nit

* new mit integration test template

* edit

* fix problem with ACL type tags and MIT testing for test_connector_deletion

* fix test_connector_deletion_for_overlapping_connectors

* disable some enterprise only tests in MIT version

* disable a bunch of user group / curator tests in MIT version

* wire off more tests

* typo fix

---------

Co-authored-by: Richard Kuo (Onyx) <rkuo@onyx.app>
Co-authored-by: Richard Kuo <rkuo@rkuo.com>

.github/workflows/pr-mit-integration-tests.yml

@@ -0,0 +1,209 @@
+name: Run MIT Integration Tests v2
+concurrency:
+  group: Run-MIT-Integration-Tests-${{ github.workflow }}-${{ github.head_ref || github.event.workflow_run.head_branch || github.run_id }}
+  cancel-in-progress: true
+
+on:
+  merge_group:
+  pull_request:
+    branches:
+      - main
+      - "release/**"
+
+env:
+  OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+  SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
+  CONFLUENCE_TEST_SPACE_URL: ${{ secrets.CONFLUENCE_TEST_SPACE_URL }}
+  CONFLUENCE_USER_NAME: ${{ secrets.CONFLUENCE_USER_NAME }}
+  CONFLUENCE_ACCESS_TOKEN: ${{ secrets.CONFLUENCE_ACCESS_TOKEN }}
+
+jobs:
+  integration-tests:
+    # See https://runs-on.com/runners/linux/
+    runs-on: [runs-on, runner=32cpu-linux-x64, "run-id=${{ github.run_id }}"]
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_TOKEN }}
+
+      # tag every docker image with "test" so that we can spin up the correct set
+      # of images during testing
+
+      # We don't need to build the Web Docker image since it's not yet used
+      # in the integration tests. We have a separate action to verify that it builds
+      # successfully.
+      - name: Pull Web Docker image
+        run: |
+          docker pull onyxdotapp/onyx-web-server:latest
+          docker tag onyxdotapp/onyx-web-server:latest onyxdotapp/onyx-web-server:test
+
+      # we use the runs-on cache for docker builds
+      # in conjunction with runs-on runners, it has better speed and unlimited caching
+      # https://runs-on.com/caching/s3-cache-for-github-actions/
+      # https://runs-on.com/caching/docker/
+      # https://github.com/moby/buildkit#s3-cache-experimental
+
+      # images are built and run locally for testing purposes. Not pushed.
+      - name: Build Backend Docker image
+        uses: ./.github/actions/custom-build-and-push
+        with:
+          context: ./backend
+          file: ./backend/Dockerfile
+          platforms: linux/amd64
+          tags: onyxdotapp/onyx-backend:test
+          push: false
+          load: true
+          cache-from: type=s3,prefix=cache/${{ github.repository }}/integration-tests/backend/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }}
+          cache-to: type=s3,prefix=cache/${{ github.repository }}/integration-tests/backend/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }},mode=max
+
+      - name: Build Model Server Docker image
+        uses: ./.github/actions/custom-build-and-push
+        with:
+          context: ./backend
+          file: ./backend/Dockerfile.model_server
+          platforms: linux/amd64
+          tags: onyxdotapp/onyx-model-server:test
+          push: false
+          load: true
+          cache-from: type=s3,prefix=cache/${{ github.repository }}/integration-tests/model-server/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }}
+          cache-to: type=s3,prefix=cache/${{ github.repository }}/integration-tests/model-server/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }},mode=max
+
+      - name: Build integration test Docker image
+        uses: ./.github/actions/custom-build-and-push
+        with:
+          context: ./backend
+          file: ./backend/tests/integration/Dockerfile
+          platforms: linux/amd64
+          tags: onyxdotapp/onyx-integration:test
+          push: false
+          load: true
+          cache-from: type=s3,prefix=cache/${{ github.repository }}/integration-tests/integration/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }}
+          cache-to: type=s3,prefix=cache/${{ github.repository }}/integration-tests/integration/,region=${{ env.RUNS_ON_AWS_REGION }},bucket=${{ env.RUNS_ON_S3_BUCKET_CACHE }},mode=max
+
+      # NOTE: Use pre-ping/null pool to reduce flakiness due to dropped connections
+      - name: Start Docker containers
+        run: |
+          cd deployment/docker_compose
+          AUTH_TYPE=basic \
+          POSTGRES_POOL_PRE_PING=true \
+          POSTGRES_USE_NULL_POOL=true \
+          REQUIRE_EMAIL_VERIFICATION=false \
+          DISABLE_TELEMETRY=true \
+          IMAGE_TAG=test \
+          INTEGRATION_TESTS_MODE=true \
+          docker compose -f docker-compose.dev.yml -p onyx-stack up -d
+        id: start_docker
+
+      - name: Wait for service to be ready
+        run: |
+          echo "Starting wait-for-service script..."
+
+          docker logs -f onyx-stack-api_server-1 &
+
+          start_time=$(date +%s)
+          timeout=300  # 5 minutes in seconds
+
+          while true; do
+            current_time=$(date +%s)
+            elapsed_time=$((current_time - start_time))
+            
+            if [ $elapsed_time -ge $timeout ]; then
+              echo "Timeout reached. Service did not become ready in 5 minutes."
+              exit 1
+            fi
+            
+            # Use curl with error handling to ignore specific exit code 56
+            response=$(curl -s -o /dev/null -w "%{http_code}" http://localhost:8080/health || echo "curl_error")
+            
+            if [ "$response" = "200" ]; then
+              echo "Service is ready!"
+              break
+            elif [ "$response" = "curl_error" ]; then
+              echo "Curl encountered an error, possibly exit code 56. Continuing to retry..."
+            else
+              echo "Service not ready yet (HTTP status $response). Retrying in 5 seconds..."
+            fi
+            
+            sleep 5
+          done
+          echo "Finished waiting for service."
+
+      - name: Start Mock Services
+        run: |
+          cd backend/tests/integration/mock_services
+          docker compose -f docker-compose.mock-it-services.yml \
+            -p mock-it-services-stack up -d
+
+      # NOTE: Use pre-ping/null to reduce flakiness due to dropped connections
+      - name: Run Standard Integration Tests
+        run: |
+          echo "Running integration tests..."
+          docker run --rm --network onyx-stack_default \
+            --name test-runner \
+            -e POSTGRES_HOST=relational_db \
+            -e POSTGRES_USER=postgres \
+            -e POSTGRES_PASSWORD=password \
+            -e POSTGRES_DB=postgres \
+            -e POSTGRES_POOL_PRE_PING=true \
+            -e POSTGRES_USE_NULL_POOL=true \
+            -e VESPA_HOST=index \
+            -e REDIS_HOST=cache \
+            -e API_SERVER_HOST=api_server \
+            -e OPENAI_API_KEY=${OPENAI_API_KEY} \
+            -e SLACK_BOT_TOKEN=${SLACK_BOT_TOKEN} \
+            -e CONFLUENCE_TEST_SPACE_URL=${CONFLUENCE_TEST_SPACE_URL} \
+            -e CONFLUENCE_USER_NAME=${CONFLUENCE_USER_NAME} \
+            -e CONFLUENCE_ACCESS_TOKEN=${CONFLUENCE_ACCESS_TOKEN} \
+            -e TEST_WEB_HOSTNAME=test-runner \
+            -e MOCK_CONNECTOR_SERVER_HOST=mock_connector_server \
+            -e MOCK_CONNECTOR_SERVER_PORT=8001 \
+            onyxdotapp/onyx-integration:test \
+            /app/tests/integration/tests \
+            /app/tests/integration/connector_job_tests
+        continue-on-error: true
+        id: run_tests
+
+      - name: Check test results
+        run: |
+          if [ ${{ steps.run_tests.outcome }} == 'failure' ]; then
+            echo "Integration tests failed. Exiting with error."
+            exit 1
+          else
+            echo "All integration tests passed successfully."
+          fi
+
+      # ------------------------------------------------------------
+      # Always gather logs BEFORE "down":
+      - name: Dump API server logs
+        if: always()
+        run: |
+          cd deployment/docker_compose
+          docker compose -f docker-compose.dev.yml -p onyx-stack logs --no-color api_server > $GITHUB_WORKSPACE/api_server.log || true
+
+      - name: Dump all-container logs (optional)
+        if: always()
+        run: |
+          cd deployment/docker_compose
+          docker compose -f docker-compose.dev.yml -p onyx-stack logs --no-color > $GITHUB_WORKSPACE/docker-compose.log || true
+
+      - name: Upload logs
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: docker-all-logs
+          path: ${{ github.workspace }}/docker-compose.log
+      # ------------------------------------------------------------
+
+      - name: Stop Docker containers
+        if: always()
+        run: |
+          cd deployment/docker_compose
+          docker compose -f docker-compose.dev.yml -p onyx-stack down -v

backend/ee/onyx/access/access.py

@@ -93,12 +93,12 @@ def _get_access_for_documents(
         )
 
         # To avoid collisions of group namings between connectors, they need to be prefixed
-        access_map[document_id] = DocumentAccess(
-            user_emails=non_ee_access.user_emails,
-            user_groups=set(user_group_info.get(document_id, [])),
+        access_map[document_id] = DocumentAccess.build(
+            user_emails=list(non_ee_access.user_emails),
+            user_groups=user_group_info.get(document_id, []),
             is_public=is_public_anywhere,
-            external_user_emails=ext_u_emails,
-            external_user_group_ids=ext_u_groups,
+            external_user_emails=list(ext_u_emails),
+            external_user_group_ids=list(ext_u_groups),
         )
     return access_map
 

backend/ee/onyx/chat/process_message.py

@@ -2,7 +2,6 @@ from ee.onyx.server.query_and_chat.models import OneShotQAResponse
 from onyx.chat.models import AllCitations
 from onyx.chat.models import LLMRelevanceFilterResponse
 from onyx.chat.models import OnyxAnswerPiece
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import QADocsResponse
 from onyx.chat.models import StreamingError
 from onyx.chat.process_message import ChatPacketStream
@@ -32,8 +31,6 @@ def gather_stream_for_answer_api(
             response.llm_selected_doc_indices = packet.llm_selected_doc_indices
         elif isinstance(packet, AllCitations):
             response.citations = packet.citations
-        elif isinstance(packet, OnyxContexts):
-            response.contexts = packet
 
     if answer:
         response.answer = answer

backend/ee/onyx/server/query_and_chat/chat_backend.py

@@ -14,7 +14,6 @@ from ee.onyx.server.query_and_chat.models import (
     BasicCreateChatMessageWithHistoryRequest,
 )
 from ee.onyx.server.query_and_chat.models import ChatBasicResponse
-from ee.onyx.server.query_and_chat.models import SimpleDoc
 from onyx.auth.users import current_user
 from onyx.chat.chat_utils import combine_message_thread
 from onyx.chat.chat_utils import create_chat_chain
@@ -56,25 +55,6 @@ logger = setup_logger()
 router = APIRouter(prefix="/chat")
 
 
-def _translate_doc_response_to_simple_doc(
-    doc_response: QADocsResponse,
-) -> list[SimpleDoc]:
-    return [
-        SimpleDoc(
-            id=doc.document_id,
-            semantic_identifier=doc.semantic_identifier,
-            link=doc.link,
-            blurb=doc.blurb,
-            match_highlights=[
-                highlight for highlight in doc.match_highlights if highlight
-            ],
-            source_type=doc.source_type,
-            metadata=doc.metadata,
-        )
-        for doc in doc_response.top_documents
-    ]
-
-
 def _get_final_context_doc_indices(
     final_context_docs: list[LlmDoc] | None,
     top_docs: list[SavedSearchDoc] | None,
@@ -111,9 +91,6 @@ def _convert_packet_stream_to_response(
         elif isinstance(packet, QADocsResponse):
             response.top_documents = packet.top_documents
 
-            # TODO: deprecate `simple_search_docs`
-            response.simple_search_docs = _translate_doc_response_to_simple_doc(packet)
-
             # This is a no-op if agent_sub_questions hasn't already been filled
             if packet.level is not None and packet.level_question_num is not None:
                 id = (packet.level, packet.level_question_num)

backend/ee/onyx/server/query_and_chat/models.py

@@ -8,7 +8,6 @@ from pydantic import model_validator
 
 from ee.onyx.server.manage.models import StandardAnswer
 from onyx.chat.models import CitationInfo
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import PersonaOverrideConfig
 from onyx.chat.models import QADocsResponse
 from onyx.chat.models import SubQuestionIdentifier
@@ -164,8 +163,6 @@ class ChatBasicResponse(BaseModel):
     cited_documents: dict[int, str] | None = None
 
     # FOR BACKWARDS COMPATIBILITY
-    # TODO: deprecate both of these
-    simple_search_docs: list[SimpleDoc] | None = None
     llm_chunks_indices: list[int] | None = None
 
     # agentic fields
@@ -220,4 +217,3 @@ class OneShotQAResponse(BaseModel):
     llm_selected_doc_indices: list[int] | None = None
     error_msg: str | None = None
     chat_message_id: int | None = None
-    contexts: OnyxContexts | None = None

backend/onyx/access/access.py

@@ -18,7 +18,7 @@ def _get_access_for_document(
         document_id=document_id,
     )
 
-    return DocumentAccess.build(
+    doc_access = DocumentAccess.build(
         user_emails=info[1] if info and info[1] else [],
         user_groups=[],
         external_user_emails=[],
@@ -26,6 +26,8 @@ def _get_access_for_document(
         is_public=info[2] if info else False,
     )
 
+    return doc_access
+
 
 def get_access_for_document(
     document_id: str,
@@ -38,12 +40,12 @@ def get_access_for_document(
 
 
 def get_null_document_access() -> DocumentAccess:
-    return DocumentAccess(
-        user_emails=set(),
-        user_groups=set(),
+    return DocumentAccess.build(
+        user_emails=[],
+        user_groups=[],
         is_public=False,
-        external_user_emails=set(),
-        external_user_group_ids=set(),
+        external_user_emails=[],
+        external_user_group_ids=[],
     )
 
 
@@ -56,18 +58,18 @@ def _get_access_for_documents(
         document_ids=document_ids,
     )
     doc_access = {
-        document_id: DocumentAccess(
-            user_emails=set([email for email in user_emails if email]),
+        document_id: DocumentAccess.build(
+            user_emails=[email for email in user_emails if email],
             # MIT version will wipe all groups and external groups on update
-            user_groups=set(),
+            user_groups=[],
             is_public=is_public,
-            external_user_emails=set(),
-            external_user_group_ids=set(),
+            external_user_emails=[],
+            external_user_group_ids=[],
         )
         for document_id, user_emails, is_public in document_access_info
     }
 
-    # Sometimes the document has not be indexed by the indexing job yet, in those cases
+    # Sometimes the document has not been indexed by the indexing job yet, in those cases
     # the document does not exist and so we use least permissive. Specifically the EE version
     # checks the MIT version permissions and creates a superset. This ensures that this flow
     # does not fail even if the Document has not yet been indexed.

backend/onyx/access/models.py

@@ -56,34 +56,46 @@ class DocExternalAccess:
         )
 
 
-@dataclass(frozen=True)
+@dataclass(frozen=True, init=False)
 class DocumentAccess(ExternalAccess):
     # User emails for Onyx users, None indicates admin
     user_emails: set[str | None]
+
     # Names of user groups associated with this document
     user_groups: set[str]
 
-    def to_acl(self) -> set[str]:
-        return set(
-            [
-                prefix_user_email(user_email)
-                for user_email in self.user_emails
-                if user_email
-            ]
-            + [prefix_user_group(group_name) for group_name in self.user_groups]
-            + [
-                prefix_user_email(user_email)
-                for user_email in self.external_user_emails
-            ]
-            + [
-                # The group names are already prefixed by the source type
-                # This adds an additional prefix of "external_group:"
-                prefix_external_group(group_name)
-                for group_name in self.external_user_group_ids
-            ]
-            + ([PUBLIC_DOC_PAT] if self.is_public else [])
+    external_user_emails: set[str]
+    external_user_group_ids: set[str]
+    is_public: bool
+
+    def __init__(self) -> None:
+        raise TypeError(
+            "Use `DocumentAccess.build(...)` instead of creating an instance directly."
         )
 
+    def to_acl(self) -> set[str]:
+        # the acl's emitted by this function are prefixed by type
+        # to get the native objects, access the member variables directly
+
+        acl_set: set[str] = set()
+        for user_email in self.user_emails:
+            if user_email:
+                acl_set.add(prefix_user_email(user_email))
+
+        for group_name in self.user_groups:
+            acl_set.add(prefix_user_group(group_name))
+
+        for external_user_email in self.external_user_emails:
+            acl_set.add(prefix_user_email(external_user_email))
+
+        for external_group_id in self.external_user_group_ids:
+            acl_set.add(prefix_external_group(external_group_id))
+
+        if self.is_public:
+            acl_set.add(PUBLIC_DOC_PAT)
+
+        return acl_set
+
     @classmethod
     def build(
         cls,
@@ -93,29 +105,32 @@ class DocumentAccess(ExternalAccess):
         external_user_group_ids: list[str],
         is_public: bool,
     ) -> "DocumentAccess":
-        return cls(
-            external_user_emails={
-                prefix_user_email(external_email)
-                for external_email in external_user_emails
-            },
-            external_user_group_ids={
-                prefix_external_group(external_group_id)
-                for external_group_id in external_user_group_ids
-            },
-            user_emails={
-                prefix_user_email(user_email)
-                for user_email in user_emails
-                if user_email
-            },
-            user_groups=set(user_groups),
-            is_public=is_public,
+        """Don't prefix incoming data wth acl type, prefix on read from to_acl!"""
+
+        obj = object.__new__(cls)
+        object.__setattr__(
+            obj, "user_emails", {user_email for user_email in user_emails if user_email}
         )
+        object.__setattr__(obj, "user_groups", set(user_groups))
+        object.__setattr__(
+            obj,
+            "external_user_emails",
+            {external_email for external_email in external_user_emails},
+        )
+        object.__setattr__(
+            obj,
+            "external_user_group_ids",
+            {external_group_id for external_group_id in external_user_group_ids},
+        )
+        object.__setattr__(obj, "is_public", is_public)
+
+        return obj
 
 
-default_public_access = DocumentAccess(
-    external_user_emails=set(),
-    external_user_group_ids=set(),
-    user_emails=set(),
-    user_groups=set(),
+default_public_access = DocumentAccess.build(
+    external_user_emails=[],
+    external_user_group_ids=[],
+    user_emails=[],
+    user_groups=[],
     is_public=True,
 )

backend/onyx/agents/agent_search/basic/utils.py

@@ -7,7 +7,6 @@ from langgraph.types import StreamWriter
 
 from onyx.agents.agent_search.shared_graph_utils.utils import write_custom_event
 from onyx.chat.models import LlmDoc
-from onyx.chat.models import OnyxContext
 from onyx.chat.stream_processing.answer_response_handler import AnswerResponseHandler
 from onyx.chat.stream_processing.answer_response_handler import CitationResponseHandler
 from onyx.chat.stream_processing.answer_response_handler import (
@@ -24,7 +23,7 @@ def process_llm_stream(
     should_stream_answer: bool,
     writer: StreamWriter,
     final_search_results: list[LlmDoc] | None = None,
-    displayed_search_results: list[OnyxContext] | list[LlmDoc] | None = None,
+    displayed_search_results: list[LlmDoc] | None = None,
 ) -> AIMessageChunk:
     tool_call_chunk = AIMessageChunk(content="")
 

backend/onyx/agents/agent_search/deep_search/initial/generate_initial_answer/nodes/generate_initial_answer.py

@@ -156,7 +156,6 @@ def generate_initial_answer(
     for tool_response in yield_search_responses(
         query=question,
         get_retrieved_sections=lambda: answer_generation_documents.context_documents,
-        get_reranked_sections=lambda: answer_generation_documents.streaming_documents,
         get_final_context_sections=lambda: answer_generation_documents.context_documents,
         search_query_info=query_info,
         get_section_relevance=lambda: relevance_list,

backend/onyx/agents/agent_search/deep_search/main/nodes/generate_validate_refined_answer.py

@@ -183,7 +183,6 @@ def generate_validate_refined_answer(
     for tool_response in yield_search_responses(
         query=question,
         get_retrieved_sections=lambda: answer_generation_documents.context_documents,
-        get_reranked_sections=lambda: answer_generation_documents.streaming_documents,
         get_final_context_sections=lambda: answer_generation_documents.context_documents,
         search_query_info=query_info,
         get_section_relevance=lambda: relevance_list,

backend/onyx/agents/agent_search/deep_search/shared/expanded_retrieval/nodes/format_results.py

@@ -57,7 +57,6 @@ def format_results(
         for tool_response in yield_search_responses(
             query=state.question,
             get_retrieved_sections=lambda: reranked_documents,
-            get_reranked_sections=lambda: state.retrieved_documents,
             get_final_context_sections=lambda: reranked_documents,
             search_query_info=query_info,
             get_section_relevance=lambda: relevance_list,

backend/onyx/agents/agent_search/orchestration/nodes/use_tool_response.py

@@ -13,9 +13,7 @@ from onyx.tools.tool_implementations.search.search_tool import (
     SEARCH_RESPONSE_SUMMARY_ID,
 )
 from onyx.tools.tool_implementations.search.search_tool import SearchResponseSummary
-from onyx.tools.tool_implementations.search.search_utils import (
-    context_from_inference_section,
-)
+from onyx.tools.tool_implementations.search.search_utils import section_to_llm_doc
 from onyx.tools.tool_implementations.search_like_tool_utils import (
     FINAL_CONTEXT_DOCUMENTS_ID,
 )
@@ -59,9 +57,7 @@ def basic_use_tool_response(
             search_response_summary = cast(SearchResponseSummary, yield_item.response)
             for section in search_response_summary.top_sections:
                 if section.center_chunk.document_id not in initial_search_results:
-                    initial_search_results.append(
-                        context_from_inference_section(section)
-                    )
+                    initial_search_results.append(section_to_llm_doc(section))
 
     new_tool_call_chunk = AIMessageChunk(content="")
     if not agent_config.behavior.skip_gen_ai_answer_generation:

backend/onyx/background/celery/tasks/connector_deletion/tasks.py

@@ -25,14 +25,8 @@ from onyx.configs.constants import OnyxRedisLocks
 from onyx.configs.constants import OnyxRedisSignals
 from onyx.db.connector import fetch_connector_by_id
 from onyx.db.connector_credential_pair import add_deletion_failure_message
-from onyx.db.connector_credential_pair import (
-    delete_connector_credential_pair__no_commit,
-)
 from onyx.db.connector_credential_pair import get_connector_credential_pair_from_id
 from onyx.db.connector_credential_pair import get_connector_credential_pairs
-from onyx.db.document import (
-    delete_all_documents_by_connector_credential_pair__no_commit,
-)
 from onyx.db.document import get_document_ids_for_connector_credential_pair
 from onyx.db.document_set import delete_document_set_cc_pair_relationship__no_commit
 from onyx.db.engine import get_session_with_current_tenant
@@ -449,27 +443,15 @@ def monitor_connector_deletion_taskset(
             connector_id_to_delete = cc_pair.connector_id
             credential_id_to_delete = cc_pair.credential_id
 
-            # Explicitly delete document by connector credential pair records before deleting the connector
-            # This is needed because connector_id is a primary key in that table and cascading deletes won't work
-            delete_all_documents_by_connector_credential_pair__no_commit(
-                db_session=db_session,
-                connector_id=connector_id_to_delete,
-                credential_id=credential_id_to_delete,
-            )
+            # No need to explicitly delete DocumentByConnectorCredentialPair records anymore
+            # as we have proper cascade relationships set up in the models
 
-            # Flush to ensure document deletion happens before connector deletion
+            # Flush to ensure all operations happen in sequence
             db_session.flush()
 
-            # Expire the cc_pair to ensure SQLAlchemy doesn't try to manage its state
-            # related to the deleted DocumentByConnectorCredentialPair during commit
-            db_session.expire(cc_pair)
+            # Delete the cc-pair directly
+            db_session.delete(cc_pair)
 
-            # finally, delete the cc-pair
-            delete_connector_credential_pair__no_commit(
-                db_session=db_session,
-                connector_id=connector_id_to_delete,
-                credential_id=credential_id_to_delete,
-            )
             # if there are no credentials left, delete the connector
             connector = fetch_connector_by_id(
                 db_session=db_session,

backend/onyx/chat/models.py

@@ -194,17 +194,6 @@ class StreamingError(BaseModel):
     stack_trace: str | None = None
 
 
-class OnyxContext(BaseModel):
-    content: str
-    document_id: str
-    semantic_identifier: str
-    blurb: str
-
-
-class OnyxContexts(BaseModel):
-    contexts: list[OnyxContext]
-
-
 class OnyxAnswer(BaseModel):
     answer: str | None
 
@@ -270,7 +259,6 @@ class PersonaOverrideConfig(BaseModel):
 AnswerQuestionPossibleReturn = (
     OnyxAnswerPiece
     | CitationInfo
-    | OnyxContexts
     | FileChatDisplay
     | CustomToolResponse
     | StreamingError

backend/onyx/chat/process_message.py

@@ -29,7 +29,6 @@ from onyx.chat.models import LLMRelevanceFilterResponse
 from onyx.chat.models import MessageResponseIDInfo
 from onyx.chat.models import MessageSpecificCitations
 from onyx.chat.models import OnyxAnswerPiece
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import PromptConfig
 from onyx.chat.models import QADocsResponse
 from onyx.chat.models import RefinedAnswerImprovement
@@ -131,7 +130,6 @@ from onyx.tools.tool_implementations.internet_search.internet_search_tool import
 from onyx.tools.tool_implementations.search.search_tool import (
     FINAL_CONTEXT_DOCUMENTS_ID,
 )
-from onyx.tools.tool_implementations.search.search_tool import SEARCH_DOC_CONTENT_ID
 from onyx.tools.tool_implementations.search.search_tool import (
     SEARCH_RESPONSE_SUMMARY_ID,
 )
@@ -300,7 +298,6 @@ def _get_force_search_settings(
 ChatPacket = (
     StreamingError
     | QADocsResponse
-    | OnyxContexts
     | LLMRelevanceFilterResponse
     | FinalUsedContextDocsResponse
     | ChatMessageDetail
@@ -919,8 +916,6 @@ def stream_chat_message_objects(
                             response=custom_tool_response.tool_result,
                             tool_name=custom_tool_response.tool_name,
                         )
-                elif packet.id == SEARCH_DOC_CONTENT_ID and include_contexts:
-                    yield cast(OnyxContexts, packet.response)
 
             elif isinstance(packet, StreamStopInfo):
                 if packet.stop_reason == StreamStopReason.FINISHED:

backend/onyx/chat/prune_and_merge.py

@@ -301,6 +301,10 @@ def prune_sections(
 
 
 def _merge_doc_chunks(chunks: list[InferenceChunk]) -> InferenceSection:
+    assert (
+        len(set([chunk.document_id for chunk in chunks])) == 1
+    ), "One distinct document must be passed into merge_doc_chunks"
+
     # Assuming there are no duplicates by this point
     sorted_chunks = sorted(chunks, key=lambda x: x.chunk_id)
 

backend/onyx/chat/stream_processing/utils.py

@@ -3,7 +3,6 @@ from collections.abc import Sequence
 from pydantic import BaseModel
 
 from onyx.chat.models import LlmDoc
-from onyx.chat.models import OnyxContext
 from onyx.context.search.models import InferenceChunk
 
 
@@ -12,7 +11,7 @@ class DocumentIdOrderMapping(BaseModel):
 
 
 def map_document_id_order(
-    chunks: Sequence[InferenceChunk | LlmDoc | OnyxContext], one_indexed: bool = True
+    chunks: Sequence[InferenceChunk | LlmDoc], one_indexed: bool = True
 ) -> DocumentIdOrderMapping:
     order_mapping = {}
     current = 1 if one_indexed else 0

backend/onyx/connectors/highspot/client.py

@@ -75,7 +75,7 @@ class HighspotClient:
 
         self.key = key
         self.secret = secret
-        self.base_url = base_url
+        self.base_url = base_url.rstrip("/") + "/"
         self.timeout = timeout
 
         # Set up session with retry logic

backend/onyx/context/search/pipeline.py

@@ -339,6 +339,12 @@ class SearchPipeline:
         self._retrieved_sections = self._get_sections()
         return self._retrieved_sections
 
+    @property
+    def merged_retrieved_sections(self) -> list[InferenceSection]:
+        """Should be used to display in the UI in order to prevent displaying
+        multiple sections for the same document as separate "documents"."""
+        return _merge_sections(sections=self.retrieved_sections)
+
     @property
     def reranked_sections(self) -> list[InferenceSection]:
         """Reranking is always done at the chunk level since section merging could create arbitrarily
@@ -415,6 +421,10 @@ class SearchPipeline:
                 raise ValueError(
                     "Basic search evaluation operation called while DISABLE_LLM_DOC_RELEVANCE is enabled."
                 )
+            # NOTE: final_context_sections must be accessed before accessing self._postprocessing_generator
+            # since the property sets the generator. DO NOT REMOVE.
+            _ = self.final_context_sections
+
             self._section_relevance = next(
                 cast(
                     Iterator[list[SectionRelevancePiece]],

backend/onyx/db/document.py

@@ -555,28 +555,6 @@ def delete_documents_by_connector_credential_pair__no_commit(
     db_session.execute(stmt)
 
 
-def delete_all_documents_by_connector_credential_pair__no_commit(
-    db_session: Session,
-    connector_id: int,
-    credential_id: int,
-) -> None:
-    """Deletes all document by connector credential pair entries for a specific connector and credential.
-    This is primarily used during connector deletion to ensure all references are removed
-    before deleting the connector itself. This is crucial because connector_id is part of the
-    primary key in DocumentByConnectorCredentialPair, and attempting to delete the Connector
-    would otherwise try to set the foreign key to NULL, which fails for primary keys.
-
-    NOTE: Does not commit the transaction, this must be done by the caller.
-    """
-    stmt = delete(DocumentByConnectorCredentialPair).where(
-        and_(
-            DocumentByConnectorCredentialPair.connector_id == connector_id,
-            DocumentByConnectorCredentialPair.credential_id == credential_id,
-        )
-    )
-    db_session.execute(stmt)
-
-
 def delete_documents__no_commit(db_session: Session, document_ids: list[str]) -> None:
     db_session.execute(delete(DbDocument).where(DbDocument.id.in_(document_ids)))
 

backend/onyx/db/models.py

@@ -694,7 +694,11 @@ class Connector(Base):
     )
     documents_by_connector: Mapped[
         list["DocumentByConnectorCredentialPair"]
-    ] = relationship("DocumentByConnectorCredentialPair", back_populates="connector")
+    ] = relationship(
+        "DocumentByConnectorCredentialPair",
+        back_populates="connector",
+        cascade="all, delete-orphan",
+    )
 
     # synchronize this validation logic with RefreshFrequencySchema etc on front end
     # until we have a centralized validation schema
@@ -748,7 +752,11 @@ class Credential(Base):
     )
     documents_by_credential: Mapped[
         list["DocumentByConnectorCredentialPair"]
-    ] = relationship("DocumentByConnectorCredentialPair", back_populates="credential")
+    ] = relationship(
+        "DocumentByConnectorCredentialPair",
+        back_populates="credential",
+        cascade="all, delete-orphan",
+    )
 
     user: Mapped[User | None] = relationship("User", back_populates="credentials")
 

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

@@ -12,7 +12,6 @@ from onyx.chat.models import AnswerStyleConfig
 from onyx.chat.models import ContextualPruningConfig
 from onyx.chat.models import DocumentPruningConfig
 from onyx.chat.models import LlmDoc
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import PromptConfig
 from onyx.chat.models import SectionRelevancePiece
 from onyx.chat.prompt_builder.answer_prompt_builder import AnswerPromptBuilder
@@ -42,9 +41,6 @@ from onyx.tools.models import SearchQueryInfo
 from onyx.tools.models import SearchToolOverrideKwargs
 from onyx.tools.models import ToolResponse
 from onyx.tools.tool import Tool
-from onyx.tools.tool_implementations.search.search_utils import (
-    context_from_inference_section,
-)
 from onyx.tools.tool_implementations.search.search_utils import llm_doc_to_dict
 from onyx.tools.tool_implementations.search_like_tool_utils import (
     build_next_prompt_for_search_like_tool,
@@ -58,7 +54,6 @@ from onyx.utils.special_types import JSON_ro
 logger = setup_logger()
 
 SEARCH_RESPONSE_SUMMARY_ID = "search_response_summary"
-SEARCH_DOC_CONTENT_ID = "search_doc_content"
 SECTION_RELEVANCE_LIST_ID = "section_relevance_list"
 SEARCH_EVALUATION_ID = "llm_doc_eval"
 QUERY_FIELD = "query"
@@ -357,13 +352,13 @@ class SearchTool(Tool[SearchToolOverrideKwargs]):
             recency_bias_multiplier=search_pipeline.search_query.recency_bias_multiplier,
         )
         yield from yield_search_responses(
-            query,
-            lambda: search_pipeline.retrieved_sections,
-            lambda: search_pipeline.reranked_sections,
-            lambda: search_pipeline.final_context_sections,
-            search_query_info,
-            lambda: search_pipeline.section_relevance,
-            self,
+            query=query,
+            # give back the merged sections to prevent duplicate docs from appearing in the UI
+            get_retrieved_sections=lambda: search_pipeline.merged_retrieved_sections,
+            get_final_context_sections=lambda: search_pipeline.final_context_sections,
+            search_query_info=search_query_info,
+            get_section_relevance=lambda: search_pipeline.section_relevance,
+            search_tool=self,
         )
 
     def final_result(self, *args: ToolResponse) -> JSON_ro:
@@ -405,7 +400,6 @@ class SearchTool(Tool[SearchToolOverrideKwargs]):
 def yield_search_responses(
     query: str,
     get_retrieved_sections: Callable[[], list[InferenceSection]],
-    get_reranked_sections: Callable[[], list[InferenceSection]],
     get_final_context_sections: Callable[[], list[InferenceSection]],
     search_query_info: SearchQueryInfo,
     get_section_relevance: Callable[[], list[SectionRelevancePiece] | None],
@@ -423,16 +417,6 @@ def yield_search_responses(
         ),
     )
 
-    yield ToolResponse(
-        id=SEARCH_DOC_CONTENT_ID,
-        response=OnyxContexts(
-            contexts=[
-                context_from_inference_section(section)
-                for section in get_reranked_sections()
-            ]
-        ),
-    )
-
     section_relevance = get_section_relevance()
     yield ToolResponse(
         id=SECTION_RELEVANCE_LIST_ID,

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

@@ -1,5 +1,4 @@
 from onyx.chat.models import LlmDoc
-from onyx.chat.models import OnyxContext
 from onyx.context.search.models import InferenceSection
 from onyx.prompts.prompt_utils import clean_up_source
 
@@ -32,10 +31,23 @@ def section_to_dict(section: InferenceSection, section_num: int) -> dict:
     return doc_dict
 
 
-def context_from_inference_section(section: InferenceSection) -> OnyxContext:
-    return OnyxContext(
-        content=section.combined_content,
+def section_to_llm_doc(section: InferenceSection) -> LlmDoc:
+    possible_link_chunks = [section.center_chunk] + section.chunks
+    link: str | None = None
+    for chunk in possible_link_chunks:
+        if chunk.source_links:
+            link = list(chunk.source_links.values())[0]
+            break
+
+    return LlmDoc(
         document_id=section.center_chunk.document_id,
+        content=section.combined_content,
+        source_type=section.center_chunk.source_type,
         semantic_identifier=section.center_chunk.semantic_identifier,
+        metadata=section.center_chunk.metadata,
+        updated_at=section.center_chunk.updated_at,
         blurb=section.center_chunk.blurb,
+        link=link,
+        source_links=section.center_chunk.source_links,
+        match_highlights=section.center_chunk.match_highlights,
     )

backend/scripts/query_time_check/seed_dummy_docs.py

@@ -78,19 +78,19 @@ def generate_dummy_chunk(
     for i in range(number_of_document_sets):
         document_set_names.append(f"Document Set {i}")
 
-    user_emails: set[str | None] = set()
-    user_groups: set[str] = set()
-    external_user_emails: set[str] = set()
-    external_user_group_ids: set[str] = set()
+    user_emails: list[str | None] = []
+    user_groups: list[str] = []
+    external_user_emails: list[str] = []
+    external_user_group_ids: list[str] = []
     for i in range(number_of_acl_entries):
-        user_emails.add(f"user_{i}@example.com")
-        user_groups.add(f"group_{i}")
-        external_user_emails.add(f"external_user_{i}@example.com")
-        external_user_group_ids.add(f"external_group_{i}")
+        user_emails.append(f"user_{i}@example.com")
+        user_groups.append(f"group_{i}")
+        external_user_emails.append(f"external_user_{i}@example.com")
+        external_user_group_ids.append(f"external_group_{i}")
 
     return DocMetadataAwareIndexChunk.from_index_chunk(
         index_chunk=chunk,
-        access=DocumentAccess(
+        access=DocumentAccess.build(
             user_emails=user_emails,
             user_groups=user_groups,
             external_user_emails=external_user_emails,

backend/tests/integration/common_utils/constants.py

@@ -6,7 +6,7 @@ API_SERVER_PROTOCOL = os.getenv("API_SERVER_PROTOCOL") or "http"
 API_SERVER_HOST = os.getenv("API_SERVER_HOST") or "localhost"
 API_SERVER_PORT = os.getenv("API_SERVER_PORT") or "8080"
 API_SERVER_URL = f"{API_SERVER_PROTOCOL}://{API_SERVER_HOST}:{API_SERVER_PORT}"
-MAX_DELAY = 45
+MAX_DELAY = 60
 
 GENERAL_HEADERS = {"Content-Type": "application/json"}
 

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

@@ -5,6 +5,7 @@ import requests
 from requests.models import Response
 
 from onyx.context.search.models import RetrievalDetails
+from onyx.context.search.models import SavedSearchDoc
 from onyx.file_store.models import FileDescriptor
 from onyx.llm.override_models import LLMOverride
 from onyx.llm.override_models import PromptOverride
@@ -97,17 +98,24 @@ class ChatSessionManager:
         for data in response_data:
             if "rephrased_query" in data:
                 analyzed.rephrased_query = data["rephrased_query"]
-            elif "tool_name" in data:
+            if "tool_name" in data:
                 analyzed.tool_name = data["tool_name"]
                 analyzed.tool_result = (
                     data.get("tool_result")
                     if analyzed.tool_name == "run_search"
                     else None
                 )
-            elif "relevance_summaries" in data:
+            if "relevance_summaries" in data:
                 analyzed.relevance_summaries = data["relevance_summaries"]
-            elif "answer_piece" in data and data["answer_piece"]:
+            if "answer_piece" in data and data["answer_piece"]:
                 analyzed.full_message += data["answer_piece"]
+            if "top_documents" in data:
+                assert (
+                    analyzed.top_documents is None
+                ), "top_documents should only be set once"
+                analyzed.top_documents = [
+                    SavedSearchDoc(**doc) for doc in data["top_documents"]
+                ]
 
         return analyzed
 

backend/tests/integration/common_utils/test_models.py

@@ -10,6 +10,7 @@ from pydantic import Field
 from onyx.auth.schemas import UserRole
 from onyx.configs.constants import QAFeedbackType
 from onyx.context.search.enums import RecencyBiasSetting
+from onyx.context.search.models import SavedSearchDoc
 from onyx.db.enums import AccessType
 from onyx.server.documents.models import DocumentSource
 from onyx.server.documents.models import IndexAttemptSnapshot
@@ -157,7 +158,7 @@ class StreamedResponse(BaseModel):
     full_message: str = ""
     rephrased_query: str | None = None
     tool_name: str | None = None
-    top_documents: list[dict[str, Any]] | None = None
+    top_documents: list[SavedSearchDoc] | None = None
     relevance_summaries: list[dict[str, Any]] | None = None
     tool_result: Any | None = None
     user: str | None = None

backend/tests/integration/tests/auth/test_saml_user_conversion.py

@@ -1,3 +1,6 @@
+import os
+
+import pytest
 import requests
 
 from onyx.auth.schemas import UserRole
@@ -6,6 +9,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.test_models import DATestUser
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="SAML tests are enterprise only",
+)
 def test_saml_user_conversion(reset: None) -> None:
     """
     Test that SAML login correctly converts users with non-authenticated roles

backend/tests/integration/tests/connector/test_connector_deletion.py

@@ -5,6 +5,7 @@ This file contains tests for the following:
     - updates the document sets and user groups to remove the connector
 - Ensure that deleting a connector that is part of an overlapping document set and/or user group works as expected
 """
+import os
 from uuid import uuid4
 
 from sqlalchemy.orm import Session
@@ -32,6 +33,13 @@ from tests.integration.common_utils.vespa import vespa_fixture
 
 
 def test_connector_deletion(reset: None, vespa_client: vespa_fixture) -> None:
+    user_group_1: DATestUserGroup
+    user_group_2: DATestUserGroup
+
+    is_ee = (
+        os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() == "true"
+    )
+
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")
     # create api key
@@ -78,16 +86,17 @@ def test_connector_deletion(reset: None, vespa_client: vespa_fixture) -> None:
 
     print("Document sets created and synced")
 
-    # create user groups
-    user_group_1: DATestUserGroup = UserGroupManager.create(
-        cc_pair_ids=[cc_pair_1.id],
-        user_performing_action=admin_user,
-    )
-    user_group_2: DATestUserGroup = UserGroupManager.create(
-        cc_pair_ids=[cc_pair_1.id, cc_pair_2.id],
-        user_performing_action=admin_user,
-    )
-    UserGroupManager.wait_for_sync(user_performing_action=admin_user)
+    if is_ee:
+        # create user groups
+        user_group_1 = UserGroupManager.create(
+            cc_pair_ids=[cc_pair_1.id],
+            user_performing_action=admin_user,
+        )
+        user_group_2 = UserGroupManager.create(
+            cc_pair_ids=[cc_pair_1.id, cc_pair_2.id],
+            user_performing_action=admin_user,
+        )
+        UserGroupManager.wait_for_sync(user_performing_action=admin_user)
 
     # inject a finished index attempt and index attempt error (exercises foreign key errors)
     with Session(get_sqlalchemy_engine()) as db_session:
@@ -147,12 +156,13 @@ def test_connector_deletion(reset: None, vespa_client: vespa_fixture) -> None:
         )
 
     # Update local records to match the database for later comparison
-    user_group_1.cc_pair_ids = []
-    user_group_2.cc_pair_ids = [cc_pair_2.id]
     doc_set_1.cc_pair_ids = []
     doc_set_2.cc_pair_ids = [cc_pair_2.id]
     cc_pair_1.groups = []
-    cc_pair_2.groups = [user_group_2.id]
+    if is_ee:
+        cc_pair_2.groups = [user_group_2.id]
+    else:
+        cc_pair_2.groups = []
 
     CCPairManager.wait_for_deletion_completion(
         cc_pair_id=cc_pair_1.id, user_performing_action=admin_user
@@ -168,11 +178,15 @@ def test_connector_deletion(reset: None, vespa_client: vespa_fixture) -> None:
         verify_deleted=True,
     )
 
+    cc_pair_2_group_name_expected = []
+    if is_ee:
+        cc_pair_2_group_name_expected = [user_group_2.name]
+
     DocumentManager.verify(
         vespa_client=vespa_client,
         cc_pair=cc_pair_2,
         doc_set_names=[doc_set_2.name],
-        group_names=[user_group_2.name],
+        group_names=cc_pair_2_group_name_expected,
         doc_creating_user=admin_user,
         verify_deleted=False,
     )
@@ -193,15 +207,19 @@ def test_connector_deletion(reset: None, vespa_client: vespa_fixture) -> None:
         user_performing_action=admin_user,
     )
 
-    # validate user groups
-    UserGroupManager.verify(
-        user_group=user_group_1,
-        user_performing_action=admin_user,
-    )
-    UserGroupManager.verify(
-        user_group=user_group_2,
-        user_performing_action=admin_user,
-    )
+    if is_ee:
+        user_group_1.cc_pair_ids = []
+        user_group_2.cc_pair_ids = [cc_pair_2.id]
+
+        # validate user groups
+        UserGroupManager.verify(
+            user_group=user_group_1,
+            user_performing_action=admin_user,
+        )
+        UserGroupManager.verify(
+            user_group=user_group_2,
+            user_performing_action=admin_user,
+        )
 
 
 def test_connector_deletion_for_overlapping_connectors(
@@ -210,6 +228,13 @@ def test_connector_deletion_for_overlapping_connectors(
     """Checks to make sure that connectors with overlapping documents work properly. Specifically, that the overlapping
     document (1) still exists and (2) has the right document set / group post-deletion of one of the connectors.
     """
+    user_group_1: DATestUserGroup
+    user_group_2: DATestUserGroup
+
+    is_ee = (
+        os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() == "true"
+    )
+
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")
     # create api key
@@ -281,47 +306,48 @@ def test_connector_deletion_for_overlapping_connectors(
         doc_creating_user=admin_user,
     )
 
-    # create a user group and attach it to connector 1
-    user_group_1: DATestUserGroup = UserGroupManager.create(
-        name="Test User Group 1",
-        cc_pair_ids=[cc_pair_1.id],
-        user_performing_action=admin_user,
-    )
-    UserGroupManager.wait_for_sync(
-        user_groups_to_check=[user_group_1],
-        user_performing_action=admin_user,
-    )
-    cc_pair_1.groups = [user_group_1.id]
+    if is_ee:
+        # create a user group and attach it to connector 1
+        user_group_1 = UserGroupManager.create(
+            name="Test User Group 1",
+            cc_pair_ids=[cc_pair_1.id],
+            user_performing_action=admin_user,
+        )
+        UserGroupManager.wait_for_sync(
+            user_groups_to_check=[user_group_1],
+            user_performing_action=admin_user,
+        )
+        cc_pair_1.groups = [user_group_1.id]
 
-    print("User group 1 created and synced")
+        print("User group 1 created and synced")
 
-    # create a user group and attach it to connector 2
-    user_group_2: DATestUserGroup = UserGroupManager.create(
-        name="Test User Group 2",
-        cc_pair_ids=[cc_pair_2.id],
-        user_performing_action=admin_user,
-    )
-    UserGroupManager.wait_for_sync(
-        user_groups_to_check=[user_group_2],
-        user_performing_action=admin_user,
-    )
-    cc_pair_2.groups = [user_group_2.id]
+        # create a user group and attach it to connector 2
+        user_group_2 = UserGroupManager.create(
+            name="Test User Group 2",
+            cc_pair_ids=[cc_pair_2.id],
+            user_performing_action=admin_user,
+        )
+        UserGroupManager.wait_for_sync(
+            user_groups_to_check=[user_group_2],
+            user_performing_action=admin_user,
+        )
+        cc_pair_2.groups = [user_group_2.id]
 
-    print("User group 2 created and synced")
+        print("User group 2 created and synced")
 
-    # verify vespa document is in the user group
-    DocumentManager.verify(
-        vespa_client=vespa_client,
-        cc_pair=cc_pair_1,
-        group_names=[user_group_1.name, user_group_2.name],
-        doc_creating_user=admin_user,
-    )
-    DocumentManager.verify(
-        vespa_client=vespa_client,
-        cc_pair=cc_pair_2,
-        group_names=[user_group_1.name, user_group_2.name],
-        doc_creating_user=admin_user,
-    )
+        # verify vespa document is in the user group
+        DocumentManager.verify(
+            vespa_client=vespa_client,
+            cc_pair=cc_pair_1,
+            group_names=[user_group_1.name, user_group_2.name],
+            doc_creating_user=admin_user,
+        )
+        DocumentManager.verify(
+            vespa_client=vespa_client,
+            cc_pair=cc_pair_2,
+            group_names=[user_group_1.name, user_group_2.name],
+            doc_creating_user=admin_user,
+        )
 
     # delete connector 1
     CCPairManager.pause_cc_pair(
@@ -354,11 +380,15 @@ def test_connector_deletion_for_overlapping_connectors(
 
     # verify the document is not in any document sets
     # verify the document is only in user group 2
+    group_names_expected = []
+    if is_ee:
+        group_names_expected = [user_group_2.name]
+
     DocumentManager.verify(
         vespa_client=vespa_client,
         cc_pair=cc_pair_2,
         doc_set_names=[],
-        group_names=[user_group_2.name],
+        group_names=group_names_expected,
         doc_creating_user=admin_user,
         verify_deleted=False,
     )

backend/tests/integration/tests/dev_apis/test_knowledge_chat.py

@@ -1,3 +1,6 @@
+import os
+
+import pytest
 import requests
 
 from onyx.configs.constants import MessageType
@@ -12,6 +15,10 @@ from tests.integration.common_utils.test_models import DATestCCPair
 from tests.integration.common_utils.test_models import DATestUser
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="/chat/send-message-simple-with-history is enterprise only",
+)
 def test_all_stream_chat_message_objects_outputs(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/dev_apis/test_simple_chat_api.py

@@ -1,5 +1,7 @@
 import json
+import os
 
+import pytest
 import requests
 
 from onyx.configs.constants import MessageType
@@ -16,10 +18,11 @@ from tests.integration.common_utils.test_models import DATestCCPair
 from tests.integration.common_utils.test_models import DATestUser
 
 
-def test_send_message_simple_with_history(reset: None) -> None:
-    # Creating an admin user (first user created is automatically an admin)
-    admin_user: DATestUser = UserManager.create(name="admin_user")
-
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="/chat/send-message-simple-with-history tests are enterprise only",
+)
+def test_send_message_simple_with_history(reset: None, admin_user: DATestUser) -> None:
     # create connectors
     cc_pair_1: DATestCCPair = CCPairManager.create_from_scratch(
         user_performing_action=admin_user,
@@ -53,18 +56,22 @@ def test_send_message_simple_with_history(reset: None) -> None:
     response_json = response.json()
 
     # Check that the top document is the correct document
-    assert response_json["simple_search_docs"][0]["id"] == cc_pair_1.documents[0].id
     assert response_json["top_documents"][0]["document_id"] == cc_pair_1.documents[0].id
 
     # assert that the metadata is correct
     for doc in cc_pair_1.documents:
         found_doc = next(
-            (x for x in response_json["simple_search_docs"] if x["id"] == doc.id), None
+            (x for x in response_json["top_documents"] if x["document_id"] == doc.id),
+            None,
         )
         assert found_doc
         assert found_doc["metadata"]["document_id"] == doc.id
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="/chat/send-message-simple-with-history tests are enterprise only",
+)
 def test_using_reference_docs_with_simple_with_history_api_flow(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")
@@ -154,6 +161,10 @@ def test_using_reference_docs_with_simple_with_history_api_flow(reset: None) ->
     assert response_json["top_documents"][0]["document_id"] == cc_pair_1.documents[2].id
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="/chat/send-message-simple-with-history tests are enterprise only",
+)
 def test_send_message_simple_with_history_strict_json(
     new_admin_user: DATestUser | None,
 ) -> None:

backend/tests/integration/tests/permissions/test_cc_pair_permissions.py

@@ -2,6 +2,8 @@
 This file takes the happy path to adding a curator to a user group and then tests
 the permissions of the curator manipulating connector-credential pairs.
 """
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -15,6 +17,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and User Group tests are enterprise only",
+)
 def test_cc_pair_permissions(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_connector_permissions.py

@@ -2,6 +2,8 @@
 This file takes the happy path to adding a curator to a user group and then tests
 the permissions of the curator manipulating connectors.
 """
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -13,6 +15,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and user group tests are enterprise only",
+)
 def test_connector_permissions(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_credential_permissions.py

@@ -2,6 +2,8 @@
 This file takes the happy path to adding a curator to a user group and then tests
 the permissions of the curator manipulating credentials.
 """
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -12,6 +14,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and user group tests are enterprise only",
+)
 def test_credential_permissions(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_doc_set_permissions.py

@@ -1,3 +1,5 @@
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -10,6 +12,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and user group tests are enterprise only",
+)
 def test_doc_set_permissions_setup(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_persona_permissions.py

@@ -4,6 +4,8 @@ This file tests the permissions for creating and editing personas for different
 - Curators can edit personas that belong exclusively to groups they curate
 - Admins can edit all personas
 """
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -13,6 +15,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and user group tests are enterprise only",
+)
 def test_persona_permissions(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_user_role_permissions.py

@@ -1,6 +1,8 @@
 """
 This file tests the ability of different user types to set the role of other users.
 """
+import os
+
 import pytest
 from requests.exceptions import HTTPError
 
@@ -10,6 +12,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator and user group tests are enterprise only",
+)
 def test_user_role_setting_permissions(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/permissions/test_whole_curator_flow.py

@@ -1,6 +1,10 @@
 """
 This test tests the happy path for curator permissions
 """
+import os
+
+import pytest
+
 from onyx.db.enums import AccessType
 from onyx.db.models import UserRole
 from onyx.server.documents.models import DocumentSource
@@ -12,6 +16,10 @@ from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.managers.user_group import UserGroupManager
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator tests are enterprise only",
+)
 def test_whole_curator_flow(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")
@@ -89,6 +97,10 @@ def test_whole_curator_flow(reset: None) -> None:
     )
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Curator tests are enterprise only",
+)
 def test_global_curator_flow(reset: None) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/query_history/test_query_history.py

@@ -1,3 +1,4 @@
+import os
 from datetime import datetime
 from datetime import timedelta
 from datetime import timezone
@@ -63,6 +64,10 @@ def setup_chat_session(reset: None) -> tuple[DATestUser, str]:
     return admin_user, str(chat_session.id)
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Chat history tests are enterprise only",
+)
 def test_chat_history_endpoints(
     reset: None, setup_chat_session: tuple[DATestUser, str]
 ) -> None:
@@ -116,6 +121,10 @@ def test_chat_history_endpoints(
     assert len(history_response.items) == 0
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Chat history tests are enterprise only",
+)
 def test_chat_history_csv_export(
     reset: None, setup_chat_session: tuple[DATestUser, str]
 ) -> None:

backend/tests/integration/tests/query_history/test_query_history_pagination.py

@@ -1,5 +1,8 @@
+import os
 from datetime import datetime
 
+import pytest
+
 from onyx.configs.constants import QAFeedbackType
 from tests.integration.common_utils.managers.query_history import QueryHistoryManager
 from tests.integration.common_utils.test_models import DAQueryHistoryEntry
@@ -47,6 +50,10 @@ def _verify_query_history_pagination(
     assert all_expected_sessions == all_retrieved_sessions
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="Query history tests are enterprise only",
+)
 def test_query_history_pagination(reset: None) -> None:
     (
         admin_user,

backend/tests/integration/tests/streaming_endpoints/conftest.py

@@ -0,0 +1,42 @@
+from collections.abc import Callable
+
+import pytest
+
+from onyx.configs.constants import DocumentSource
+from tests.integration.common_utils.managers.api_key import APIKeyManager
+from tests.integration.common_utils.managers.cc_pair import CCPairManager
+from tests.integration.common_utils.managers.document import DocumentManager
+from tests.integration.common_utils.test_models import DATestAPIKey
+from tests.integration.common_utils.test_models import DATestUser
+from tests.integration.common_utils.test_models import SimpleTestDocument
+
+
+DocumentBuilderType = Callable[[list[str]], list[SimpleTestDocument]]
+
+
+@pytest.fixture
+def document_builder(admin_user: DATestUser) -> DocumentBuilderType:
+    api_key: DATestAPIKey = APIKeyManager.create(
+        user_performing_action=admin_user,
+    )
+
+    # create connector
+    cc_pair_1 = CCPairManager.create_from_scratch(
+        source=DocumentSource.INGESTION_API,
+        user_performing_action=admin_user,
+    )
+
+    def _document_builder(contents: list[str]) -> list[SimpleTestDocument]:
+        # seed documents
+        docs: list[SimpleTestDocument] = [
+            DocumentManager.seed_doc_with_content(
+                cc_pair=cc_pair_1,
+                content=content,
+                api_key=api_key,
+            )
+            for content in contents
+        ]
+
+        return docs
+
+    return _document_builder

backend/tests/integration/tests/streaming_endpoints/test_chat_stream.py

@@ -5,12 +5,11 @@ import pytest
 from tests.integration.common_utils.constants import API_SERVER_URL
 from tests.integration.common_utils.managers.chat import ChatSessionManager
 from tests.integration.common_utils.managers.llm_provider import LLMProviderManager
-from tests.integration.common_utils.managers.user import UserManager
 from tests.integration.common_utils.test_models import DATestUser
+from tests.integration.tests.streaming_endpoints.conftest import DocumentBuilderType
 
 
-def test_send_message_simple_with_history(reset: None) -> None:
-    admin_user: DATestUser = UserManager.create(name="admin_user")
+def test_send_message_simple_with_history(reset: None, admin_user: DATestUser) -> None:
     LLMProviderManager.create(user_performing_action=admin_user)
 
     test_chat_session = ChatSessionManager.create(user_performing_action=admin_user)
@@ -24,6 +23,44 @@ def test_send_message_simple_with_history(reset: None) -> None:
     assert len(response.full_message) > 0
 
 
+def test_send_message__basic_searches(
+    reset: None, admin_user: DATestUser, document_builder: DocumentBuilderType
+) -> None:
+    MESSAGE = "run a search for 'test'"
+    SHORT_DOC_CONTENT = "test"
+    LONG_DOC_CONTENT = "blah blah blah blah" * 100
+
+    LLMProviderManager.create(user_performing_action=admin_user)
+
+    short_doc = document_builder([SHORT_DOC_CONTENT])[0]
+
+    test_chat_session = ChatSessionManager.create(user_performing_action=admin_user)
+    response = ChatSessionManager.send_message(
+        chat_session_id=test_chat_session.id,
+        message=MESSAGE,
+        user_performing_action=admin_user,
+    )
+    assert response.top_documents is not None
+    assert len(response.top_documents) == 1
+    assert response.top_documents[0].document_id == short_doc.id
+
+    # make sure this doc is really long so that it will be split into multiple chunks
+    long_doc = document_builder([LONG_DOC_CONTENT])[0]
+
+    # new chat session for simplicity
+    test_chat_session = ChatSessionManager.create(user_performing_action=admin_user)
+    response = ChatSessionManager.send_message(
+        chat_session_id=test_chat_session.id,
+        message=MESSAGE,
+        user_performing_action=admin_user,
+    )
+    assert response.top_documents is not None
+    assert len(response.top_documents) == 2
+    # short doc should be more relevant and thus first
+    assert response.top_documents[0].document_id == short_doc.id
+    assert response.top_documents[1].document_id == long_doc.id
+
+
 @pytest.mark.skip(
     reason="enable for autorun when we have a testing environment with semantically useful data"
 )

backend/tests/integration/tests/usergroup/test_user_group_deletion.py

@@ -8,6 +8,10 @@ This tests the deletion of a user group with the following foreign key constrain
 - token_rate_limit (Not Implemented)
 - persona
 """
+import os
+
+import pytest
+
 from onyx.server.documents.models import DocumentSource
 from tests.integration.common_utils.managers.cc_pair import CCPairManager
 from tests.integration.common_utils.managers.credential import CredentialManager
@@ -25,6 +29,10 @@ from tests.integration.common_utils.test_models import DATestUserGroup
 from tests.integration.common_utils.vespa import vespa_fixture
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="User group tests are enterprise only",
+)
 def test_user_group_deletion(reset: None, vespa_client: vespa_fixture) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/integration/tests/usergroup/test_usergroup_syncing.py

@@ -1,3 +1,7 @@
+import os
+
+import pytest
+
 from onyx.server.documents.models import DocumentSource
 from tests.integration.common_utils.constants import NUM_DOCS
 from tests.integration.common_utils.managers.api_key import APIKeyManager
@@ -11,6 +15,10 @@ from tests.integration.common_utils.test_models import DATestUserGroup
 from tests.integration.common_utils.vespa import vespa_fixture
 
 
+@pytest.mark.skipif(
+    os.environ.get("ENABLE_PAID_ENTERPRISE_EDITION_FEATURES", "").lower() != "true",
+    reason="User group tests are enterprise only",
+)
 def test_removing_connector(reset: None, vespa_client: vespa_fixture) -> None:
     # Creating an admin user (first user created is automatically an admin)
     admin_user: DATestUser = UserManager.create(name="admin_user")

backend/tests/unit/onyx/chat/conftest.py

@@ -9,8 +9,6 @@ from onyx.chat.chat_utils import llm_doc_from_inference_section
 from onyx.chat.models import AnswerStyleConfig
 from onyx.chat.models import CitationConfig
 from onyx.chat.models import LlmDoc
-from onyx.chat.models import OnyxContext
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import PromptConfig
 from onyx.chat.prompt_builder.answer_prompt_builder import AnswerPromptBuilder
 from onyx.configs.constants import DocumentSource
@@ -19,7 +17,6 @@ from onyx.context.search.models import InferenceSection
 from onyx.llm.interfaces import LLM
 from onyx.llm.interfaces import LLMConfig
 from onyx.tools.models import ToolResponse
-from onyx.tools.tool_implementations.search.search_tool import SEARCH_DOC_CONTENT_ID
 from onyx.tools.tool_implementations.search.search_tool import SearchTool
 from onyx.tools.tool_implementations.search_like_tool_utils import (
     FINAL_CONTEXT_DOCUMENTS_ID,
@@ -120,24 +117,7 @@ def mock_search_results(
 
 
 @pytest.fixture
-def mock_contexts(mock_inference_sections: list[InferenceSection]) -> OnyxContexts:
-    return OnyxContexts(
-        contexts=[
-            OnyxContext(
-                content=section.combined_content,
-                document_id=section.center_chunk.document_id,
-                semantic_identifier=section.center_chunk.semantic_identifier,
-                blurb=section.center_chunk.blurb,
-            )
-            for section in mock_inference_sections
-        ]
-    )
-
-
-@pytest.fixture
-def mock_search_tool(
-    mock_contexts: OnyxContexts, mock_search_results: list[LlmDoc]
-) -> MagicMock:
+def mock_search_tool(mock_search_results: list[LlmDoc]) -> MagicMock:
     mock_tool = MagicMock(spec=SearchTool)
     mock_tool.name = "search"
     mock_tool.build_tool_message_content.return_value = "search_response"
@@ -146,7 +126,6 @@ def mock_search_tool(
         json.loads(doc.model_dump_json()) for doc in mock_search_results
     ]
     mock_tool.run.return_value = [
-        ToolResponse(id=SEARCH_DOC_CONTENT_ID, response=mock_contexts),
         ToolResponse(id=FINAL_CONTEXT_DOCUMENTS_ID, response=mock_search_results),
     ]
     mock_tool.tool_definition.return_value = {

backend/tests/unit/onyx/chat/test_answer.py

@@ -19,7 +19,6 @@ from onyx.chat.models import AnswerStyleConfig
 from onyx.chat.models import CitationInfo
 from onyx.chat.models import LlmDoc
 from onyx.chat.models import OnyxAnswerPiece
-from onyx.chat.models import OnyxContexts
 from onyx.chat.models import PromptConfig
 from onyx.chat.models import StreamStopInfo
 from onyx.chat.models import StreamStopReason
@@ -33,7 +32,6 @@ from onyx.tools.force import ForceUseTool
 from onyx.tools.models import ToolCallFinalResult
 from onyx.tools.models import ToolCallKickoff
 from onyx.tools.models import ToolResponse
-from onyx.tools.tool_implementations.search.search_tool import SEARCH_DOC_CONTENT_ID
 from onyx.tools.tool_implementations.search_like_tool_utils import (
     FINAL_CONTEXT_DOCUMENTS_ID,
 )
@@ -141,7 +139,6 @@ def test_basic_answer(answer_instance: Answer, mocker: MockerFixture) -> None:
 def test_answer_with_search_call(
     answer_instance: Answer,
     mock_search_results: list[LlmDoc],
-    mock_contexts: OnyxContexts,
     mock_search_tool: MagicMock,
     force_use_tool: ForceUseTool,
     expected_tool_args: dict,
@@ -197,25 +194,21 @@ def test_answer_with_search_call(
         tool_name="search", tool_args=expected_tool_args
     )
     assert output[1] == ToolResponse(
-        id=SEARCH_DOC_CONTENT_ID,
-        response=mock_contexts,
-    )
-    assert output[2] == ToolResponse(
         id="final_context_documents",
         response=mock_search_results,
     )
-    assert output[3] == ToolCallFinalResult(
+    assert output[2] == ToolCallFinalResult(
         tool_name="search",
         tool_args=expected_tool_args,
         tool_result=[json.loads(doc.model_dump_json()) for doc in mock_search_results],
     )
-    assert output[4] == OnyxAnswerPiece(answer_piece="Based on the search results, ")
+    assert output[3] == OnyxAnswerPiece(answer_piece="Based on the search results, ")
     expected_citation = CitationInfo(citation_num=1, document_id="doc1")
-    assert output[5] == expected_citation
-    assert output[6] == OnyxAnswerPiece(
+    assert output[4] == expected_citation
+    assert output[5] == OnyxAnswerPiece(
         answer_piece="the answer is abc[[1]](https://example.com/doc1). "
     )
-    assert output[7] == OnyxAnswerPiece(answer_piece="This is some other stuff.")
+    assert output[6] == OnyxAnswerPiece(answer_piece="This is some other stuff.")
 
     expected_answer = (
         "Based on the search results, "
@@ -268,7 +261,6 @@ def test_answer_with_search_call(
 def test_answer_with_search_no_tool_calling(
     answer_instance: Answer,
     mock_search_results: list[LlmDoc],
-    mock_contexts: OnyxContexts,
     mock_search_tool: MagicMock,
 ) -> None:
     answer_instance.graph_config.tooling.tools = [mock_search_tool]
@@ -288,30 +280,26 @@ def test_answer_with_search_no_tool_calling(
     output = list(answer_instance.processed_streamed_output)
 
     # Assertions
-    assert len(output) == 8
+    assert len(output) == 7
     assert output[0] == ToolCallKickoff(
         tool_name="search", tool_args=DEFAULT_SEARCH_ARGS
     )
     assert output[1] == ToolResponse(
-        id=SEARCH_DOC_CONTENT_ID,
-        response=mock_contexts,
-    )
-    assert output[2] == ToolResponse(
         id=FINAL_CONTEXT_DOCUMENTS_ID,
         response=mock_search_results,
     )
-    assert output[3] == ToolCallFinalResult(
+    assert output[2] == ToolCallFinalResult(
         tool_name="search",
         tool_args=DEFAULT_SEARCH_ARGS,
         tool_result=[json.loads(doc.model_dump_json()) for doc in mock_search_results],
     )
-    assert output[4] == OnyxAnswerPiece(answer_piece="Based on the search results, ")
+    assert output[3] == OnyxAnswerPiece(answer_piece="Based on the search results, ")
     expected_citation = CitationInfo(citation_num=1, document_id="doc1")
-    assert output[5] == expected_citation
-    assert output[6] == OnyxAnswerPiece(
+    assert output[4] == expected_citation
+    assert output[5] == OnyxAnswerPiece(
         answer_piece="the answer is abc[[1]](https://example.com/doc1). "
     )
-    assert output[7] == OnyxAnswerPiece(answer_piece="This is some other stuff.")
+    assert output[6] == OnyxAnswerPiece(answer_piece="This is some other stuff.")
 
     expected_answer = (
         "Based on the search results, "

backend/tests/unit/onyx/chat/test_skip_gen_ai.py

@@ -79,7 +79,7 @@ def test_skip_gen_ai_answer_generation_flag(
     for res in results:
         print(res)
 
-    expected_count = 4 if skip_gen_ai_answer_generation else 5
+    expected_count = 3 if skip_gen_ai_answer_generation else 4
     assert len(results) == expected_count
     if not skip_gen_ai_answer_generation:
         mock_llm.stream.assert_called_once()

web/src/app/chat/input/AgenticToggle.tsx

@@ -91,7 +91,7 @@ export function AgenticToggle({
         >
           <div className="flex items-center space-x-2 mb-3">
             <h3 className="text-sm font-semibold text-neutral-900">
-              Agent Search (BETA)
+              Agent Search
             </h3>
           </div>
           <p className="text-xs text-neutral-600  dark:text-neutral-700 mb-2">

web/src/app/globals.css

@@ -34,11 +34,12 @@
     /* -------------------------------------------------------
      * 2. Keep special, custom, or near-duplicate background
      * ------------------------------------------------------- */
-    --background: #fefcfa; /* slightly off-white, keep it */
+    --background: #fefcfa; /* slightly off-white */
+    --background-50: #fffdfb; /* a little lighter than background but not quite white */
     --input-background: #fefcfa;
     --input-border: #f1eee8;
     --text-text: #f4f2ed;
-    --background-dark: #e9e6e0;
+    --background-dark: #141414;
     --new-background: #ebe7de;
     --new-background-light: #d9d1c0;
     --background-chatbar: #f5f3ee;
@@ -234,6 +235,7 @@
 
     --text-text: #1d1d1d;
     --background-dark: #252525;
+    --background-50: #252525;
 
     /* --new-background: #fff; */
     --new-background: #2c2c2c;

web/src/components/admin/users/SignedUpUserTable.tsx

@@ -181,7 +181,7 @@ const SignedUpUserTable = ({
                 : "All Roles"}
             </SelectValue>
           </SelectTrigger>
-          <SelectContent className="bg-background">
+          <SelectContent className="bg-background-50">
             {Object.entries(USER_ROLE_LABELS)
               .filter(([role]) => role !== UserRole.EXT_PERM_USER)
               .map(([role, label]) => (

web/src/components/search/DocumentDisplay.tsx

@@ -26,7 +26,13 @@ export const buildDocumentSummaryDisplay = (
   matchHighlights: string[],
   blurb: string
 ) => {
-  if (!matchHighlights || matchHighlights.length === 0) {
+  // if there are no match highlights, or if it's really short, just use the blurb
+  // this is to prevent the UI from showing something like `...` for the summary
+  const MIN_MATCH_HIGHLIGHT_LENGTH = 5;
+  if (
+    !matchHighlights ||
+    matchHighlights.length <= MIN_MATCH_HIGHLIGHT_LENGTH
+  ) {
     return blurb;
   }
 

web/tailwind-themes/tailwind.config.js

@@ -108,6 +108,7 @@ module.exports = {
         "accent-background": "var(--accent-background)",
         "accent-background-hovered": "var(--accent-background-hovered)",
         "accent-background-selected": "var(--accent-background-selected)",
+        "background-50": "var(--background-50)",
         "background-dark": "var(--off-white)",
         "background-100": "var(--neutral-100-border-light)",
         "background-125": "var(--neutral-125)",