Merge pull request #29 from 2405-team3/evals-ui

feat: evals UI dashboard

eval_pg_utils.py

@@ -9,10 +9,6 @@
 from dotenv import load_dotenv
 load_dotenv()
 
-def filter_by_chatbot(data, chatbot_id):
-    filtered_data = [entry for entry in data if entry[2].get('chatbot_id') == chatbot_id]
-    return filtered_data
-
 def connect_to_db():
     try:
         conn = psycopg2.connect(

evals.py

@@ -83,10 +83,50 @@ def evaluate_and_store_benchmark_data_entry(entry):
 
     pg.insert_scored_benchmark_data(scored_entry)
 
-def get_chat_history(chatbot_id):
+def get_chat_history():
     data = pg.get_data_from('chat_history')
-    data = pg.filter_by_chatbot(data, chatbot_id)
-    return data
+
+    json_data_list = []
+
+    for tuple in data:
+        json_data = {
+            'pg_id': tuple[0],
+            'chatbot_id': tuple[2]['chatbot_id'],
+            'time': tuple[1],
+            'input': tuple[2]['input'],
+            'output': tuple[2]['output'],
+            'context': tuple[2]['context'],
+            'faithfulness': tuple[2]['scores']['faithfulness'],
+            'answer_relevancy': tuple[2]['scores']['answer_relevancy'],
+        }
+        json_data_list.append(json_data)
+
+    return json_data_list
+
+# [
+#     7,
+#     "2024-08-02T19:55:41.279549",
+#     {
+#         "input": "how did giraffes necks get so long",
+#         "output": "Empty Response",
+#         "scores": {
+#             "faithfulness": 0,
+#             "answer_relevancy": 0
+#         },
+#         "context": "",
+#         "chatbot_id": "test1"
+#     }
+# ]
+# const data = [
+#     {
+#         "chatbot_id": "chatbot1",
+#         "input": "how tall are banana trees",
+#         "output": "Banana trees can reach up to 25 feet tall.",
+#         "faithfulness": 0.0,
+#         "answer_relevancy": 0.9541919130393199,
+#         "context": "This is a bunch of sample context"
+#     }
+# ]
 
 def get_benchmark_data():
     data = pg.get_data_from('benchmark_data')

server.py

@@ -149,9 +149,9 @@ async def post_chatbots(request: Request):
     return pipeline_json
 
 
-@app.get('/api/history/{chatbot_id}')
-async def get_evals(chatbot_id):
-    data = evals.get_chat_history(chatbot_id)
+@app.get('/api/history')
+async def get_evals():
+    data = evals.get_chat_history()
     return {"table_data": data}
 
 

systemd/run_system.sh

@@ -1,6 +1,8 @@
 #! /bin/bash
 
 # these commands are used to "reset" systemd after making changes to the unit file (e.g., test.service)
+# sudo systemctl stop test.service
+
 sudo systemctl daemon-reload
 sudo systemctl enable test.service
 sudo systemctl start test.service

ui/package.json

@@ -11,6 +11,7 @@
   },
   "dependencies": {
     "@hookform/resolvers": "^3.9.0",
+    "@radix-ui/react-checkbox": "^1.1.1",
     "@radix-ui/react-dropdown-menu": "^2.1.1",
     "@radix-ui/react-label": "^2.1.0",
     "@radix-ui/react-scroll-area": "^1.1.0",
@@ -19,6 +20,7 @@
     "@radix-ui/react-slot": "^1.1.0",
     "@radix-ui/react-switch": "^1.1.0",
     "@tanstack/react-query": "^5.51.15",
+    "@tanstack/react-table": "^8.20.1",
     "axios": "^1.7.2",
     "class-variance-authority": "^0.7.0",
     "clsx": "^2.1.1",

ui/src/App.tsx

@@ -6,6 +6,7 @@ import { PageKnowledgeBase } from "./components/pages/PageKnowledgeBase.tsx";
 import { PageChatbots } from "./components/pages/PageChatbots.tsx";
 import { PageChatbot } from "./components/pages/PageChatbot.tsx";
 import { PageEvaluations } from "./components/pages/PageEvaluations.tsx";
+// import { DataTableDemo } from "./components/pages/DataTableDemo.tsx"; // testing only
 
 const queryClient = new QueryClient();