k

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

backend/alembic/versions/f13db29f3101_add_composite_index_for_last_modified_.py

@@ -0,0 +1,27 @@
+"""Add composite index for last_modified and last_synced to document
+
+Revision ID: f13db29f3101
+Revises: b388730a2899
+Create Date: 2025-02-18 22:48:11.511389
+
+"""
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision = "f13db29f3101"
+down_revision = "acaab4ef4507"
+branch_labels: str | None = None
+depends_on: str | None = None
+
+
+def upgrade() -> None:
+    op.create_index(
+        "ix_document_sync_status",
+        "document",
+        ["last_modified", "last_synced"],
+        unique=False,
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_document_sync_status", table_name="document")

backend/onyx/auth/email_utils.py

@@ -10,6 +10,7 @@ from onyx.configs.app_configs import SMTP_PORT
 from onyx.configs.app_configs import SMTP_SERVER
 from onyx.configs.app_configs import SMTP_USER
 from onyx.configs.app_configs import WEB_DOMAIN
+from onyx.configs.constants import AuthType
 from onyx.configs.constants import TENANT_ID_COOKIE_NAME
 from onyx.db.models import User
 
@@ -187,23 +188,51 @@ def send_subscription_cancellation_email(user_email: str) -> None:
     send_email(user_email, subject, html_content, text_content)
 
 
-def send_user_email_invite(user_email: str, current_user: User) -> None:
+def send_user_email_invite(
+    user_email: str, current_user: User, auth_type: AuthType
+) -> None:
     subject = "Invitation to Join Onyx Organization"
     heading = "You've Been Invited!"
-    message = (
-        f"<p>You have been invited by {current_user.email} to join an organization on Onyx.</p>"
-        "<p>To join the organization, please click the button below to set a password "
-        "or login with Google and complete your registration.</p>"
-    )
+
+    # the exact action taken by the user, and thus the message, depends on the auth type
+    message = f"<p>You have been invited by {current_user.email} to join an organization on Onyx.</p>"
+    if auth_type == AuthType.CLOUD:
+        message += (
+            "<p>To join the organization, please click the button below to set a password "
+            "or login with Google and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.BASIC:
+        message += (
+            "<p>To join the organization, please click the button below to set a password "
+            "and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.GOOGLE_OAUTH:
+        message += (
+            "<p>To join the organization, please click the button below to login with Google "
+            "and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.OIDC or auth_type == AuthType.SAML:
+        message += (
+            "<p>To join the organization, please click the button below to"
+            " complete your registration.</p>"
+        )
+    else:
+        raise ValueError(f"Invalid auth type: {auth_type}")
+
     cta_text = "Join Organization"
     cta_link = f"{WEB_DOMAIN}/auth/signup?email={user_email}"
     html_content = build_html_email(heading, message, cta_text, cta_link)
+
+    # text content is the fallback for clients that don't support HTML
+    # not as critical, so not having special cases for each auth type
     text_content = (
         f"You have been invited by {current_user.email} to join an organization on Onyx.\n"
         "To join the organization, please visit the following link:\n"
         f"{WEB_DOMAIN}/auth/signup?email={user_email}\n"
-        "You'll be asked to set a password or login with Google to complete your registration."
     )
+    if auth_type == AuthType.CLOUD:
+        text_content += "You'll be asked to set a password or login with Google to complete your registration."
+
     send_email(user_email, subject, html_content, text_content)
 
 

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

@@ -140,7 +140,7 @@ def on_task_postrun(
         f"{f'for tenant_id={tenant_id}' if tenant_id else ''}"
     )
 
-    r = get_redis_client()
+    r = get_redis_client(tenant_id=tenant_id)
 
     if task_id.startswith(RedisConnectorCredentialPair.PREFIX):
         r.srem(RedisConnectorCredentialPair.get_taskset_key(), task_id)

backend/onyx/chat/chat_utils.py

@@ -190,7 +190,8 @@ def create_chat_chain(
             and previous_message.message_type == MessageType.ASSISTANT
             and mainline_messages
         ):
-            mainline_messages[-1] = current_message
+            if current_message.refined_answer_improvement:
+                mainline_messages[-1] = current_message
         else:
             mainline_messages.append(current_message)
 

backend/onyx/chat/models.py

@@ -142,6 +142,15 @@ class MessageResponseIDInfo(BaseModel):
     reserved_assistant_message_id: int
 
 
+class AgentMessageIDInfo(BaseModel):
+    level: int
+    message_id: int
+
+
+class AgenticMessageResponseIDInfo(BaseModel):
+    agentic_message_ids: list[AgentMessageIDInfo]
+
+
 class StreamingError(BaseModel):
     error: str
     stack_trace: str | None = None

backend/onyx/chat/process_message.py

@@ -11,6 +11,8 @@ from onyx.agents.agent_search.orchestration.nodes.call_tool import ToolCallExcep
 from onyx.chat.answer import Answer
 from onyx.chat.chat_utils import create_chat_chain
 from onyx.chat.chat_utils import create_temporary_persona
+from onyx.chat.models import AgenticMessageResponseIDInfo
+from onyx.chat.models import AgentMessageIDInfo
 from onyx.chat.models import AgentSearchPacket
 from onyx.chat.models import AllCitations
 from onyx.chat.models import AnswerPostInfo
@@ -308,6 +310,7 @@ ChatPacket = (
     | CustomToolResponse
     | MessageSpecificCitations
     | MessageResponseIDInfo
+    | AgenticMessageResponseIDInfo
     | StreamStopInfo
     | AgentSearchPacket
 )
@@ -1035,6 +1038,7 @@ def stream_chat_message_objects(
         next_level = 1
         prev_message = gen_ai_response_message
         agent_answers = answer.llm_answer_by_level()
+        agentic_message_ids = []
         while next_level in agent_answers:
             next_answer = agent_answers[next_level]
             info = info_by_subq[
@@ -1059,17 +1063,18 @@ def stream_chat_message_objects(
                 refined_answer_improvement=refined_answer_improvement,
                 is_agentic=True,
             )
+            agentic_message_ids.append(
+                AgentMessageIDInfo(level=next_level, message_id=next_answer_message.id)
+            )
             next_level += 1
             prev_message = next_answer_message
 
         logger.debug("Committing messages")
         db_session.commit()  # actually save user / assistant message
 
-        msg_detail_response = translate_db_message_to_chat_message_detail(
-            gen_ai_response_message
-        )
+        yield AgenticMessageResponseIDInfo(agentic_message_ids=agentic_message_ids)
 
-        yield msg_detail_response
+        yield translate_db_message_to_chat_message_detail(gen_ai_response_message)
     except Exception as e:
         error_msg = str(e)
         logger.exception(error_msg)

backend/onyx/connectors/factory.py

@@ -195,11 +195,16 @@ def validate_ccpair_for_user(
         db_session,
         get_editable=False,
     )
-    if not credential:
-        raise ValueError("Credential not found")
+
     if not connector:
         raise ValueError("Connector not found")
 
+    if connector.source == DocumentSource.INGESTION_API:
+        return
+
+    if not credential:
+        raise ValueError("Credential not found")
+
     try:
         runnable_connector = instantiate_connector(
             db_session=db_session,

backend/onyx/connectors/gitbook/connector.py

@@ -229,16 +229,20 @@ class GitbookConnector(LoadConnector, PollConnector):
 
         try:
             content = self.client.get(f"/spaces/{self.space_id}/content")
-            pages = content.get("pages", [])
-
+            pages: list[dict[str, Any]] = content.get("pages", [])
             current_batch: list[Document] = []
-            for page in pages:
-                updated_at = datetime.fromisoformat(page["updatedAt"])
 
+            while pages:
+                page = pages.pop(0)
+
+                updated_at_raw = page.get("updatedAt")
+                if updated_at_raw is None:
+                    # if updatedAt is not present, that means the page has never been edited
+                    continue
+
+                updated_at = datetime.fromisoformat(updated_at_raw)
                 if start and updated_at < start:
-                    if current_batch:
-                        yield current_batch
-                    return
+                    continue
                 if end and updated_at > end:
                     continue
 
@@ -250,6 +254,8 @@ class GitbookConnector(LoadConnector, PollConnector):
                     yield current_batch
                     current_batch = []
 
+                pages.extend(page.get("pages", []))
+
             if current_batch:
                 yield current_batch
 

backend/onyx/connectors/google_drive/connector.py

@@ -568,6 +568,7 @@ class GoogleDriveConnector(LoadConnector, PollConnector, SlimConnector):
         callback: IndexingHeartbeatInterface | None = None,
     ) -> GenerateSlimDocumentOutput:
         slim_batch = []
+        # ERROR here.
         for file in self._fetch_drive_items(
             is_slim=True,
             start=start,

backend/onyx/db/document.py

@@ -60,9 +60,8 @@ def count_documents_by_needs_sync(session: Session) -> int:
     This function executes the query and returns the count of
     documents matching the criteria."""
 
-    count = (
-        session.query(func.count(DbDocument.id.distinct()))
-        .select_from(DbDocument)
+    return (
+        session.query(DbDocument.id)
         .join(
             DocumentByConnectorCredentialPair,
             DbDocument.id == DocumentByConnectorCredentialPair.id,
@@ -73,63 +72,53 @@ def count_documents_by_needs_sync(session: Session) -> int:
                 DbDocument.last_synced.is_(None),
             )
         )
-        .scalar()
+        .count()
     )
 
-    return count
-
 
 def construct_document_select_for_connector_credential_pair_by_needs_sync(
     connector_id: int, credential_id: int
 ) -> Select:
-    initial_doc_ids_stmt = select(DocumentByConnectorCredentialPair.id).where(
-        and_(
-            DocumentByConnectorCredentialPair.connector_id == connector_id,
-            DocumentByConnectorCredentialPair.credential_id == credential_id,
-        )
-    )
-
-    stmt = (
+    return (
         select(DbDocument)
-        .where(
-            DbDocument.id.in_(initial_doc_ids_stmt),
-            or_(
-                DbDocument.last_modified
-                > DbDocument.last_synced,  # last_modified is newer than last_synced
-                DbDocument.last_synced.is_(None),  # never synced
-            ),
+        .join(
+            DocumentByConnectorCredentialPair,
+            DbDocument.id == DocumentByConnectorCredentialPair.id,
+        )
+        .where(
+            and_(
+                DocumentByConnectorCredentialPair.connector_id == connector_id,
+                DocumentByConnectorCredentialPair.credential_id == credential_id,
+                or_(
+                    DbDocument.last_modified > DbDocument.last_synced,
+                    DbDocument.last_synced.is_(None),
+                ),
+            )
         )
-        .distinct()
     )
 
-    return stmt
-
 
 def construct_document_id_select_for_connector_credential_pair_by_needs_sync(
     connector_id: int, credential_id: int
 ) -> Select:
-    initial_doc_ids_stmt = select(DocumentByConnectorCredentialPair.id).where(
-        and_(
-            DocumentByConnectorCredentialPair.connector_id == connector_id,
-            DocumentByConnectorCredentialPair.credential_id == credential_id,
-        )
-    )
-
-    stmt = (
+    return (
         select(DbDocument.id)
-        .where(
-            DbDocument.id.in_(initial_doc_ids_stmt),
-            or_(
-                DbDocument.last_modified
-                > DbDocument.last_synced,  # last_modified is newer than last_synced
-                DbDocument.last_synced.is_(None),  # never synced
-            ),
+        .join(
+            DocumentByConnectorCredentialPair,
+            DbDocument.id == DocumentByConnectorCredentialPair.id,
+        )
+        .where(
+            and_(
+                DocumentByConnectorCredentialPair.connector_id == connector_id,
+                DocumentByConnectorCredentialPair.credential_id == credential_id,
+                or_(
+                    DbDocument.last_modified > DbDocument.last_synced,
+                    DbDocument.last_synced.is_(None),
+                ),
+            )
         )
-        .distinct()
     )
 
-    return stmt
-
 
 def get_all_documents_needing_vespa_sync_for_cc_pair(
     db_session: Session, cc_pair_id: int

backend/onyx/db/models.py

@@ -570,6 +570,14 @@ class Document(Base):
         back_populates="documents",
     )
 
+    __table_args__ = (
+        Index(
+            "ix_document_sync_status",
+            last_modified,
+            last_synced,
+        ),
+    )
+
 
 class Tag(Base):
     __tablename__ = "tag"

backend/onyx/llm/models.py

@@ -23,6 +23,7 @@ class PreviousMessage(BaseModel):
     message_type: MessageType
     files: list[InMemoryChatFile]
     tool_call: ToolCallFinalResult | None
+    refined_answer_improvement: bool | None
 
     @classmethod
     def from_chat_message(
@@ -47,6 +48,7 @@ class PreviousMessage(BaseModel):
             )
             if chat_message.tool_call
             else None,
+            refined_answer_improvement=chat_message.refined_answer_improvement,
         )
 
     def to_langchain_msg(self) -> BaseMessage:

backend/onyx/main.py

@@ -14,6 +14,7 @@ from fastapi import Request
 from fastapi import status
 from fastapi.exceptions import RequestValidationError
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.openapi.utils import get_openapi
 from fastapi.responses import JSONResponse
 from httpx_oauth.clients.google import GoogleOAuth2
 from sentry_sdk.integrations.fastapi import FastApiIntegration
@@ -28,8 +29,6 @@ from onyx.auth.users import auth_backend
 from onyx.auth.users import create_onyx_oauth_router
 from onyx.auth.users import fastapi_users
 from onyx.configs.app_configs import APP_API_PREFIX
-from onyx.configs.app_configs import APP_HOST
-from onyx.configs.app_configs import APP_PORT
 from onyx.configs.app_configs import AUTH_RATE_LIMITING_ENABLED
 from onyx.configs.app_configs import AUTH_TYPE
 from onyx.configs.app_configs import DISABLE_GENERATIVE_AI
@@ -105,7 +104,6 @@ from onyx.utils.telemetry import get_or_generate_uuid
 from onyx.utils.telemetry import optional_telemetry
 from onyx.utils.telemetry import RecordType
 from onyx.utils.variable_functionality import fetch_versioned_implementation
-from onyx.utils.variable_functionality import global_version
 from onyx.utils.variable_functionality import set_is_ee_based_on_env_variable
 from shared_configs.configs import CORS_ALLOWED_ORIGIN
 from shared_configs.configs import MULTI_TENANT
@@ -263,7 +261,12 @@ def log_http_error(request: Request, exc: Exception) -> JSONResponse:
 
 
 def get_application() -> FastAPI:
-    application = FastAPI(title="Onyx Backend", version=__version__, lifespan=lifespan)
+    application = FastAPI(
+        title="Onyx Backend",
+        version=__version__,
+        openapi_url="/openapi.json",
+        lifespan=lifespan,
+    )
     if SENTRY_DSN:
         sentry_sdk.init(
             dsn=SENTRY_DSN,
@@ -397,24 +400,48 @@ def get_application() -> FastAPI:
     if LOG_ENDPOINT_LATENCY:
         add_latency_logging_middleware(application, logger)
 
-    # Ensure all routes have auth enabled or are explicitly marked as public
+    # # Ensure all routes have auth enabled or are explicitly marked as public
     check_router_auth(application)
 
+    # @application.get("/")
+    # async def root():
+    #     return {"message": "Hello World"}
+
+    def custom_openapi():
+        if application.openapi_schema:
+            return application.openapi_schema
+        openapi_schema = get_openapi(
+            title=application.title,
+            version=application.version,
+            description="Your API description here",
+            routes=application.routes,
+        )
+        # Force OpenAPI version 2.0 instead of 3.1.0
+        application.openapi_schema = openapi_schema
+        return application.openapi_schema
+
+    application.openapi = custom_openapi
+
     return application
 
 
 # NOTE: needs to be outside of the `if __name__ == "__main__"` block so that the
 # app is exportable
 set_is_ee_based_on_env_variable()
+# app = get_application()
 app = fetch_versioned_implementation(module="onyx.main", attribute="get_application")
 
+# from fastapi import FastAPI
+
+# app = FastAPI()
+
+# @app.get("/")
+# async def root():
+#     return {"message": "Hello World"}
+
+# @app.get("/hi")
+# async def docs():
+#     return {"message": "Hello World"}
 
 if __name__ == "__main__":
-    logger.notice(
-        f"Starting Onyx Backend version {__version__} on http://{APP_HOST}:{str(APP_PORT)}/"
-    )
-
-    if global_version.is_ee_version():
-        logger.notice("Running Enterprise Edition")
-
-    uvicorn.run(app, host=APP_HOST, port=APP_PORT)
+    uvicorn.run(app, host="0.0.0.0", port=8000)

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

@@ -1,61 +1,47 @@
 from fastapi import APIRouter
-from fastapi import Depends
-from fastapi import HTTPException
-from fastapi_users.exceptions import InvalidPasswordException
-from sqlalchemy.orm import Session
 
-from onyx.auth.users import current_admin_user
-from onyx.auth.users import current_user
-from onyx.auth.users import get_user_manager
-from onyx.auth.users import User
-from onyx.auth.users import UserManager
-from onyx.db.engine import get_session
-from onyx.db.users import get_user_by_email
-from onyx.server.features.password.models import ChangePasswordRequest
-from onyx.server.features.password.models import UserResetRequest
-from onyx.server.features.password.models import UserResetResponse
 
 router = APIRouter(prefix="/password")
 
 
-@router.post("/change-password")
-async def change_my_password(
-    form_data: ChangePasswordRequest,
-    user_manager: UserManager = Depends(get_user_manager),
-    current_user: User = Depends(current_user),
-) -> None:
-    """
-    Change the password for the current user.
-    """
-    try:
-        await user_manager.change_password_if_old_matches(
-            user=current_user,
-            old_password=form_data.old_password,
-            new_password=form_data.new_password,
-        )
-    except InvalidPasswordException as e:
-        raise HTTPException(status_code=400, detail=str(e.reason))
-    except Exception as e:
-        raise HTTPException(
-            status_code=500, detail=f"An unexpected error occurred: {str(e)}"
-        )
+# @router.post("/change-password")
+# async def change_my_password(
+#     form_data: ChangePasswordRequest,
+#     user_manager: UserManager = Depends(get_user_manager),
+#     current_user: User = Depends(current_user),
+# ) -> None:
+#     """
+#     Change the password for the current user.
+#     """
+#     try:
+#         await user_manager.change_password_if_old_matches(
+#             user=current_user,
+#             old_password=form_data.old_password,
+#             new_password=form_data.new_password,
+#         )
+#     except InvalidPasswordException as e:
+#         raise HTTPException(status_code=400, detail=str(e.reason))
+#     except Exception as e:
+#         raise HTTPException(
+#             status_code=500, detail=f"An unexpected error occurred: {str(e)}"
+#         )
 
 
-@router.post("/reset_password")
-async def admin_reset_user_password(
-    user_reset_request: UserResetRequest,
-    user_manager: UserManager = Depends(get_user_manager),
-    db_session: Session = Depends(get_session),
-    _: User = Depends(current_admin_user),
-) -> UserResetResponse:
-    """
-    Reset the password for a user (admin only).
-    """
-    user = get_user_by_email(user_reset_request.user_email, db_session)
-    if not user:
-        raise HTTPException(status_code=404, detail="User not found")
-    new_password = await user_manager.reset_password_as_admin(user.id)
-    return UserResetResponse(
-        user_id=str(user.id),
-        new_password=new_password,
-    )
+# @router.post("/reset_password")
+# async def admin_reset_user_password(
+#     user_reset_request: UserResetRequest,
+#     user_manager: UserManager = Depends(get_user_manager),
+#     db_session: Session = Depends(get_session),
+#     _: User = Depends(current_admin_user),
+# ) -> UserResetResponse:
+#     """
+#     Reset the password for a user (admin only).
+#     """
+#     user = get_user_by_email(user_reset_request.user_email, db_session)
+#     if not user:
+#         raise HTTPException(status_code=404, detail="User not found")
+#     new_password = await user_manager.reset_password_as_admin(user.id)
+#     return UserResetResponse(
+#         user_id=str(user.id),
+#         new_password=new_password,
+#     )

backend/onyx/server/manage/users.py

@@ -311,19 +311,23 @@ def bulk_invite_users(
     all_emails = list(set(new_invited_emails) | set(initial_invited_users))
     number_of_invited_users = write_invited_users(all_emails)
 
+    # send out email invitations if enabled
+    if ENABLE_EMAIL_INVITES:
+        try:
+            for email in new_invited_emails:
+                send_user_email_invite(email, current_user, AUTH_TYPE)
+        except Exception as e:
+            logger.error(f"Error sending email invite to invited users: {e}")
+
     if not MULTI_TENANT:
         return number_of_invited_users
+
+    # for billing purposes, write to the control plane about the number of new users
     try:
         logger.info("Registering tenant users")
         fetch_ee_implementation_or_noop(
             "onyx.server.tenants.billing", "register_tenant_users", None
         )(tenant_id, get_total_users_count(db_session))
-        if ENABLE_EMAIL_INVITES:
-            try:
-                for email in new_invited_emails:
-                    send_user_email_invite(email, current_user)
-            except Exception as e:
-                logger.error(f"Error sending email invite to invited users: {e}")
 
         return number_of_invited_users
     except Exception as e:

backend/onyx/server/settings/models.py

@@ -45,7 +45,7 @@ class Settings(BaseModel):
     gpu_enabled: bool | None = None
     application_status: ApplicationStatus = ApplicationStatus.ACTIVE
     anonymous_user_enabled: bool | None = None
-    pro_search_disabled: bool | None = None
+    pro_search_enabled: bool | None = None
 
     temperature_override_enabled: bool = False
     auto_scroll: bool = False

backend/requirements/ee.txt

@@ -1,3 +1,4 @@
 cohere==5.6.1
 posthog==3.7.4
 python3-saml==1.15.0
+xmlsec==1.3.14

deployment/data/nginx/app.conf.template

@@ -4,6 +4,24 @@ log_format custom_main '$remote_addr - $remote_user [$time_local] "$request" '
                 '"$http_user_agent" "$http_x_forwarded_for" '
                 'rt=$request_time';
 
+# Map X-Forwarded-Proto or fallback to $scheme
+map $http_x_forwarded_proto $forwarded_proto {
+    default $http_x_forwarded_proto;
+    ""      $scheme;
+}
+
+# Map X-Forwarded-Host or fallback to $host
+map $http_x_forwarded_host $forwarded_host {
+    default $http_x_forwarded_host;
+    ""      $host;
+}
+
+# Map X-Forwarded-Port or fallback to server port
+map $http_x_forwarded_port $forwarded_port {
+    default $http_x_forwarded_port;
+    ""      $server_port;
+}
+
 upstream api_server {
     # fail_timeout=0 means we always retry an upstream even if it failed
     # to return a good HTTP response
@@ -21,8 +39,7 @@ upstream web_server {
 }
 
 server {
-    listen 80;
-    server_name ${DOMAIN};
+    listen 80 default_server;
 
     client_max_body_size 5G;    # Maximum upload size
 
@@ -36,8 +53,9 @@ server {
         # misc headers
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Proto $forwarded_proto;
+        proxy_set_header X-Forwarded-Host $forwarded_host; 
+        proxy_set_header X-Forwarded-Port $forwarded_port;
         proxy_set_header Host $host;
 
         # need to use 1.1 to support chunked transfers
@@ -54,8 +72,9 @@ server {
         # misc headers
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Proto $forwarded_proto;
+        proxy_set_header X-Forwarded-Host $forwarded_host; 
+        proxy_set_header X-Forwarded-Port $forwarded_port;
         proxy_set_header Host $host;
 
         proxy_http_version 1.1;
@@ -72,14 +91,25 @@ server {
 }
 
 server {
-    listen 443 ssl;
-    server_name ${DOMAIN};
+    listen 443 ssl default_server;
 
     client_max_body_size 5G;    # Maximum upload size
     
     location / {
+        # misc headers
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        # don't use forwarded schema, host, or port here - this is the entry point
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Port $server_port;
+        proxy_set_header Host $host;
+
         proxy_http_version 1.1;
         proxy_buffering off;
+        # we don't want nginx trying to do something clever with
+        # redirects, we set the Host: header above already.
+        proxy_redirect off;
         proxy_pass http://localhost:80;
     }
 

deployment/data/nginx/app.conf.template.dev

@@ -21,8 +21,7 @@ upstream web_server {
 }
 
 server {
-    listen 80;
-    server_name ${DOMAIN};
+    listen 80 default_server;
 
     client_max_body_size 5G;    # Maximum upload size    
 
@@ -37,7 +36,8 @@ server {
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
         proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
         proxy_set_header Host $host;
 
         # need to use 1.1 to support chunked transfers
@@ -55,7 +55,8 @@ server {
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
         proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
         proxy_set_header Host $host;
 
         proxy_http_version 1.1;

deployment/data/nginx/app.conf.template.no-letsencrypt

@@ -4,6 +4,24 @@ log_format custom_main '$remote_addr - $remote_user [$time_local] "$request" '
                 '"$http_user_agent" "$http_x_forwarded_for" '
                 'rt=$request_time';
 
+# Map X-Forwarded-Proto or fallback to $scheme
+map $http_x_forwarded_proto $forwarded_proto {
+    default $http_x_forwarded_proto;
+    ""      $scheme;
+}
+
+# Map X-Forwarded-Host or fallback to $host
+map $http_x_forwarded_host $forwarded_host {
+    default $http_x_forwarded_host;
+    ""      $host;
+}
+
+# Map X-Forwarded-Port or fallback to server port
+map $http_x_forwarded_port $forwarded_port {
+    default $http_x_forwarded_port;
+    ""      $server_port;
+}
+
 upstream api_server {
     # fail_timeout=0 means we always retry an upstream even if it failed
     # to return a good HTTP response
@@ -21,8 +39,7 @@ upstream web_server {
 }
 
 server {
-    listen 80;
-    server_name ${DOMAIN};
+    listen 80 default_server;
 
     client_max_body_size 5G;    # Maximum upload size
 
@@ -36,8 +53,9 @@ server {
         # misc headers
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Proto $forwarded_proto;
+        proxy_set_header X-Forwarded-Host $forwarded_host; 
+        proxy_set_header X-Forwarded-Port $forwarded_port;
         proxy_set_header Host $host;
 
         # need to use 1.1 to support chunked transfers
@@ -54,8 +72,9 @@ server {
         # misc headers
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Proto $forwarded_proto;
+        proxy_set_header X-Forwarded-Host $forwarded_host; 
+        proxy_set_header X-Forwarded-Port $forwarded_port;
         proxy_set_header Host $host;
 
         proxy_http_version 1.1;
@@ -68,14 +87,25 @@ server {
 }
 
 server {
-    listen 443 ssl;
-    server_name ${DOMAIN};
+    listen 443 ssl default_server;
 
     client_max_body_size 5G;    # Maximum upload size
     
     location / {
+        # misc headers
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        # don't use forwarded schema, host, or port here - this is the entry point
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host; 
+        proxy_set_header X-Forwarded-Port $server_port;
+        proxy_set_header Host $host;
+
         proxy_http_version 1.1;
         proxy_buffering off;
+        # we don't want nginx trying to do something clever with
+        # redirects, we set the Host: header above already.
+        proxy_redirect off;
         proxy_pass http://localhost:80;
     }
 

deployment/data/nginx/run-nginx.sh

@@ -1,5 +1,5 @@
 # fill in the template
-envsubst '$DOMAIN $SSL_CERT_FILE_NAME $SSL_CERT_KEY_FILE_NAME' < "/etc/nginx/conf.d/$1" > /etc/nginx/conf.d/app.conf
+envsubst '$SSL_CERT_FILE_NAME $SSL_CERT_KEY_FILE_NAME' < "/etc/nginx/conf.d/$1" > /etc/nginx/conf.d/app.conf
 
 # wait for the api_server to be ready
 echo "Waiting for API server to boot up; this may take a minute or two..."

web/next.config.js

@@ -68,6 +68,22 @@ const nextConfig = {
       },
     ];
   },
+  async rewrites() {
+    return [
+      {
+        source: "/api/docs/:path*", // catch /api/docs and /api/docs/...
+        destination: "http://localhost:8000/docs/:path*",
+      },
+      {
+        source: "/api/docs", // if you also need the exact /api/docs
+        destination: "http://localhost:8000/docs",
+      },
+      {
+        source: "/openapi.json",
+        destination: "http://localhost:8000/openapi.json",
+      },
+    ];
+  },
 };
 
 // Sentry configuration for error monitoring:

web/package-lock.json

@@ -75,6 +75,7 @@
         "semver": "^7.5.4",
         "sharp": "^0.33.5",
         "stripe": "^17.0.0",
+        "swagger-ui-dist": "^5.19.0",
         "swr": "^2.1.5",
         "tailwind-merge": "^2.5.4",
         "tailwindcss": "^3.3.1",
@@ -6180,6 +6181,13 @@
       "integrity": "sha512-qC/xYId4NMebE6w/V33Fh9gWxLgURiNYgVNObbJl2LZv0GUUItCcCqC5axQSwRaAgaxl2mELq1rMzlswaQ0Zxg==",
       "dev": true
     },
+    "node_modules/@scarf/scarf": {
+      "version": "1.4.0",
+      "resolved": "https://registry.npmjs.org/@scarf/scarf/-/scarf-1.4.0.tgz",
+      "integrity": "sha512-xxeapPiUXdZAE3che6f3xogoJPeZgig6omHEy1rIY5WVsB3H2BHNnZH+gHG6x91SCWyQCzWGsuL2Hh3ClO5/qQ==",
+      "hasInstallScript": true,
+      "license": "Apache-2.0"
+    },
     "node_modules/@segment/analytics-core": {
       "version": "1.4.1",
       "resolved": "https://registry.npmjs.org/@segment/analytics-core/-/analytics-core-1.4.1.tgz",
@@ -20122,6 +20130,15 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
+    "node_modules/swagger-ui-dist": {
+      "version": "5.19.0",
+      "resolved": "https://registry.npmjs.org/swagger-ui-dist/-/swagger-ui-dist-5.19.0.tgz",
+      "integrity": "sha512-bSVZeYaqanMFeW5ZY3+EejFbsjkjazYxm1I7Lz3xayYz5XU3m2aUzvuPC0jI95WCQdduszHYV3ER4buQoy8DXA==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@scarf/scarf": "=1.4.0"
+      }
+    },
     "node_modules/swr": {
       "version": "2.2.5",
       "resolved": "https://registry.npmjs.org/swr/-/swr-2.2.5.tgz",

web/package.json

@@ -78,6 +78,7 @@
     "semver": "^7.5.4",
     "sharp": "^0.33.5",
     "stripe": "^17.0.0",
+    "swagger-ui-dist": "^5.19.0",
     "swr": "^2.1.5",
     "tailwind-merge": "^2.5.4",
     "tailwindcss": "^3.3.1",

web/src/app/admin/settings/SettingsForm.tsx

@@ -240,11 +240,11 @@ export function SettingsForm() {
       />
 
       <Checkbox
-        label="Pro Search Disabled"
-        sublabel="If set, users will not be able to use Pro Search."
-        checked={settings.pro_search_disabled ?? false}
+        label="Agent Search"
+        sublabel="If set, users will be able to use Agent Search."
+        checked={settings.pro_search_enabled ?? true}
         onChange={(e) =>
-          handleToggleSettingsField("pro_search_disabled", e.target.checked)
+          handleToggleSettingsField("pro_search_enabled", e.target.checked)
         }
       />
 

web/src/app/admin/settings/interfaces.ts

@@ -10,7 +10,7 @@ export interface Settings {
   notifications: Notification[];
   needs_reindexing: boolean;
   gpu_enabled: boolean;
-  pro_search_disabled: boolean | null;
+  pro_search_enabled: boolean | null;
   application_status: ApplicationStatus;
   auto_scroll: boolean;
   temperature_override_enabled: boolean;

web/src/app/api/[...path]/route.ts

@@ -61,20 +61,20 @@ export async function OPTIONS(
 }
 
 async function handleRequest(request: NextRequest, path: string[]) {
-  if (
-    process.env.NODE_ENV !== "development" &&
-    // NOTE: Set this environment variable to 'true' for preview environments
-    // Where you want finer-grained control over API access
-    process.env.OVERRIDE_API_PRODUCTION !== "true"
-  ) {
-    return NextResponse.json(
-      {
-        message:
-          "This API is only available in development mode. In production, something else (e.g. nginx) should handle this.",
-      },
-      { status: 404 }
-    );
-  }
+  // if (
+  //   process.env.NODE_ENV !== "development" &&
+  //   // NOTE: Set this environment variable to 'true' for preview environments
+  //   // Where you want finer-grained control over API access
+  //   process.env.OVERRIDE_API_PRODUCTION !== "true"
+  // ) {
+  //   return NextResponse.json(
+  //     {
+  //       message:
+  //         "This API is only available in development mode. In production, something else (e.g. nginx) should handle this.",
+  //     },
+  //     { status: 404 }
+  //   );
+  // }
 
   try {
     const backendUrl = new URL(`${INTERNAL_URL}/${path.join("/")}`);
@@ -87,14 +87,17 @@ async function handleRequest(request: NextRequest, path: string[]) {
       backendUrl.searchParams.append(key, value);
     });
 
-    const response = await fetch(backendUrl, {
+    const fetchDetails = {
       method: request.method,
       headers: request.headers,
       body: request.body,
       signal: request.signal,
       // @ts-ignore
       duplex: "half",
-    });
+    };
+    console.log(fetchDetails);
+    console.log("BACKEND URL: ", backendUrl);
+    const response = await fetch(backendUrl, fetchDetails);
 
     // Check if the response is a stream
     if (

web/src/app/chat/ChatPage.tsx

@@ -23,6 +23,7 @@ import {
   SubQuestionDetail,
   constructSubQuestions,
   DocumentsResponse,
+  AgenticMessageResponseIDInfo,
 } from "./interfaces";
 
 import Prism from "prismjs";
@@ -1280,6 +1281,8 @@ export function ChatPage({
     let toolCall: ToolCallMetadata | null = null;
     let isImprovement: boolean | undefined = undefined;
     let isStreamingQuestions = true;
+    let includeAgentic = false;
+    let secondLevelMessageId: number | null = null;
 
     let initialFetchDetails: null | {
       user_message_id: number;
@@ -1336,7 +1339,7 @@ export function ChatPage({
           searchParams.get(SEARCH_PARAM_NAMES.SYSTEM_PROMPT) || undefined,
         useExistingUserMessage: isSeededChat,
         useLanggraph:
-          !settings?.settings.pro_search_disabled &&
+          settings?.settings.pro_search_enabled &&
           proSearchEnabled &&
           retrievalEnabled,
       });
@@ -1417,6 +1420,17 @@ export function ChatPage({
             resetRegenerationState();
           } else {
             const { user_message_id, frozenMessageMap } = initialFetchDetails;
+            if (Object.hasOwn(packet, "agentic_message_ids")) {
+              const agenticMessageIds = (packet as AgenticMessageResponseIDInfo)
+                .agentic_message_ids;
+              const level1MessageId = agenticMessageIds.find(
+                (item) => item.level === 1
+              )?.message_id;
+              if (level1MessageId) {
+                secondLevelMessageId = level1MessageId;
+                includeAgentic = true;
+              }
+            }
 
             setChatState((prevState) => {
               if (prevState.get(chatSessionIdRef.current!) === "loading") {
@@ -1568,7 +1582,10 @@ export function ChatPage({
                   };
                 }
               );
-            } else if (Object.hasOwn(packet, "error")) {
+            } else if (
+              Object.hasOwn(packet, "error") &&
+              (packet as any).error != null
+            ) {
               if (
                 sub_questions.length > 0 &&
                 sub_questions
@@ -1580,8 +1597,8 @@ export function ChatPage({
                 setAgenticGenerating(false);
                 setAlternativeGeneratingAssistant(null);
                 setSubmittedMessage("");
-                return;
-                // throw new Error((packet as StreamingError).error);
+
+                throw new Error((packet as StreamingError).error);
               } else {
                 error = (packet as StreamingError).error;
                 stackTrace = (packet as StreamingError).stack_trace;
@@ -1664,6 +1681,19 @@ export function ChatPage({
                 second_level_generating: second_level_generating,
                 agentic_docs: agenticDocs,
               },
+              ...(includeAgentic
+                ? [
+                    {
+                      messageId: secondLevelMessageId!,
+                      message: second_level_answer,
+                      type: "assistant" as const,
+                      files: [],
+                      toolCall: null,
+                      parentMessageId:
+                        initialFetchDetails.assistant_message_id!,
+                    },
+                  ]
+                : []),
             ]);
           }
         }
@@ -2692,6 +2722,11 @@ export function ChatPage({
                                     ? messageHistory[i + 1]?.documents
                                     : undefined;
 
+                                const nextMessage =
+                                  messageHistory[i + 1]?.type === "assistant"
+                                    ? messageHistory[i + 1]
+                                    : undefined;
+
                                 return (
                                   <div
                                     className="text-text"
@@ -2720,7 +2755,10 @@ export function ChatPage({
                                             selectedMessageForDocDisplay ==
                                               secondLevelMessage?.messageId)
                                         }
-                                        isImprovement={message.isImprovement}
+                                        isImprovement={
+                                          message.isImprovement ||
+                                          nextMessage?.isImprovement
+                                        }
                                         secondLevelGenerating={
                                           (message.second_level_generating &&
                                             currentSessionChatState !==

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

@@ -805,13 +805,12 @@ export function ChatInputBar({
                 )}
               </div>
               <div className="flex items-center my-auto">
-                {retrievalEnabled &&
-                  !settings?.settings.pro_search_disabled && (
-                    <AgenticToggle
-                      proSearchEnabled={proSearchEnabled}
-                      setProSearchEnabled={setProSearchEnabled}
-                    />
-                  )}
+                {retrievalEnabled && settings?.settings.pro_search_enabled && (
+                  <AgenticToggle
+                    proSearchEnabled={proSearchEnabled}
+                    setProSearchEnabled={setProSearchEnabled}
+                  />
+                )}
                 <button
                   id="onyx-chat-input-send-button"
                   className={`cursor-pointer ${

web/src/app/chat/interfaces.ts

@@ -155,6 +155,15 @@ export interface MessageResponseIDInfo {
   reserved_assistant_message_id: number;
 }
 
+export interface AgentMessageIDInfo {
+  level: number;
+  message_id: number;
+}
+
+export interface AgenticMessageResponseIDInfo {
+  agentic_message_ids: AgentMessageIDInfo[];
+}
+
 export interface DocumentsResponse {
   top_documents: OnyxDocument[];
   rephrased_query: string | null;

web/src/app/chat/lib.tsx

@@ -25,6 +25,7 @@ import {
   RetrievalType,
   StreamingError,
   ToolCallMetadata,
+  AgenticMessageResponseIDInfo,
 } from "./interfaces";
 import { Persona } from "../admin/assistants/interfaces";
 import { ReadonlyURLSearchParams } from "next/navigation";
@@ -154,7 +155,8 @@ export type PacketType =
   | AgentAnswerPiece
   | SubQuestionPiece
   | ExtendedToolResponse
-  | RefinedAnswerImprovement;
+  | RefinedAnswerImprovement
+  | AgenticMessageResponseIDInfo;
 
 export async function* sendMessage({
   regenerate,

web/src/app/layout.tsx

@@ -21,11 +21,9 @@ import { fetchAssistantData } from "@/lib/chat/fetchAssistantdata";
 import { AppProvider } from "@/components/context/AppProvider";
 import { PHProvider } from "./providers";
 import { getCurrentUserSS } from "@/lib/userSS";
-import CardSection from "@/components/admin/CardSection";
 import { Suspense } from "react";
 import PostHogPageView from "./PostHogPageView";
 import Script from "next/script";
-import { LogoType } from "@/components/logo/Logo";
 import { Hanken_Grotesk } from "next/font/google";
 import { WebVitals } from "./web-vitals";
 import { ThemeProvider } from "next-themes";

web/src/components/settings/lib.ts

@@ -51,7 +51,7 @@ export async function fetchSettingsSS(): Promise<CombinedSettings | null> {
           notifications: [],
           needs_reindexing: false,
           anonymous_user_enabled: false,
-          pro_search_disabled: false,
+          pro_search_enabled: true,
           temperature_override_enabled: true,
         };
       } else {
@@ -95,8 +95,8 @@ export async function fetchSettingsSS(): Promise<CombinedSettings | null> {
       }
     }
 
-    if (enterpriseSettings && settings.pro_search_disabled == null) {
-      settings.pro_search_disabled = true;
+    if (settings.pro_search_enabled == null) {
+      settings.pro_search_enabled = true;
     }
 
     const webVersion = getWebVersion();