Update app.py

app.py

@@ -1,14 +1,24 @@
+import os
 import gradio as gr
 import aiohttp
 import asyncio
-import json, os
+import json
+from datasets import Dataset, DatasetDict, load_dataset, load_from_disk
+from huggingface_hub import HfApi, HfFolder
 
+HEA_API_TOKEN = os.environ.get("HF_API_TOKEN")
 LLM_API = os.environ.get("LLM_API")
 LLM_URL = os.environ.get("LLM_URL")
+USER_ID = "HuggingFace Space"
+HfFolder.save_token(HF_API_TOKEN)
+DATASET_NAME = os.environ.get("DATASET_NAME")
 
-USER_ID = "HuggingFace Space"  # Placeholder user ID
+try:
+    dataset = load_dataset(DATASET_NAME)
+except:
+    dataset = DatasetDict({"feedback": Dataset.from_dict({"user_input": [], "response": [], "feedback_type": [], "improvement": []})})
 
-async def send_chat_message(LLM_URL, LLM_API, user_input):
+async def send_chat_message(user_input):
     payload = {
         "inputs": {},
         "query": user_input,
@@ -16,55 +26,87 @@ async def send_chat_message(LLM_URL, LLM_API, user_input):
         "conversation_id": "",
         "user": USER_ID,
     }
-    print("Sending chat message payload:", payload)  # Debug information
+    print("Sending chat message payload:", payload)
 
     async with aiohttp.ClientSession() as session:
-        async with session.post(
-            url=f"{LLM_URL}/chat-messages",
-            headers={"Authorization": f"Bearer {LLM_API}"},
-            json=payload,
-            timeout=aiohttp.ClientTimeout(total=60)
-        ) as response:
-            if response.status != 200:
-                print(f"Error: {response.status}")
-                return f"Error: {response.status}"
-
-            # Handle the stream of events
-            full_response = []
-            async for line in response.content:
-                line = line.decode('utf-8').strip()
-                if not line:
-                    continue
-                if "data: " not in line:
-                    continue
-                try:
-                    print("Received line:", line)  # Debug information
-                    data = json.loads(line.split("data: ")[1])
-                    if "answer" in data:
-                        full_response.append(data["answer"])
-                except (IndexError, json.JSONDecodeError) as e:
-                    print(f"Error parsing line: {line}, error: {e}")  # Debug information
-                    continue
-
-            if full_response:
-                return ''.join(full_response).strip()
-            else:
-                return "Error: No thought found in the response"
+        try:
+            async with session.post(
+                url=f"{LLM_URL}/chat-messages",
+                headers={"Authorization": f"Bearer {LLM_API}"},
+                json=payload,
+                timeout=aiohttp.ClientTimeout(total=60)
+            ) as response:
+                if response.status != 200:
+                    print(f"Error: {response.status}")
+                    return f"Error: {response.status}"
+
+                full_response = []
+                async for line in response.content:
+                    line = line.decode('utf-8').strip()
+                    if not line:
+                        continue
+                    if "data: " not in line:
+                        continue
+                    try:
+                        data = json.loads(line.split("data: ")[1])
+                        if "answer" in data:
+                            full_response.append(data["answer"])
+                    except (IndexError, json.JSONDecodeError) as e:
+                        print(f"Error parsing line: {line}, error: {e}")
+                        continue
+
+                if full_response:
+                    return ''.join(full_response).strip()
+                else:
+                    return "Error: No thought found in the response"
+        except Exception as e:
+            print(f"Exception: {e}")
+            return f"Exception: {e}"
 
 async def handle_input(user_input):
-    chat_response = await send_chat_message(LLM_URL, LLM_API, user_input)
-    print("Chat response:", chat_response)  # Debug information
+    print(f"Handling input: {user_input}")
+    chat_response = await send_chat_message(user_input)
+    print("Chat response:", chat_response)
     return chat_response
 
 def run_sync(user_input):
+    print(f"Running sync with input: {user_input}")
     return asyncio.run(handle_input(user_input))
 
-# Define Gradio interface
-user_input = gr.Textbox(label='歡迎問我關於「高熵合金」(High-entropy alloys) 的各種疑難雜症')
-examples = [
-    ["AlCoCrFeNi HEA coating 可用怎樣的實驗方法做到 ?"],
-    ["請問high entropy nitride coatings的形成，主要可透過那些元素來熱這個材料形成熱穩定?"]
-]
+def save_feedback(user_input, response, feedback_type, improvement):
+    feedback = {
+        "user_input": user_input,
+        "response": response,
+        "feedback_type": feedback_type,
+        "improvement": improvement
+    }
+    print(f"Saving feedback: {feedback}")
+    # Append to the dataset
+    new_data = {"user_input": [user_input], "response": [response], "feedback_type": [feedback_type], "improvement": [improvement]}
+    global dataset
+    dataset["feedback"] = dataset["feedback"].add_item(new_data)
+    dataset.push_to_hub(DATASET_NAME)
+
+def handle_feedback(response, feedback_type, improvement):
+    feedback = {
+        "response": response,
+        "feedback_type": feedback_type,
+        "improvement": improvement
+    }
+    save_feedback(response, feedback_type, improvement)
+    return "Thank you for your feedback!"
+
+def handle_user_input(user_input):
+    print(f"User input: {user_input}")
+    return run_sync(user_input)
+
+# 读取并显示反馈内容的函数
+def show_feedback():
+    try:
+        feedbacks = dataset["feedback"].to_pandas().to_dict(orient="records")
+        return feedbacks
+    except Exception as e:
+        return f"Error: {e}"
 
 TITLE = """<h1 align="center">Large Language Model (LLM) Playground 💬 <a href='https://support.maicoin.com/zh-TW/support/home' target='_blank'>Cryptocurrency Exchange FAQ</a></h1>"""
 SUBTITLE = """<h2 align="center"><a href='https://www.twman.org' target='_blank'>TonTon Huang Ph.D. @ 2024/06 </a><br></h2>"""
@@ -72,16 +114,41 @@ LINKS = """<a href='https://blog.twman.org/2021/04/ASR.html' target='_blank'>那
 <a href='https://blog.twman.org/2023/07/wsl.html' target='_blank'>用PaddleOCR的PPOCRLabel來微調醫療診斷書和收據</a> | <a href='https://blog.twman.org/2023/07/HugIE.html' target='_blank'>基於機器閱讀理解和指令微調的統一信息抽取框架之診斷書醫囑資訊擷取分析</a><br>
 <a href='https://huggingface.co/spaces/DeepLearning101/High-Entropy-Alloys-FAQ/blob/main/reference.txt' target='_blank'>「高熵合金」(High-entropy alloys) 參考論文</a><br>"""
 
-with gr.Blocks() as iface:
+iface = gr.Blocks()
+
+with iface:
     gr.HTML(TITLE)
     gr.HTML(SUBTITLE)
     gr.HTML(LINKS)
-    gr.Interface(
-        fn=run_sync,
-        inputs=user_input,
-        outputs="text",
-        examples=examples,
-        allow_flagging="never"
+    with gr.Row():
+        user_input = gr.Textbox(label='歡迎問我關於「高熵合金」(High-entropy alloys) 的各種疑難雜症', lines=2, placeholder="在此輸入問題...")
+        submit_button = gr.Button("提交")
+    with gr.Row():
+        response_output = gr.Textbox(label='模型回應', interactive=False)
+    with gr.Row():
+        like_button = gr.Button("👍")
+        dislike_button = gr.Button("👎")
+        improvement_input = gr.Textbox(label='改進建議', placeholder='請輸入如何改進模型回應的建議...')
+    with gr.Row():
+        feedback_output = gr.Textbox(label='反饋結果', interactive=False)
+    with gr.Row():
+        show_feedback_button = gr.Button("查看所有反饋")
+        feedback_display = gr.JSON(label='所有反饋')
+
+    submit_button.click(fn=handle_user_input, inputs=user_input, outputs=response_output)
+
+    like_button.click(
+        fn=lambda response, improvement: handle_feedback(response, "like", improvement),
+        inputs=[response_output, improvement_input],
+        outputs=feedback_output
     )
 
-iface.launch()
+    dislike_button.click(
+        fn=lambda response, improvement: handle_feedback(response, "dislike", improvement),
+        inputs=[response_output, improvement_input],
+        outputs=feedback_output
+    )
+
+    show_feedback_button.click(fn=show_feedback, outputs=feedback_display)
+
+iface.launch()