextual_pruning_config,
         )
 
         search_query_info = SearchQueryInfo(
@@ -448,7 +447,6 @@ class SearchTool(Tool[SearchToolOverrideKwargs]):
             db_session=self.db_session,
             bypass_acl=self.bypass_acl,
             prompt_config=self.prompt_config,
-            contextual_pruning_config=self.contextual_pruning_config,
         )
 
         # Log what we're doing

backend/tests/integration/common_utils/managers/document.py

@@ -165,18 +165,17 @@ class DocumentManager:
             doc["fields"]["document_id"]: doc["fields"] for doc in retrieved_docs_dict
         }
 
-        # NOTE(rkuo): too much log spam
         # Left this here for debugging purposes.
-        # import json
+        import json
 
-        # print("DEBUGGING DOCUMENTS")
-        # print(retrieved_docs)
-        # for doc in retrieved_docs.values():
-        #     printable_doc = doc.copy()
-        #     print(printable_doc.keys())
-        #     printable_doc.pop("embeddings")
-        #     printable_doc.pop("title_embedding")
-        #     print(json.dumps(printable_doc, indent=2))
+        print("DEBUGGING DOCUMENTS")
+        print(retrieved_docs)
+        for doc in retrieved_docs.values():
+            printable_doc = doc.copy()
+            print(printable_doc.keys())
+            printable_doc.pop("embeddings")
+            printable_doc.pop("title_embedding")
+            print(json.dumps(printable_doc, indent=2))
 
         for document in cc_pair.documents:
             retrieved_doc = retrieved_docs.get(document.id)

backend/tests/integration/common_utils/managers/index_attempt.py

@@ -1,4 +1,3 @@
-import time
 from datetime import datetime
 from datetime import timedelta
 from urllib.parse import urlencode
@@ -192,7 +191,7 @@ class IndexAttemptManager:
         user_performing_action: DATestUser | None = None,
     ) -> None:
         """Wait for an IndexAttempt to complete"""
-        start = time.monotonic()
+        start = datetime.now()
         while True:
             index_attempt = IndexAttemptManager.get_index_attempt_by_id(
                 index_attempt_id=index_attempt_id,
@@ -204,7 +203,7 @@ class IndexAttemptManager:
                 print(f"IndexAttempt {index_attempt_id} completed")
                 return
 
-            elapsed = time.monotonic() - start
+            elapsed = (datetime.now() - start).total_seconds()
             if elapsed > timeout:
                 raise TimeoutError(
                     f"IndexAttempt {index_attempt_id} did not complete within {timeout} seconds"

backend/tests/integration/common_utils/managers/user.py

@@ -313,3 +313,29 @@ class UserManager:
         )
         response.raise_for_status()
         return UserInfo(**response.json())
+
+    @staticmethod
+    def invite_users(
+        user_performing_action: DATestUser,
+        emails: list[str],
+    ) -> int:
+        response = requests.put(
+            url=f"{API_SERVER_URL}/manage/admin/users",
+            json={"emails": emails},
+            headers=user_performing_action.headers,
+        )
+        response.raise_for_status()
+        return response.json()
+
+    @staticmethod
+    def remove_invited_user(
+        user_performing_action: DATestUser,
+        user_email: str,
+    ) -> int:
+        response = requests.patch(
+            url=f"{API_SERVER_URL}/manage/admin/remove-invited-user",
+            json={"user_email": user_email},
+            headers=user_performing_action.headers,
+        )
+        response.raise_for_status()
+        return response.json()

backend/tests/integration/common_utils/reset.py

@@ -22,6 +22,7 @@ from onyx.document_index.document_index_utils import get_multipass_config
 from onyx.document_index.vespa.index import DOCUMENT_ID_ENDPOINT
 from onyx.document_index.vespa.index import VespaIndex
 from onyx.indexing.models import IndexingSetting
+from onyx.redis.redis_pool import get_redis_client
 from onyx.setup import setup_postgres
 from onyx.setup import setup_vespa
 from onyx.utils.logger import setup_logger
@@ -237,6 +238,12 @@ def reset_vespa() -> None:
             time.sleep(5)
 
 
+def reset_redis() -> None:
+    """Reset the Redis database."""
+    redis_client = get_redis_client()
+    redis_client.flushall()
+
+
 def reset_postgres_multitenant() -> None:
     """Reset the Postgres database for all tenants in a multitenant setup."""
 
@@ -341,6 +348,8 @@ def reset_all() -> None:
     reset_postgres()
     logger.info("Resetting Vespa...")
     reset_vespa()
+    logger.info("Resetting Redis...")
+    reset_redis()
 
 
 def reset_all_multitenant() -> None:

backend/tests/integration/tests/auth/test_user_invitation.py

@@ -0,0 +1,38 @@
+import pytest
+from requests import HTTPError
+
+from onyx.auth.schemas import UserRole
+from tests.integration.common_utils.managers.user import UserManager
+from tests.integration.common_utils.test_models import DATestUser
+
+
+def test_inviting_users_flow(reset: None) -> None:
+    """
+    Test that verifies the functionality around inviting users:
+      1. Creating an admin user
+      2. Admin inviting a new user
+      3. Invited user successfully signing in
+      4. Non-invited user attempting to sign in (should result in an error)
+    """
+    # 1) Create an admin user (the first user created is automatically admin)
+    admin_user: DATestUser = UserManager.create(name="admin_user")
+    assert admin_user is not None
+    assert UserManager.is_role(admin_user, UserRole.ADMIN)
+
+    # 2) Admin invites a new user
+    invited_email = "invited_user@test.com"
+    invite_response = UserManager.invite_users(admin_user, [invited_email])
+
+    assert invite_response == 1
+
+    # 3) The invited user successfully registers/logs in
+    invited_user: DATestUser = UserManager.create(
+        name="invited_user", email=invited_email
+    )
+    assert invited_user is not None
+    assert invited_user.email == invited_email
+    assert UserManager.is_role(invited_user, UserRole.BASIC)
+
+    # 4) A non-invited user attempts to sign in/register (should fail)
+    with pytest.raises(HTTPError):
+        UserManager.create(name="uninvited_user", email="uninvited_user@test.com")

web/src/app/admin/assistants/AssistantEditor.tsx

@@ -242,7 +242,15 @@ export function AssistantEditor({
     enabledToolsMap[tool.id] = personaCurrentToolIds.includes(tool.id);
   });
 
-  const { selectedFiles, selectedFolders } = useDocumentsContext();
+  const {
+    selectedFiles,
+    selectedFolders,
+    addSelectedFile,
+    removeSelectedFile,
+    addSelectedFolder,
+    removeSelectedFolder,
+    clearSelectedItems,
+  } = useDocumentsContext();
 
   const [showVisibilityWarning, setShowVisibilityWarning] = useState(false);