Improvements to Query History (#17)

* Add option to download query-history as a CSV

* Add user email + more complete timestamp

backend/ee/danswer/db/query_history.py

@@ -4,27 +4,29 @@ from typing import cast
 from typing import Literal
 
 from sqlalchemy import or_
+from sqlalchemy import Select
 from sqlalchemy import select
+from sqlalchemy.orm import aliased
 from sqlalchemy.orm import Session
 from sqlalchemy.orm.attributes import InstrumentedAttribute
 
 from danswer.configs.constants import QAFeedbackType
 from danswer.db.models import QueryEvent
+from danswer.db.models import User
 
 SortByOptions = Literal["time_created", "feedback"]
 
 
-def fetch_query_history(
+def build_query_history_query(
-    db_session: Session,
     start: datetime.datetime,
     end: datetime.datetime,
-    query: str | None = None,
+    query: str | None,
-    feedback_type: QAFeedbackType | None = None,
+    feedback_type: QAFeedbackType | None,
-    sort_by_field: SortByOptions = "time_created",
+    sort_by_field: SortByOptions,
-    sort_by_direction: Literal["asc", "desc"] = "desc",
+    sort_by_direction: Literal["asc", "desc"],
-    offset: int = 0,
+    offset: int,
-    limit: int = 500,
+    limit: int | None,
-) -> Sequence[QueryEvent]:
+) -> Select[tuple[QueryEvent]]:
     stmt = (
         select(QueryEvent)
         .where(
@@ -41,7 +43,10 @@ def fetch_query_history(
     else:
         stmt = stmt.order_by(order_by_field.desc())
 
-    stmt = stmt.offset(offset).limit(limit)
+    if offset:
+        stmt = stmt.offset(offset)
+    if limit:
+        stmt = stmt.limit(limit)
 
     if query:
         stmt = stmt.where(
@@ -54,4 +59,59 @@ def fetch_query_history(
     if feedback_type:
         stmt = stmt.where(QueryEvent.feedback == feedback_type)
 
+    return stmt
+
+
+def fetch_query_history(
+    db_session: Session,
+    start: datetime.datetime,
+    end: datetime.datetime,
+    query: str | None = None,
+    feedback_type: QAFeedbackType | None = None,
+    sort_by_field: SortByOptions = "time_created",
+    sort_by_direction: Literal["asc", "desc"] = "desc",
+    offset: int = 0,
+    limit: int | None = 500,
+) -> Sequence[QueryEvent]:
+    stmt = build_query_history_query(
+        start=start,
+        end=end,
+        query=query,
+        feedback_type=feedback_type,
+        sort_by_field=sort_by_field,
+        sort_by_direction=sort_by_direction,
+        offset=offset,
+        limit=limit,
+    )
+
     return db_session.scalars(stmt).all()
+
+
+def fetch_query_history_with_user_email(
+    db_session: Session,
+    start: datetime.datetime,
+    end: datetime.datetime,
+    query: str | None = None,
+    feedback_type: QAFeedbackType | None = None,
+    sort_by_field: SortByOptions = "time_created",
+    sort_by_direction: Literal["asc", "desc"] = "desc",
+    offset: int = 0,
+    limit: int | None = 500,
+) -> Sequence[tuple[QueryEvent, str | None]]:
+    subquery = build_query_history_query(
+        start=start,
+        end=end,
+        query=query,
+        feedback_type=feedback_type,
+        sort_by_field=sort_by_field,
+        sort_by_direction=sort_by_direction,
+        offset=offset,
+        limit=limit,
+    ).subquery()
+    subquery_alias = aliased(QueryEvent, subquery)
+
+    stmt_with_user_email = select(subquery_alias, User.email).join(  # type: ignore
+        User, subquery_alias.user_id == User.id, isouter=True
+    )
+
+    return db_session.execute(stmt_with_user_email).all()  # type: ignore

backend/ee/danswer/server/query_history/api.py

@@ -1,10 +1,14 @@
+import csv
+import io
 from collections.abc import Iterable
 from datetime import datetime
 from datetime import timedelta
+from datetime import timezone
 
 from fastapi import APIRouter
 from fastapi import Depends
 from fastapi import HTTPException
+from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
 from sqlalchemy.orm import Session
 
@@ -15,7 +19,9 @@ from danswer.db.engine import get_session
 from danswer.db.feedback import fetch_query_event_by_id
 from danswer.db.models import Document
 from ee.danswer.db.document import fetch_documents_from_ids
-from ee.danswer.db.query_history import fetch_query_history
+from ee.danswer.db.query_history import (
+    fetch_query_history_with_user_email,
+)
 
 
 router = APIRouter()
@@ -31,6 +37,7 @@ class AbridgedSearchDoc(BaseModel):
 
 class QuerySnapshot(BaseModel):
     id: int
+    user_email: str | None
     query: str
     llm_answer: str | None
     retrieved_documents: list[AbridgedSearchDoc]
@@ -39,10 +46,14 @@ class QuerySnapshot(BaseModel):
 
     @classmethod
     def build(
-        cls, query_event: db_models.QueryEvent, documents: Iterable[Document]
+        cls,
+        query_event: db_models.QueryEvent,
+        user_email: str | None,
+        documents: Iterable[Document],
     ) -> "QuerySnapshot":
         return cls(
             id=query_event.id,
+            user_email=user_email,
             query=query_event.query,
             llm_answer=query_event.llm_answer,
             retrieved_documents=[
@@ -57,25 +68,43 @@ class QuerySnapshot(BaseModel):
             time_created=query_event.time_created,
         )
 
+    def to_json(self) -> dict[str, str]:
+        return {
+            "id": str(self.id),
+            "query": self.query,
+            "user_email": self.user_email or "",
+            "llm_answer": self.llm_answer or "",
+            "retrieved_documents": "|".join(
+                [
+                    doc.link or doc.semantic_identifier
+                    for doc in self.retrieved_documents
+                ]
+            ),
+            "feedback": self.feedback.value if self.feedback else "",
+            "time_created": str(self.time_created),
+        }
 
-@router.get("/admin/query-history")
+
-def get_query_history(
+def fetch_and_process_query_history(
-    feedback_type: QAFeedbackType | None = None,
+    db_session: Session,
-    start: datetime | None = None,
+    start: datetime | None,
-    end: datetime | None = None,
+    end: datetime | None,
-    _: db_models.User | None = Depends(current_admin_user),
+    feedback_type: QAFeedbackType | None,
-    db_session: Session = Depends(get_session),
+    limit: int | None = 500,
 ) -> list[QuerySnapshot]:
-    query_history = fetch_query_history(
+    query_history_with_user_email = fetch_query_history_with_user_email(
         db_session=db_session,
         start=start
-        or (datetime.utcnow() - timedelta(days=30)),  # default is 30d lookback
+        or (
-        end=end or datetime.utcnow(),
+            datetime.now(tz=timezone.utc) - timedelta(days=30)
+        ),  # default is 30d lookback
+        end=end or datetime.now(tz=timezone.utc),
         feedback_type=feedback_type,
+        limit=limit,
     )
 
     all_relevant_document_ids: set[str] = set()
-    for query_event in query_history:
+    for query_event, _ in query_history_with_user_email:
         all_relevant_document_ids = all_relevant_document_ids.union(
             query_event.retrieved_document_ids or []
         )
@@ -87,7 +116,7 @@ def get_query_history(
     }
 
     query_snapshots: list[QuerySnapshot] = []
-    for query_event in query_history:
+    for query_event, user_email in query_history_with_user_email:
         unique_document_ids = set(query_event.retrieved_document_ids or [])
         documents = [
             document_id_to_document[doc_id]
@@ -95,11 +124,29 @@ def get_query_history(
             if doc_id in document_id_to_document
         ]
         query_snapshots.append(
-            QuerySnapshot.build(query_event=query_event, documents=documents)
+            QuerySnapshot.build(
+                query_event=query_event, user_email=user_email, documents=documents
+            )
         )
     return query_snapshots
 
 
+@router.get("/admin/query-history")
+def get_query_history(
+    feedback_type: QAFeedbackType | None = None,
+    start: datetime | None = None,
+    end: datetime | None = None,
+    _: db_models.User | None = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+) -> list[QuerySnapshot]:
+    return fetch_and_process_query_history(
+        db_session=db_session,
+        start=start,
+        end=end,
+        feedback_type=feedback_type,
+    )
+
+
 @router.get("/admin/query-history/{query_id}")
 def get_query(
     query_id: int,
@@ -113,4 +160,40 @@ def get_query(
     documents = fetch_documents_from_ids(
         db_session, query_event.retrieved_document_ids or []
     )
-    return QuerySnapshot.build(query_event=query_event, documents=documents)
+    return QuerySnapshot.build(
+        query_event=query_event,
+        user_email=query_event.user.email if query_event.user else None,
+        documents=documents,
+    )
+
+
+@router.get("/admin/query-history-csv")
+def get_query_history_as_csv(
+    _: db_models.User | None = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+) -> StreamingResponse:
+    complete_query_history = fetch_and_process_query_history(
+        db_session=db_session,
+        start=datetime.fromtimestamp(0, tz=timezone.utc),
+        end=datetime.now(tz=timezone.utc),
+        feedback_type=None,
+        limit=None,
+    )
+
+    # Create an in-memory text stream
+    stream = io.StringIO()
+    writer = csv.DictWriter(stream, fieldnames=list(QuerySnapshot.__fields__.keys()))
+    writer.writeheader()
+    for row in complete_query_history:
+        writer.writerow(row.to_json())
+
+    # Reset the stream's position to the start
+    stream.seek(0)
+
+    return StreamingResponse(
+        iter([stream.getvalue()]),
+        media_type="text/csv",
+        headers={
+            "Content-Disposition": "attachment;filename=danswer_query_history.csv"
+        },
+    )

web/src/app/ee/admin/performance/analytics/types.ts

@@ -21,6 +21,7 @@ export interface AbridgedSearchDoc {
 export interface QuerySnapshot {
   id: number;
   query: string;
+  user_email: string | null;
   llm_answer: string;
   retrieved_documents: AbridgedSearchDoc[];
   time_created: string;

web/src/app/ee/admin/performance/query-history/DownloadAsCSV.tsx

@@ -0,0 +1,13 @@
+import { FiDownload } from "react-icons/fi";
+
+export function DownloadAsCSV() {
+  return (
+    <a
+      href="/api/admin/query-history-csv"
+      className="text-gray-300 flex ml-auto py-2 px-4 border border-gray-800 h-fit cursor-pointer hover:bg-gray-800 text-sm"
+    >
+      <FiDownload className="my-auto mr-2" />
+      Download as CSV
+    </a>
+  );
+}

web/src/app/ee/admin/performance/query-history/QueryHistoryTable.tsx

@@ -14,7 +14,10 @@ import { Divider } from "@tremor/react";
 import { Select, SelectItem } from "@tremor/react";
 import { ThreeDotsLoader } from "@/components/Loading";
 import { QuerySnapshot } from "../analytics/types";
-import { timestampToDateString } from "@/lib/dateUtils";
+import {
+  timestampToDateString,
+  timestampToReadableDate,
+} from "@/lib/dateUtils";
 import { FiBook, FiFrown, FiMinus, FiSmile } from "react-icons/fi";
 import { useState } from "react";
 import { Feedback } from "@/lib/types";
@@ -22,6 +25,7 @@ import { DateRangeSelector } from "../DateRangeSelector";
 import { PageSelector } from "@/components/PageSelector";
 import Link from "next/link";
 import { FeedbackBadge } from "./FeedbackBadge";
+import { DownloadAsCSV } from "./DownloadAsCSV";
 
 const NUM_IN_PAGE = 20;
 
@@ -54,7 +58,10 @@ function QueryHistoryTableRow({
       <TableCell>
         <FeedbackBadge feedback={querySnapshot.feedback} />
       </TableCell>
-      <TableCell>{timestampToDateString(querySnapshot.time_created)}</TableCell>
+      <TableCell>{querySnapshot.user_email || "-"}</TableCell>
+      <TableCell>
+        {timestampToReadableDate(querySnapshot.time_created)}
+      </TableCell>
       {/* Wrapping in <td> to avoid console warnings */}
       <td className="w-0 p-0">
         <Link
@@ -114,13 +121,20 @@ export function QueryHistoryTable() {
     <Card className="mt-8">
       {queryHistoryData ? (
         <>
-          <div className="gap-y-3 flex flex-col">
+          <div className="flex">
-            <SelectFeedbackType
+            <div className="gap-y-3 flex flex-col">
-              value={selectedFeedbackType || "all"}
+              <SelectFeedbackType
-              onValueChange={setSelectedFeedbackType}
+                value={selectedFeedbackType || "all"}
-            />
+                onValueChange={setSelectedFeedbackType}
+              />
 
-            <DateRangeSelector value={timeRange} onValueChange={setTimeRange} />
+              <DateRangeSelector
+                value={timeRange}
+                onValueChange={setTimeRange}
+              />
+            </div>
+
+            <DownloadAsCSV />
           </div>
           <Divider />
           <Table className="mt-5">
@@ -130,6 +144,7 @@ export function QueryHistoryTable() {
                 <TableHeaderCell>LLM Answer</TableHeaderCell>
                 <TableHeaderCell>Retrieved Documents</TableHeaderCell>
                 <TableHeaderCell>Feedback</TableHeaderCell>
+                <TableHeaderCell>User</TableHeaderCell>
                 <TableHeaderCell>Date</TableHeaderCell>
               </TableRow>
             </TableHead>

web/src/app/ee/admin/performance/query-history/[id]/page.tsx

@@ -1,10 +1,11 @@
-import { Bold, Text, Card, Title, Divider } from "@tremor/react";
+import { Bold, Text, Card, Title, Divider, Italic } from "@tremor/react";
 import { QuerySnapshot } from "../../analytics/types";
 import { buildUrl } from "@/lib/utilsSS";
 import { BackButton } from "./BackButton";
 import { FiBook } from "react-icons/fi";
 import { processCookies } from "@/lib/userSS";
 import { cookies } from "next/headers";
+import { timestampToReadableDate } from "@/lib/dateUtils";
 
 export default async function QueryPage({
   params,
@@ -26,6 +27,11 @@ export default async function QueryPage({
       <Card className="mt-4">
         <Title>Query Details</Title>
 
+        <Text className="flex flex-wrap whitespace-normal mt-1 text-xs">
+          {queryEvent.user_email || "-"},{" "}
+          {timestampToReadableDate(queryEvent.time_created)}
+        </Text>
+
         <Divider />
 
         <div className="flex flex-col gap-y-3">

web/src/lib/dateUtils.ts

@@ -16,3 +16,26 @@ export const timestampToDateString = (timestamp: string) => {
     .padStart(2, "0")}`;
   return formattedDate;
 };
+
+// Options for formatting the date
+const dateOptions: Intl.DateTimeFormatOptions = {
+  year: "numeric",
+  month: "2-digit",
+  day: "2-digit",
+};
+
+// Options for formatting the time
+const timeOptions: Intl.DateTimeFormatOptions = {
+  hour: "numeric",
+  minute: "2-digit",
+  hour12: true, // Use 12-hour format with AM/PM
+};
+
+export const timestampToReadableDate = (timestamp: string) => {
+  const date = new Date(timestamp);
+  return (
+    date.toLocaleDateString(undefined, dateOptions) +
+    ", " +
+    date.toLocaleTimeString(undefined, timeOptions)
+  );
+};