first cut at anonymizing query history (#4123)

Co-authored-by: Richard Kuo <rkuo@rkuo.com>
This commit is contained in:
rkuo-danswer
2025-02-26 13:32:01 -08:00
committed by GitHub
parent ad0721ecd8
commit 25389c5120
8 changed files with 91 additions and 16 deletions

View File

@@ -2,6 +2,7 @@ import csv
import io
from datetime import datetime
from datetime import timezone
from http import HTTPStatus
from uuid import UUID
from fastapi import APIRouter
@@ -21,8 +22,10 @@ from ee.onyx.server.query_history.models import QuestionAnswerPairSnapshot
from onyx.auth.users import current_admin_user
from onyx.auth.users import get_display_email
from onyx.chat.chat_utils import create_chat_chain
from onyx.configs.app_configs import ONYX_QUERY_HISTORY_TYPE
from onyx.configs.constants import MessageType
from onyx.configs.constants import QAFeedbackType
from onyx.configs.constants import QueryHistoryType
from onyx.configs.constants import SessionType
from onyx.db.chat import get_chat_session_by_id
from onyx.db.chat import get_chat_sessions_by_user
@@ -35,6 +38,8 @@ from onyx.server.query_and_chat.models import ChatSessionsResponse
router = APIRouter()
ONYX_ANONYMIZED_EMAIL = "anonymous@anonymous.invalid"
def fetch_and_process_chat_session_history(
db_session: Session,
@@ -107,6 +112,17 @@ def get_user_chat_sessions(
_: User | None = Depends(current_admin_user),
db_session: Session = Depends(get_session),
) -> ChatSessionsResponse:
# we specifically don't allow this endpoint if "anonymized" since
# this is a direct query on the user id
if ONYX_QUERY_HISTORY_TYPE in [
QueryHistoryType.DISABLED,
QueryHistoryType.ANONYMIZED,
]:
raise HTTPException(
status_code=HTTPStatus.FORBIDDEN,
detail="Per user query history has been disabled by the administrator.",
)
try:
chat_sessions = get_chat_sessions_by_user(
user_id=user_id, deleted=False, db_session=db_session, limit=0
@@ -141,6 +157,12 @@ def get_chat_session_history(
_: User | None = Depends(current_admin_user),
db_session: Session = Depends(get_session),
) -> PaginatedReturn[ChatSessionMinimal]:
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
raise HTTPException(
status_code=HTTPStatus.FORBIDDEN,
detail="Query history has been disabled by the administrator.",
)
page_of_chat_sessions = get_page_of_chat_sessions(
page_num=page_num,
page_size=page_size,
@@ -157,11 +179,16 @@ def get_chat_session_history(
feedback_filter=feedback_type,
)
minimal_chat_sessions: list[ChatSessionMinimal] = []
for chat_session in page_of_chat_sessions:
minimal_chat_session = ChatSessionMinimal.from_chat_session(chat_session)
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
minimal_chat_session.user_email = ONYX_ANONYMIZED_EMAIL
minimal_chat_sessions.append(minimal_chat_session)
return PaginatedReturn(
items=[
ChatSessionMinimal.from_chat_session(chat_session)
for chat_session in page_of_chat_sessions
],
items=minimal_chat_sessions,
total_items=total_filtered_chat_sessions_count,
)
@@ -172,6 +199,12 @@ def get_chat_session_admin(
_: User | None = Depends(current_admin_user),
db_session: Session = Depends(get_session),
) -> ChatSessionSnapshot:
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
raise HTTPException(
status_code=HTTPStatus.FORBIDDEN,
detail="Query history has been disabled by the administrator.",
)
try:
chat_session = get_chat_session_by_id(
chat_session_id=chat_session_id,
@@ -193,6 +226,9 @@ def get_chat_session_admin(
f"Could not create snapshot for chat session with id '{chat_session_id}'",
)
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
snapshot.user_email = ONYX_ANONYMIZED_EMAIL
return snapshot
@@ -203,6 +239,12 @@ def get_query_history_as_csv(
end: datetime | None = None,
db_session: Session = Depends(get_session),
) -> StreamingResponse:
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
raise HTTPException(
status_code=HTTPStatus.FORBIDDEN,
detail="Query history has been disabled by the administrator.",
)
complete_chat_session_history = fetch_and_process_chat_session_history(
db_session=db_session,
start=start or datetime.fromtimestamp(0, tz=timezone.utc),
@@ -213,6 +255,9 @@ def get_query_history_as_csv(
question_answer_pairs: list[QuestionAnswerPairSnapshot] = []
for chat_session_snapshot in complete_chat_session_history:
if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
chat_session_snapshot.user_email = ONYX_ANONYMIZED_EMAIL
question_answer_pairs.extend(
QuestionAnswerPairSnapshot.from_chat_session_snapshot(chat_session_snapshot)
)